212 100 1MB
English Pages 426 Year 2014
Anna-Maria De Cesare (Ed.) Frequency, Forms and Functions of Cleft Constructions in Romance and Germanic
Trends in Linguistics Studies and Monographs
Editor Volker Gast Editorial Board Walter Bisang Jan Terje Faarlund Hans Henrich Hock Natalia Levshina Heiko Narrog Matthias Schlesewsky Amir Zeldes Niina Ning Zhang Editor responsible for this volume Jan Terje Faarlund
Volume 281
Frequency, Forms and Functions of Cleft Constructions in Romance and Germanic
Contrastive, Corpus-based studies
Edited by Anna-Maria De Cesare
ISBN 978-3-11-036107-0 e-ISBN (PDF) 978-3-11-036187-2 e-ISBN (EPUB) 978-3-11-039474-0 ISSN 1861-4302 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.dnb.de. © 2014 Walter de Gruyter GmbH, Berlin/Munich/Boston Typesetting: PTP-Berlin Protago-TEX-Production GmbH, Berlin Printing: CPI books GmbH, Leck ♾ Printed on acid-free paper Printed in Germany www.degruyter.com
Table of Contents Anna-Maria De Cesare Introduction | 1
Part I. Italian Cleft Constructions in a Contrastive Perspective Anna-Maria De Cesare Cleft constructions in a contrastive perspective. Towards an operational taxonomy | 9 Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini Form and frequency of Italian Cleft constructions in a corpus of electronic news. A contrastive perspective with French, Spanish, German and English | 49 Davide Garassino Cleft sentences. Italian-English in contrast | 101 Laura Baranzini Pseudo-cleft sentences. Italian-French in contrast | 139 Rocío Agar Marco Pseudo-cleft sentences. Italian-Spanish in contrast | 177 Iørn Korzen Cleft sentences. Italian-Danish in contrast | 217 Giovanna Brianti Cleft sentences. A translation perspective on Italian and French | 277
Part II. Romance and Germanic Cleft Constructions in Contrast Elisabeth Stark Frequency, form and function of Cleft constructions in the Swiss SMS corpus | 325
vi
Table of Contents
Vahram Atayan and Ursula Wienen Inferential cleft constructions in translation. French c’est que in political texts | 345 Volker Gast and Natalia Levshina Motivating w(h)-Clefts in English and German: A hypothesis-driven parallel corpus study | 377 Subject index | 415
Anna-Maria De Cesare
Introduction 1 Cleft constructions in the Romance and Germanic languages: A contrastive perspective Cleft constructions form a class of syntactic structures that include at least three distinct subclasses known as Cleft sentences (or it-clefts), Pseudo-cleft sentences (or wh-clefts) and Reverse pseudo-cleft sentences (or Reverse wh-cleft). The examples given below, taken from Lambrecht (2001: 246), illustrate these constructions in English, but the same three types of clefts exist in a wide array of languages (cf. for instance Smits 1989, Lambrecht 2001, Miller 2006, the papers included in Hartmann and Veenstrav 2013 and Wehr in press): (1)
It is champagne (that) I like.
Cleft sentence / it-cleft
(2)
What I like is champagne.
Pseudo-cleft sentence / wh-cleft
(3)
Champagne is what I like.
Reverse pseudo-cleft sentence / Reverse wh-cleft
In contrast to most of the literature published to date, which describes Cleft constructions by focusing solely on a single language, all the contributions included in this volume focus on at least two languages. The languages taken into account belong to two genetic families: the Romance languages (with discussion of Italian, French, and Spanish data) and the Germanic languages (with data from English, German, Swiss German and Danish). Analyzing clefts cross-linguistically has numerous advantages: from a descriptive point of view, one can highlight the main similarities and differences between clefts of two or more languages and detect phenomena that one would not be able to observe easily when working on a single language; from a theoretical point of view, one can sharpen conceptual tools and test linguistic hypotheses based on the data of one (main) language; finally, from a practical point of view, it allows to produce results that are valuable outputs for different areas of applied linguistics, such as language teaching and translation studies. A clear understanding of the options available when learning the syntactic structures of a foreign language or when translating clefts
Anna-Maria De Cesare, University of Basle
2
Anna-Maria De Cesare
constitutes important knowledge for avoiding calques and interferences and thus for producing natural texts in the target language.
2 Methodology used in the contributions of this volume All the contributions included in this volume are corpus-based. They rely on the analysis of a large and diverse body of authentic linguistic data. The main data is extracted from either comparable corpora (cf. from the Swiss SMS-corpus, Mr. Bean-corpus, ICOCP-corpus) or parallel language corpora (EUROPARL, Le Monde diplomatique). The choice of working with authentic occurrences of clefts rather than with invented examples alone has obvious benefits. Using authentic data allows an analysis of clefts in context, which is a basic requirement if one wants to better understand their discourse functions. In order to allow as much comparability as possible, the papers included in this volume address similar issues: they describe the frequency, formal (syntactic categories and grammatical functions of the cleft constituent), informational (at different levels of information structure, in particular in terms of the activation state of the cleft constituent and the cleft clause) and textual (i.e. discourse functions) properties of one or more cleft types. Moreover, many articles included in the volume are based on the same data (EUROPARL, Le Monde diplomatique, ICOCP-corpus), which allows interesting comparisons of the results obtained for a single type of cleft in different languages, for different clefts in a single language and for a single cleft type in different text types of a single language.
3 Structure of the book The book is divided into two parts. Part I of the volume (entitled Italian Cleft constructions in a contrastive perspective) includes seven contributions in which the Italian language is discussed in a contrastive perspective with another Romance and/or Germanic language. Part II of the book (Romance and Germanic Cleft constructions in contrast) includes three contributions devoted to Cleft constructions in several Romance and/or Germanic languages. The first paper of Part I, by the editor of this volume, proposes a new and more operational taxononmy of clefts, which should be easy to apply when working cross-linguistically with languages that differ from English. The following paper,
Introduction
3
by Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini, describes the forms and assesses the frequency of different types of clefts in five languages (Italian, Spanish, French, German, and English) on the basis of a corpus of online journalistic texts. Based on the same corpus, the following three papers – by Davide Garassino on Cleft sentences in Italian and English, by Laura Baranzini on Pseudo-cleft sentences in Italian and French, and by Rocío Agar Marco on Pseudo-cleft sentences in Italian and Spanish – highlight some of the main similarities and differences in the frequency, forms, and functions of clefts in Italian from a contrastive perspective with English, French, and Spanish, respectively. The two last papers of Part I of the volume provide additional finegrained accounts of Cleft sentences: Iørn Korzen describes Clefts in Italian and Danish using the EUROPARL corpus, while Giovanna Brianti takes a translation perspective on Clefts found in Italian and in French journalistic texts. Part II of the volume opens with the contribution by Elisabeth Stark, who analyzes the frequency, form and function of Cleft constructions in French, Italian, German and Swiss German on the basis of the Swiss-SMS-corpus. The following paper, by Vahram Atayan and Ursula Wienen, studies Inferential clefts in French and in a translation perspective with Italian, Spanish, German and English by looking at data from both Le Monde Diplomatique and the EUROPARL corpus. The volume closes with the contribution by Volker Gast and Natalia Levshina, which is devoted to Pseudo-cleft sentences in English and German and is based on the EUROPARL corpus.
4 Descriptive, theoretical and methodological issues discussed in the present volume The present volume addresses descriptive, theoretical and practical issues related to the use of Cleft constructions in contemporary varieties of Romance and Germanic languages. By presenting, in a single book, data from both the Romance and the Germanic languages, one of the main objectives of the volume is to increase the discussion and knowledge transfer between scholars of close, but institutionally distinct disciplines. As some of the papers included in this volume show, a detailed account of lesser-known Cleft constructions in the Romance and the Germanic languages as well as of Cleft constructions belonging to languages which have a somewhat peripheral status in the current, international linguistic debate is essential for fine-tuning our understanding of the raison d’être of the cleaving mechanism and, more generally, of the macro-communicative strategy of information highlighting.
4
Anna-Maria De Cesare
Cleft constructions have been – and still are – at the heart of numerous descriptive and theoretical linguistic debates taking place within different formal and functional frameworks. These constructions pose for instance notorious challenges to current syntactic theories in respect to the best ways of accounting for their derivation pattern. Although central in the linguistic debate, this volume does not address this issue (we thus refer to some of the latest studies on this aspect, namely Belletti 2008, Frascarelli 2010, Reeve 2012 as well as several papers edited by Hartmann and Veenstra 2013, in particular Frascarelli and Ramaglia 2013, devoted to the derivation of (pseudo)cleft constructions, and Sleeman 2013 who accounts for the licensing of infinitival subject relative clauses by Italian cleft constituents). It should also be noted that, although some of the papers included in this volume make important claims on the semantics of clefts, no study is specifically devoted to the different meaning component associated with clefts, in particular to the nature and status of the exhaustiveness component (on these issues, cf. among others Horn 1981, Dufter 2009, De Cesare and Garassino in press). The questions addressed in this volume consider the frequency, distribution (both textual and geographic), forms (described in terms of surface structure) and functions of Cleft constructions. Specifically, from a descriptive point of view, this volume: 1. sheds light on the frequency of different Cleft constructions across two or more languages that belong to the same or to distinct genetic families; several text types, registers and genres are explored (most of the attention is devoted to journalistic prose – both in the form of daily and monthly press – and to political texts, in particular to the speeches delivered at the European parliament; one contribution, by Stark, explores the use of clefts in text messages); 2. includes detailed descriptions of cleft types that have not yet been much at the center of attention: Atayan & Wienen study Inferential clefts, a type of construction that has not been the object of much scrutiny; De Cesare et al. devote some space to Italian Reverse-cleft sentences, a type of cleft that is again not as well studied as classic Cleft sentences; to a lesser extent, De Cesare et al. also discuss Reverse pseudo-cleft sentences; 3. offers rich linguistic data on Cleft constructions that are only marginally taken into account in the general literature on these syntactic structures, such as Italian and Spanish Pseudo-clefts (De Cesare et al.), as well as German Cleft constructions in general (De Cesare et al.; Gast & Levshina); 4. deepens our knowledge of the forms and functions of Cleft constructions in individual Romance and Germanic languages, while at the same time highlighting some of the main formal and functional similarities and differences between Cleft constructions in the Romance and Germanic languages, as well as within the Romance and the Germanic language families.
Introduction
5
From a theoretical point of view, this volume sheds new light on: 1. the intension of the class of Cleft constructions (i.e., on their basic formal, semantic and pragmatic properties), as well as its extension (i.e., on the external and internal boundaries of this group of syntactic forms; cf. De Cesare; Agar Marco); 2. the relationship between the linguistic form of a given cleft and its information and discourse properties (focusing function, etc.; cf. Agar Marco; Garassino); 3. the so-called compensation mechanism principle (cf. De Cesare et al.; Brianti; Gast & Levshina), which stipulates that only languages with (relatively) rigid constituent order (French and English etc.) possess different types of Cleft constructions and regularly use these clefts, while languages with free or semi-free constituent order (Italian, Spanish, Portuguese) rely on more economic focusing devices to achieve information highlighting, namely prosody (primarily in spoken communication), clause internal reorganization of basic constituents (fronting of the direct / indirect object, subject inversion); 4. the use of clefts in V2 languages such as German and Danish (cf. Gast & Levshina and Korzen, respectively). In addition to descriptive and theoretical issues, this volume also addresses important methodological questions. Crucially, several contributions discuss the most appropriate corpora to use when working with more than one language simultaneously, weighing the pros and cons of using comparable as opposed to parallel (i.e., translation) corpora (cf. Brianti).
Acknowledgments This volume has been conceived within the framework of the research project Italian Constituent Order in a Contrastive Perspective (ICOCP), based at the University of Basle and funded by the generous support of the Swiss National Science Foundation (research project PP00P1-133716/1). The first five contributions included in the volume present results obtained within the ICOCP project. I am most grateful to all the members and ex-members of the ICOCP research group – i.e., listed in alphabetical order, Rocío Agar Marco, Ana Albom, Laura Baranzini, Doriana Cimmino, and Davide Garassino – for their collaborative work and help in the preparation of the manuscript. A special thank you also goes to Claudia Ricci, who translated the papers by Rocío Agar Marco and Laura Baranzini from Italian into English.
6
Anna-Maria De Cesare
I also wish to thank the colleagues who accepted to review one or more of the papers included in this volume: Jörn Albrecht, Margarita Borreguero Zuloaga, Andreea Calude, Alberto Gil, Michael Herslund, Florent Perek, Tinka Reichmann, Nadine Rentel, Ursula Reutner, Carlo Enrico Roggia and Marie-Noëlle Roubaud. My acknowlegments also go to two additional anonymous reviewers, chosen by one of the editors of the series Trends in Linguistics, for their valuable comments on an earlier draft of the manuscript. Last but not least, I would like to thank all the authors of this volume for having participated in this project with enthusiasm.
References Belletti, Adriana. 2008. The CP of Clefts. Rivista di Grammatica Generativa 33. 191–204. De Cesare, Anna-Maria & Davide Garassino. In press. On the status of exhaustiveness in cleft sentences: A data-driven and cross-linguistic study of English also- / only-clefts and Italian anche- / solo-clefts. Folia Linguistica, pp. 36. Dufter, Andreas. 2009. Clefting and discourse organization: Comparing Germanic and Romance. In Andreas Dufter & Daniel Jacob (eds.), Focus and Background in Romance Languages, 83–121. Amsterdam & Philadelphia: John Benjamins. Frascarelli, Mara. 2010. Narrow Focus, clefting and predicate inversion. Lingua 120 (9). 2121–2147. Frascarelli, Mara & Francesca Ramaglia. 2013. (Pseudo)clefts at the syntax-prosody-discourse interface. In Katharina Hartmann & Tonjes Veenstra (eds.), Cleft Structures. 97–138. Amsterdam & Philadelphia: John Benjamins. Hartmann, Katharina & Tonjes Veenstrav (eds.). 2013. Cleft structures. Amsterdam & Philadelphia: John Benjamins. Horn, Laurence R. 1981. Exhaustiveness and the semantics of clefts. In Victoria Burke & James Pustejovsky (eds.), Papers from the 11th Annual Meeting of NELS, 124–142. Amherst: GLSA. Lambrecht, Knud. 2001. A framework for the analysis of cleft constructions. Linguistics 39. 463–561. Miller, Jim. 2006. Focus in the languages of Europe. In Giuliano Bernini & Marcia L. Schwartz (eds.), Pragmatic Organization of Discourse in the Languages of Europe, 121–214. Berlin & New York: Mouton de Gruyter. Reeve, Matthew. 2012. Clefts and their Relatives. Amsterdam & Philadelphia: John Benjamins. Sleeman, Petra. 2013. Italian clefts and the licensing of infinitival subject relatives. In Katharina Hartmann & Tonjes Veenstra (eds.), Cleft Structures. 319–342. Amsterdam & Philadelphia: John Benjamins. Smits, Rik J. C. 1989. Eurogrammar: The Relative and Cleft Constructions of the Germanic and Romance Languages. Dordrecht: Foris. Wehr, Barbara. In press. Cleft constructions from a typological perspective. In Sabine Diao-Klaeger & Britta Thörle (eds.), Linguistique interactionnelle contrastive. Grammaire et interaction dans les langues romanes. Tübingen: Stauffenburg.
Part I. Italian Cleft Constructions in a Contrastive Perspective
Anna-Maria De Cesare*
Cleft constructions in a contrastive perspective Towards an operational taxonomy
1 Introduction It is well known and accepted in the literature dealing with a wide array of typologically different languages that Cleft constructions are a family of forms that subsume three main types (for studies considering languages of several genetic families, see for instance Lambrecht 2001; Miller 2006): Cleft sentences (also called it-clefts), Pseudo-cleft sentences (i.e. wh-clefts) and Reverse pseudo-cleft sentences (or reverse wh-clefts).¹ These forms, illustrated in (1) to (3)² on the basis of Lambrecht (2001: 467), share a set of central formal and semantic properties. They are biclausal structures (i.e., they contain two clauses, while being semantically equivalent to a monoclausal sentence: I like champagne) and have a specificational (also called identificational) meaning (i.e. they specify a value for a variable: I like something and this something is champagne). As far as their function
* The research presented in this paper has been funded by the Swiss National Science Foundation (Project PP00P1-133716/1, Italian Constituent Order in a Contrastive Perspective, in short ICOCP). I wish to thank Andreea Calude and two anonymous reviewers for their helpful comments on a previous draft of this paper. 1 As well as these three types of clefts, there are other cleft forms, for instance Inferential clefts (see Atayan and Wienen in this volume), Pseudo-conditional clefts (if I insist it’s because I think I am right), Presentational clefts (lt. ho il piede che è gonfio ‘I have a swollen foot’) etc. As the status of the clefts illustrated in (1) to (3) is already problematic cross-linguistically, and because the contributions offered in this volume mainly involve the clefts illustrated in (1) and (2), we will not say much on these other types of clefts here. In addition, we will only focus on clefts based on the copula be (or equivalent forms in other languages), as the status of syntactic structures with other verb forms is debated. 2 The examples provided in this paper will be taken either from the existing literature (maintaining, whenever there are some, the small capitals indicating the accented part of the cleft) or from other sources (mainly from written texts; the label ICOCP refers to a corpus described in detail in De Cesare et al. in this volume). These sources will always be provided in parentheses after the example or specified before the example(s). Examples with no reference to a source have to be interpreted as invented by the author of this paper. Anna-Maria De Cesare, University of Basle
10
Anna-Maria De Cesare
is concerned, clefts are typically associated with focusing. As can be observed in examples (1) to (3), the focus tends to coincide with the element highlighted with the small capitals; the small capitals also indicate the location of the main sentence accent: (1)
It is champagne (that) I like.
Cleft sentence / it-cleft
(2)
What I like is champagne.
Pseudo-cleft sentence / wh-cleft
(3)
Champagne is what I like.
Reverse pseudo-cleft sentence / Reverse wh-cleft
In this paper, we will use the following terminology to refer to the main components of these clefts: we will refer to the main verb (a form of the verb be in examples [1] to [3]) with the term copula; to the highlighted element (i.e. champagne) with the term cleft constituent and to the relative clause (that / what / ø I like) with the term cleft clause.³ As we will see in this chapter, both the taxonomy and the labels used in the literature to refer to each subtype of cleft are based primarily on the form of the different cleft types and, crucially, the basis of the formal description is the English language. This anglocentric view poses a number of important problems for the description of languages that differ from English, while at the same time making any large-scale cross-linguistic attempt to evaluate the frequency of clefts, analyze the forms available in the repertoire of different language families and describe the functions of these clefts quite intricate. The aim of the present study is thus to propose a more operational taxonomy of Cleft constructions, i.e., a taxonomy that can easily be used when working cross-linguistically both
3 The terminology used in the literature to refer to the cleft components varies quite significantly (on this issue, also see note 23). The term cleft / clefted constituent, which is used for instance in Hedberg (1988) and Calude (2009), is called differently in other studies, in particular according to the point of view that is adopted: Collins (1991: 2), for instance, prefers to use the term highlighted element, which he considers to be “neutral as to the semantic/syntactic/textual/logical role of the constituent in question” (p. 217). In his view, this element should be called identifier in propositional semantic terms, complement of the copula be (or post-copular constituent) in syntactic terms, new or comment in textual terms and focus in logical terms (cf. Collins 1991: 217). Again in line with Hedberg (1988) and Calude (2009), the same is true for what we call cleft clause: Declerck (1984) calls this part of the cleft wh-that clause, and Collins (1991) and Lambrecht (2001) label it relative clause. All of these labels have advantages and disadvantages that we cannot discuss in detail here. In a way, these labels are therefore to be interpreted as a practical, compromise solution.
Cleft constructions in a contrastive perspective
11
with English and languages that differ from English. Our proposal is to classify Cleft constructions in the first place according to the position occupied by the so-called cleft constituent (i.e., the noun champagne in examples [1] to [3]) in the syntactic structure. We will therefore differentiate between three types of Cleft constructions: with initial cleft constituent (as in example [3]), with medial cleft constituent (as in example [1]) and with final cleft constituent (as in example [2]). This taxonomy is inspired by a proposal made by Calude (2009: 143) for English, which we have extended here to other languages. In this paper, we will focus primarily on Romance (mainly Italian, French and Spanish) and Germanic languages (mainly English and German). Amongst the languages taken into account in this study, special attention will be given to Italian. This article is organized as follows: Section 2 presents the classic taxonomy of clefts and discusses some of the problems encountered with its application to languages that differ from English; in Section 3 a new taxonomy of clefts based on the position of the cleft constituent in the syntactic structure is proposed and some arguments supporting this new taxonomy are provided; Section 4 concludes by summarizing the main findings of the paper and pointing out some open questions.
2 Cleft construction: The classic taxonomy A great number of studies identify two main types of clefts: Cleft sentences and Pseudo-cleft sentences (cf., among many others, the classic studies from Prince 1978; Sornicola 1988; Smits 1989; more recently, see Roggia 2008, 2009 and Panunzi 2009 on Italian; as well as Valentini 2012 on clefts in Bergamasco, a Gallo-Italian dialect spoken in Northern Italy). Thus, in the first place, the main distinction to be made is between these two types of clefts. In a second step, a distinction is made between two types of Pseudo-clefts: Pseudo-clefts proper (or non-inverted, according to the characterization of Declerck 1984 and Geluykens 1988) and Reverse pseudo-clefts.⁴ The classic taxonomy of Cleft constructions is recalled in Figure 1.
4 Cf. Geluykens (1988: 823): “Two formal types of CC [cleft construction] will be distinguished, the IT cleft […] and the pseudocleft; the latter has a subtype with inverted order”.
12
Anna-Maria De Cesare
CLEFT CONSTRUCTIONS
Cleft sentences (IT-clefts)
Pseudo-cleft sentences (WH-clefts)
Pseudo-clefts proper (WH-clefts proper)
Reverse pseudo-clefts (reverse WH-clefts)
Figure 1: Cleft constructions: The classic taxonomy.
In both steps of the distinction, the decisive criteria are formal: Cleft sentences and Pseudo-cleft sentences are distinguished on the basis of the form that introduces the whole structure (it vs. what); in turn, the two types of pseudo-clefts are distinguished on the basis of the location of the wh-clause in the structure: it is found in initial and final position, respectively.⁵ These three subtypes of clefts differ on the basis of a number of other formal properties, such as the form that introduces the cleft clause (that vs. what), and the position of the cleft constituent (in medial, final and initial position in Cleft sentences, Pseudo-clefts proper and Reverse pseudo-clefts, respectively).⁶ In the following paragraphs, these
5 The element taken as a reference point in accounting for the linear order of Reverse pseudoclefts is not always the same. In some studies, the reference point coincides with the cleft constituent (cf. Biber et al. 1999: 960); in others, it coincides with the copula (cf. Lambrecht 2001: 467: “sentences such as Jespersen’s [1937] above-quoted Champagne is what I like best, in which the headless relative appears in postcopular position, were called ‘Reverse pseudo-clefts’”). 6 All three cleft types also differ in regard to the syntactic categories and functions of the cleft constituent. Cf. Prince (1978: 884): “Both [it-clefts and wh-clefts] readily accept an NP […], but an ADV [It was then that I became a young revolutionary] or a PP [It is against pardoning these that many protest] commonly occur mainly (though not only) in it-clefts, while a VP [What that does is tend to rob Ervin and the Grand Jury with yet a third investigation group] or an S [What you are saying is that the President was involved] commonly occur in WH-clefts”; Cf. also Biber et al. (1999: 960): “Wh-clefts are less flexible than it-clefts in that they cannot be used to focus on a prepositional phrase, an adverb phrase, or an adverbial clause […]. Wh-clefts, on the other hand, permit focus on a nominal clause and on the verb plus accompanying elements in the predicate […]; this possibility is excluded with it-clefts”.
Cleft constructions in a contrastive perspective
13
properties will be described in more detail. Specifically, Section 2.1 will focus on the forms that introduce the Cleft sentences and Pseudo-cleft sentences and Section 2.2 discusses the forms that introduce the cleft clause of these two types of clefts. Along the way, and in particular in Section 2.3, we will point out some problems related to the cross-linguistic classification of clefts based on the criteria we just mentioned.
2.1 Distinguishing Cleft constructions on the basis of the introducer 2.1.1 From English to other languages The first criterion that is used in the classification of clefts is the form that introduces the cleft structure as a whole. As we have seen in examples (1) and (2), recalled below in (4) and (5a), in English we have the following picture: a Cleft sentence is introduced by the pronoun it,⁷ while a Pseudo-cleft sentence is opened by a free relative pronoun,⁸ i.e. by what in (5a), or by who (5b), when (5c), why (5d), where (5e) and how (5f); according to Collins (1991: 27), a Pseudo-cleft can also be opened by which. (4)
It is champagne (that) I like.
Cleft sentence / it-cleft
(5)
a.
Pseudo-cleft sentence / wh-cleft
What I like is champagne.
b. Who(m) she saw last week was an old friend. c.
When the band went home was after dark.
d. Why I left was because of my headache.
7 Lambrecht (2001: 468) also considers the cases in which the pronoun introducing a cleft is not empty: “Exceptionally, it can be a semantically nonempty pronoun (I, you, etc.) which loses some or all of its meaning within the [cleft construction], or which is semantically redundant with an element elsewhere in the sentence (cf. [I have my NEIGHBOR who’s black]).” We will not discuss these cases here. 8 In this chapter, we will use the labels free relative pronoun (in other chapters also fused relative pronoun) and wh-pronoun interchangeably. As we can see in examples (5), we are referring to a relative clause opened by a pronoun that does not have an antecedent. A list of free relative pronouns used in English, German, Italian, French, Spanish is provided in Table 2 of § 3.1.2.
14
Anna-Maria De Cesare
e.
Where my parents went on holidays was their bach.
f.
How I chose it was by looking at the entire list of courses. (Examples b–f are from Calude 2009: 132)
In German, the situation is very similar to English, as the class of Cleft sentences is introduced by the pronoun es ‘it’ and the class of Pseudo-clefts by a free relative pronoun (by the form was ‘what’, wer ‘who’, wo ‘where’, etc.; on these forms, see Gast and Levshina in this volume): (6)
Es war Hans, der mir half. It was Hans REL me helped ‘It was Hans who helped me.’ (Erdmann 1990a: 69)
(7)
Was ihn störte, war ihre Gleichgültigkeit. What him bothered was her indifference ‘What bothered him was her indifference.’ (Erdmann 1990a: 69)
In the Romance languages, the picture is somewhat different. Although Clefts and Pseudo-clefts are of course also identified as two stable Cleft constructions classes in Italian, French and Spanish, the forms that introduce them do not necessarily coincide with an empty pronoun and a free relative pronoun, respectively. In French, a parallel to English clefts can be drawn only for the category of Cleft sentences: French Cleft sentences are opened by the pronoun ce ‘it’, as in (8), while Pseudo-clefts can theoretically be opened by a free relative pronoun such as qui ‘who’ (9), où ‘where’ (10), comment ‘how’ (11), pourquoi ‘why’ (12), quand ‘when’ (13) (cf. Roubaud 2000 for authentic examples taken mainly from spoken registers), but start most of the time with the complex pronoun ce qui ‘what’ (for quantitative data related to a corpus of electronic news, see De Cesare et al. in this volume and Baranzini in this volume): (8)
c’ est mon pied qui me it is my foot that me ‘it’s my foot that hurts.’
fait mal. hurts (Lambrecht 2001: 486)
Cleft constructions in a contrastive perspective
(9)
qui Nixon a choisi c’ est who Nixon chose it is ‘who Nixon chose was Agnew’
15
Agnew Agnew (Roubaud 2000: 26)
(10)
où je vais c’ est au cours de linguistique where I go it is to the class of linguistics ‘where I go is to the linguistics class’ (Roubaud 2000: 9)
(11)
comment il l’ a su c’ est qu’ elle lui a donné how he it found out it is REL she him gave le numéro the number ‘how he came to know it is that she gave him the number’ (Roubaud 2000: 14)
(12)
pourquoi il a fait ça c’ est parce qu’ il est fou why he did this it is because he is crazy ‘why he did this is because he’s crazy’ (Roubaud 2000: 14)
(13)
quand ils m’ ont énervé c’ est quand ils m’ when they me have bothered it is when they me ont répondu answered ‘when they got on my nerves is when they answered me’ (Roubaud 2000: 14)
(14)
ce qui m’ intéresse c’ est le pognon what me interests it is money ‘what I am interested in is money’ (Roubaud 2000: 13)
In Italian and Spanish, a parallel to English is even more far-reaching. As we already said, in these two languages, Cleft Sentences are not opened by any form of pronoun and the situation regarding the class of Pseudo-clefts is similar to French: Pseudo-clefts can be introduced either by a free relative pronoun (It. chi ‘who’, Sp. quien ‘who’; note that there is no Romance equivalent to the widespread E. what) or – and this is the preferred option (cf. De Cesare et al. in this
16
Anna-Maria De Cesare
volume) – by a relative headed by another pronoun (cf. It. quello che ‘what’, Sp. lo que ‘what’, etc.). Here are some representative examples of clefts and Pseudoclefts in Italian and Spanish, respectively: (15)
[ø] è mio padre che ha aperto la finestra [ø] is my father that opened the window ‘it’s my father that/who opened the window’ (Berretta 2002: 16)
(16)
[ø] es este coche el que compré [ø] is this car that bought.1sg ‘it’s this car that I bought’ (Pinedo 2000: 131)
(17)
chi va a Roma sono io who goes to Rome is I ‘the one who goes to Rome is me’ (Sornicola 1988: 345)
(18)
quien va a Roma soy yo who goes to Rome is I ‘the one who goes to Rome is me’ (Sornicola 1988: 345)
(19)
quello che mi è venuto in mente è di partire domani what me came to mind is to leave tomorrow ‘what came to my mind is to leave tomorrow’ (Salvi 1991: 180)
(20)
lo que compré es what bought.1sg is ‘what I bought is this car’
este coche this car
In light of these cross-linguistic grammatical differences, we consider that it is better to avoid using the labels it-cleft and wh-cleft in the description of the languages that differ from English (e.g. Italian, French, and Spanish) because these labels are not transparent (either) and therefore their use is relatively counterintuitive. The label it-cleft is not appropriate for pro-drop languages such as Italian and Spanish, because these languages are not opened by any form of pronoun. In our view, stating, as Lambrecht (2001) does, that the matrix subject
Cleft constructions in a contrastive perspective
17
of an it-cleft coincides with an inflectional morpheme in Italian and Spanish (p. 464) demands to stretch the notion of “matrix subject” as well as the label it-clefts to a point where they no longer seem operational. In turn, the label whcleft is problematic because in the case of the Romance languages it is also supposed to refer to constructions that are not opened by a wh-word (i.e., by a free relative pronoun of the form chi / qui / quien ‘who’ or dove / où / donde ‘where’ etc. in Italian, French and Spanish, respectively) but by a complex pronoun, such as It. quello che, Fr. ce que, Sp. lo que ‘what’. Using the other set of labels – i.e., Cleft sentence, Pseudo-cleft sentence and Reverse pseudo-cleft sentence – is not unproblematic either. Besides their lack of transparency and the fact that, for some, the two last terms misguidedly lead to the view that the syntactic structures to which they refer are not real clefts, they cover syntactic forms that differ quite a bit from one language to another (a fact that is also true for the alternative set of labels).
2.1.2 On the external and internal boundaries of the classic taxonomy of clefts As is well known, the boundaries of the subclasses of clefts, in particular of Cleft sentences and Pseudo-cleft sentences, are problematic. A first debated issue is whether temporal syntactic structures such as (21) to (25) are clefts or not. Some authors take them into account, while others do not (see Berretta 1994 for Italian; Grewendorf and Poletto 1989: 115; on temporal clefts in general, also see Valentini 2012). It is not clear if these structures are real clefts or if they are predicative constructions in which the initial pronoun is anaphoric (cf. E. it, Fr. ce ‘it’, G. es ‘it’) or else in which there is an unexpressed subject that could be made explicit through the use of a demonstrative pronoun (as in Italian questa è la seconda volta che… ‘this is the second time that’; the same is true for Spanish): (21)
it was the last time she would bow to leadership pressure. (ICOCP, AP)
(22)
È la seconda volta in tre mesi che il tribunale è colpito da una deflagrazione (ICOCP, ANSA) ‘It is the second time in three months that the court is hit by a blast’
(23)
Ce n’est pas la première fois qu’Elio Di Rupo dramatise la situation politique. (ICOCP, lalibre.be) ‘It is not the first time that Elio Di Rupo dramatizes the political situation.’
18
Anna-Maria De Cesare
(24)
Es la primera vez que la madre y su entorno son tan claros en sus acusaciones. (ICOCP, elpais.com) ‘It is the first time that the mother and the ones around her are so clear in their allegations.’
(25)
Es war das erste Mal, dass er öffentlich in Erscheinung trat (ICOCP, swissinfo.ch) ‘It was the first time that he made a public appearance’
Very much debated are in particular the boundaries of the class of Pseudo-cleft sentences.⁹ Classification problems are encountered, for instance, with the syntactic structures illustrated in (26) and (27), i.e., with forms of clefts that are not opened by one of the English free relatives what, who, which, where, why, how, and when: (26)
All the car needs is a new battery. (Collins 1991: 32)
(27)
all that is needed is some space, a mat and a few favorite things (chicagotribune.com)
If it is generally recognized that these syntactic structures are clefts, their specific nature is debated. In some studies, clefts opened by all belong to the class of Pseudo-clefts; in Collins (1991: 27), for instance, this cleft type forms a subtype of Pseudo-cleft called all-cleft¹⁰ (see also Biber et al. 1999: 961). In other studies, these clefts form an independent category; in Erdmann (1990a), they are called fokussierende all-Sätze (i.e. focusing all-clefts). Note that Erdmann (1990a) extends this proposal to German clefts such as (28). However, in German, an alles-cleft ‘allcleft’ does not match neatly the English all-cleft, as only in the former language is the cleft opened by both a complex form, based on alles ‘all’ and a wh-form (alles was ‘all what’). Consider the different acceptability of examples (28) and (29), as opposed to (26) and (27):
9 On this issue, see in particular the chapter entitled The ‘th-cleft’ debate in Calude (2009: 130– 144). 10 For Collins (1991: 32), there is a close semantic relationship between all-clefts and th-clefts (i.e., clefts introduced by the definite article and followed by a generic noun; we will say more on these clefts below) with a relative clause headed by the noun phrase the only thing. Consider the following two examples: all you need is love; the only thing you need is love. Accordingly, for Collins (1991), both all-clefts and th-clefts opened by the noun phrase the only thing are considered subtypes of Pseudo-cleft sentences.
Cleft constructions in a contrastive perspective
(28)
Alles was du brauchst, ist All what you need.2sg is ‘All you need is love.’
19
Liebe. love (Erdmann 1990a: 69)
(29) * Alles du brauchst, ist Liebe. All you need.2sg is love ‘All you need is love.’ (30)
All you need is love.
In the Romance languages, English all-clefts do not have a one-to-one structural equivalent either. From a semantic and functional point of view, English all-clefts correspond to Romance clefts introduced by a relative headed by a complex pronominal form (It. tutto quello che, lit. ‘all what’ / tutto ciò che, lit. ‘all what’, Fr. tout ce que, lit. ‘all what’, and Sp. todo lo que, lit. again, ‘all what’; note that all the following examples are quotations): (31)
“Tutto quello che voglio è essere dimenticata.” (ilgiornale.it) “All what want.prs1sg is to be forgotten “ ‘All I want is to be forgotten.’ ”
(32)
“Tout ce que je veux, c’est que Gottéron “All what I want it is that Gottéron “‘All I what is that Gottéron wins the title!’”
(33)
“Todo lo que quiero es huir “All what want.prs1sg is flee “‘All I want is flee this city’”
gagne le titre!” wins the title (laliberte.ch)
de esta ciudad” from this city (elpais.com)
In addition to this problematic case, in the literature there is also no agreement on the status of the syntactic structures opened by a generic noun phrase, such as the one in the following example: (34)
The one I like is Robert (Lambrecht 2001: 469)
20
Anna-Maria De Cesare
In some studies (see for instance Prince 1978),¹¹ these syntactic structures are not taken into account, while in other studies they belong to a new, independent category of clefts. In the classification of Cleft constructions proposed by Erdmann (1990a), for instance, these clefts form a fifth group called focusing copular sentences with generalized nouns (G. “generalisierenden Nomen”) and are claimed to occur in both English and German.¹² Here are two parallel examples in English and in German provided by the author: (35)
The one who helped you was me (Erdmann 1990a: 70)
(36)
Der dir geholfen hat, war ich (Erdmann 1990a: 70)
In yet other studies, syntactic structures of the type illustrated in (35) and (36) are considered to be Pseudo-cleft sentences (cf. Akmajian 1979: 18, cited in Erdmann 1990b: 175). However, their status in this class varies in the literature, as they are presented in some cases as prototypical examples of Pseudo-clefts (in Sornicola 1988: 343, the first example provided to illustrate the class of Pseudo-clefts is the one who wrote the book is me) and in others only as a peripheral subtype of Pseudo-clefts. As mentioned earlier (cf. footnote 10), these structures belong to the subtype of Pseudo-clefts called th-clefts notably by Collins (1991: 27).¹³ This label is based on the fact that these clefts are opened by the definite article the, and followed by a generic noun corresponding to a free relative (what = the + thing; who = the + one; where = the + place; when = the + time; why = the + reason; how = the + way; for German, Erdmann 1990a mentions the sequence der + Grund ‘the reason’ coinciding with the free relative warum).¹⁴ 11 On the clefts of the type illustrated in (34), Prince (1978: 883, N. 1) states the following: “I am (perhaps arbitrarily) defining WH-clefts as sentences of the form What S – Ci is / was Ci, where S – Ci = Sentence minus Constituenti. That is, I am excluding all clefts whose subject clause has a lexical head (the one, the thing…)”. 12 In Erdmann’s (1990a) taxonomy, we find it-clefts, wh-clefts, Reverse wh-clefts, all-clefts and clefts with generalized nouns. 13 Note that the label th-cleft is already used in Horn (1981: 132) to describe examples such as The thing that Mary ate was a pizza. In addition, it should be noted that, in the literature, the label th-cleft is also applied to a completely different type of cleft. In Ball’s (1977) study, it refers to a type of Cleft sentence opened by a demonstrative pronoun (rather than by the form it): that was Timothy laughing at you (ex. from Calude 2009: 135). 14 Another issue is whether syntactic structures opened by generic heads that are modified by an adjective or an adverb should be considered clefts or not. Collins (1991), for instance, considers as clefts syntactic structures headed by lexically empty pro-forms (the one, the thing, the place etc.) that are modified by numerals and quantifiers (The second thing which the university does is to give its students a special experience in which they gain an abiding insight into a university’s
Cleft constructions in a contrastive perspective
21
Syntactic and semantic structures such as (34) to (36) are also attested in the Romance languages. However, the forms that can be considered to be typical in Italian, French and Spanish are not the same. As can be observed on the basis of the following examples, in Italian and French we find a construction headed by the generic noun phrase la persona ‘the person’ and followed by a relative pronoun (It. che, Fr. que, both corresponding to E. ‘that’), while Spanish prefers a form in which the generic lexical noun is not expressed: (37)
la persona che ha aperto la finestra è tuo padre ‘the person who opened the window is your father’
(38)
la personne que Nixon a choisie c’est Agnew (Roubaud 2000: 26) ‘the person that Nixon chose is Agnew’
(39)
la que me falta es Nuria Espert (revistavanityfair.es, 15.11.2012) ‘the one I am missing is Nuria Espert’
The inclusion of syntactic structures such as (34) to (39) in the class of clefts is debated, as it is sometimes claimed that these constructions could be interpreted as monoclausal, predicative structures, which involve an anaphoric noun phrase (the one, la persona ‘the person’ etc.), rather than as biclausal, specificational Cleft constructions, based on a non-anaphoric noun phrase (see, in particular, Calude 2009: 132). Even more debated is whether other classes of lexical nouns can be accepted to head Pseudo-clefts or to be involved in a Cleft sentence as well as a Reverse pseudo-cleft sentence. In some studies, it is proposed that the list of lexical nouns involved in a cleft (i.e., opening a cleft clause) should not be restricted whatsoever or should not be restricted to the generic nouns coinciding with a wh-word (what = the + thing; who = the + one etc.). This view is adopted, for instance, by Halliday (1967: 234) for English and by Salvi (1991: 180) for Italian.¹⁵ Thus, both authors consider that structures such as (40) to (43) are instances of Pseudo-cleft perspective; The only time I can RELAX is when I ACCEPT this, p. 31). Other modifiers, such as evaluative adjectives (The best thing I can do is lie still, p. 31) are not included in his study. The criterion on which this choice is based regards the availability of a non-cleft version (p. 31): according to Collins (1991), only the syntactic structures that have a monoclausal equivalent should be regarded as real instances of clefts (for Collins, the monoclausal equivalents of the two examples provided above are secondly, the university gives; I can relax only when I accept this, respectively; cf. Collins 1991: 31). 15 Also see Panunzi (2010: 118), who provides examples of Pseudo-clefts opened by quello che and ciò che ‘what’, la cosa che ‘the thing that’, but also la storia che ‘the story that’ (cf. la storia
22
Anna-Maria De Cesare
sentences (for Salvi 1991 the noun heading what he considers to be a Pseudo-cleft should minimally function as hyperonym in relation to the postcopular element, as in the color > red): (40)
The colour I like best is red.
(41)
The movie I like best is Easy Rider.
(42)
Il colore che preferisco è il rosso. The color that prefer.prs.1sg is the red ‘The color I prefer is red.’ (Salvi 1991: 180)
(43)
Il mese in cui verrà è marzo. The month in which will.come.3sg is March ‘The month in which s/he will come is March.’ (Salvi 1991: 177)
This view, however, is very controversial because with lexically less generic nouns than one, thing etc., i.e., with nouns such as color, movie etc., it is even harder to draw a line between biclausal, specificational constructions and monoclausal sentences (on this issue, cf. for instance Berretta 1996: 121 for Italian; the example she provides is the following: La ragazza che ti ho presentato è la mia compagna di scuola ‘the young woman that I introduced you to is my schoolmate’). If the structures given in (34) to (43) seem to be biclausal and specificational constructions, from a purely syntactic point of view they differ from the traditional Pseudo-clefts, opened by a free wh-word. An important argument to keep them separate from the classic Pseudo-clefts is that, besides the fact that they are not headed by the same type of generic pronouns, they cannot be declefted easily, as shown in (44) and (45). In these cases, either important informations of the original propositional content are lost in the syntactic conversion (see [44a] in respect to [44b]) or the original sentence has to be greatly restructured from a syntactic point of view (see [45], in which the adjective last becomes an adverb):¹⁶
che vogliamo raccontarvi / non è quella di guerre commerciali, lit. ‘the story that we would like to tell you / is not the one concerning commercial wars’). 16 Cf. also Calude (2009: 133), who claims that some propositional content can be lost in the conversion. Additional observations on this issue can be found in Gast and Wiechmann (2012) on declefting English and German Pseudo-clefts.
Cleft constructions in a contrastive perspective
(44)
a.
23
Il colore che preferisco è il rosso. The color that prefer.prs.1sg is the red ‘The color I prefer is red.’ (Salvi 1991: 180)
b. Preferisco prefer.prs.1sg ‘I prefer red.’ (45)
a.
il rosso. the red
The last person to die while active in the major leagues was Nick Adenhart (ICOCP, nytimes.com)
b. Nick Adenhart died last while active in the major leagues Based on their differences with classic Pseudo-clefts, we thus believe that the syntactic structures opened by a semantically (non) generic noun such as the ones illustrated in (37) to (45) are at best peripheral to the class of Cleft constructions. That said, structures such as (37) to (45) seem to be functionally similar to Pseudo-clefts and it can be interesting to consider them in pragmatic and textual accounts of clefts (as does Agar Marco in this volume). Classic clefts also bear some functional resemblance with syntactic structures (also considered clefts by some researchers) in which the cleft clause is opened by a (non) generic noun. Consider the text provided in (46). In this text, we find two similar Cleft constructions (both with the cleft constituent after the copula) that occur in two consecutive Utterances. As can be easily noted, although the two Cleft constructions are syntactically different – and would therefore be associated with different cleft types in some works – they are associated with the same information pattern and perform exactly the same discourse function. In both cases, the cleft constituent coincides with given information, while the rest is new. In addition, both constructions highlight the anaphoric and topical, post-copular element (the referent lui ‘he’), while the rest (i.e., the cleft clause) reformulates and specifies the content conveyed in the Comment part of the first proposition of the reported speech (must explain what truly happened in 2006): (46)
“Guido Rossi deve spiegare cosa davvero accadde nel 2006. È lui il primo tra tutti che deve pubblicamente spiegare che cosa è realmente accaduto allora, assumendosi le proprie responsabilità. È lui – aveva concluso Diego Della Valle – che ha il dovere di ricostruire i fatti e darne spiegazione pubblica a tutti quelli che vogliono conoscere la verita”. (ICOCP, leggo.it)
24
Anna-Maria De Cesare
‘“Guido Rossi must explain what truly happened in 2006. He’s the one that must publicly explain what really happened then […]. He’s […] the one that must reconstruct the facts and explain them publicly to all the people who want to know the truth.”’ È lui il primo tra tutti che deve pubblicamente spiegare Is him the first one among all that must publicly explain che cosa è realmente accaduto allora what is really happened then ‘He’s the one that must publicly explain what really happened then’ È lui […] che ha il dovere di ricostruire i fatti Is he […] that must reconstruct the facts ‘He’s […] the one that must reconstruct the facts’
2.2 Distinguishing Cleft constructions on the basis of the cleft clause As we pointed out earlier, two other formal criteria related to the cleft clause are used in the classification of clefts: the nature of the cleft clause introducer and the position of the cleft constituent in regards to the rest of the structure.¹⁷
2.2.1 Nature of the cleft clause introducer and position of the cleft constituent On the basis of examples (1) to (3), recalled below for the sake of clarity, we could describe English clefts in the following way: the cleft clause of English Cleft sentences can be opened either by a generic / opaque form (that) or by a null form (ø); by contrast, the cleft clause of Pseudo-cleft sentences, including the reverse type, is opened by a wh-word (E. what, who, which, where, why, how, when; see Collins 1991: 27): 17 A debated issue is whether we can have clefts lacking a subordinate clause (cf. Mikkelsen 2007). In some cases, as in (a) below, where we have two consecutive clefts, it is quite clear that the structure in bold is indeed a so-called “truncated cleft”; in other cases, it is not so easy to decide. (a) One of the major changes seen by FareShare and organisations like it is in the type of people they are now feeding. Where once it was single homeless and the chronically destitute now it’s increasingly families and working people who have fallen on hard times. (ICOCP, guardian.co.uk)
Cleft constructions in a contrastive perspective
25
(47)
It is champagne (that) I like.
Cleft sentence / it-cleft
(48)
What I like is champagne.
Pseudo-cleft sentence / wh-cleft
(49)
Champagne is what I like.
Reverse pseudo-cleft sentence / Reverse wh-cleft
In Italian and French, too, there seems to be a close connection between the form that introduces the cleft clause and the cleft type: (50)
È lo champagne che mi piace. Is the champagne that I like ‘It is champagne that I like.’
(51)
Quello che mi piace è lo champagne. What I like is the champagne ‘What I like is champagne.’
(52)
C’est le champagne que j’aime. It is the champagne that I like ‘It is champagne that I like.’
(53)
Ce que j’aime c’est le champagne. What I like it is the champagne.
Cleft sentence
Pseudo-cleft sentence
Cleft sentence
Pseudo-cleft sentence
However, the idea that there is a tight connection between the form that introduces the cleft clause and the cleft type does not always hold across all the languages considered and is therefore not possible to rely on in contrastive studies. It is known, for instance, that in some languages there is no difference between the form that opens the cleft clause of what is nevertheless sometimes assumed to be three different types of clefts. As can be observed on the basis of the following examples (these are alternative cleft versions conveying the basic meaning John lost his keys), this is the case in Spanish (note that the labels referring to the three cleft types that are given under parenthesis are from Pinedo 2000); thus, in Spanish, the nature of a cleft type depends solely on the position of the cleft constituent in regards to the rest of the structure (cf. Van den Steen 2005: 278):
26
(54)
Anna-Maria De Cesare
Fueron las llaves lo que perdió Juan. Were.3pl the keys what lost.3sg John ‘It was the keys that John lost.’ (it-cleft; Pinedo 2000: 130)
(55)
Lo que perdió Juan fueron las llaves. What lost.3sg John were.3pl the keys ‘What John lost was the keys.’ (Canonical pseudo-cleft; Pinedo 2000: 131)
(56)
Eso fue lo que perdió Juan. This was.3sg what lost.3sg John ‘That’s what John lost’ (Inverted pseudo-cleft; Pinedo 2000: 131; small capitals are mine)
In contrast with English and other Romance languages, in the variety of Spanish used in Spain, clefts cannot be introduced by the generic / opaque que (cf. Dufter 2010 and Metzeltin 2010: 115):¹⁸ (57) * Fueron las llaves que perdió Were.3pl the keys that lost.3sg (58) * Que perdió Juan That lost.3sg John
Juan. John
fueron las llaves. were.3pl the keys
(59) * Eso fue que perdió Juan. This was.3sg that lost.3sg John As reported by Pinedo (2000), the lack of grammatical marks to distinguish Cleft sentences (54) from Pseudo-cleft sentences (55) and the fact that the form that opens a Spanish cleft clause (lo que) is closer to English Pseudo-clefts than to English Cleft sentences has led to different accounts of the Spanish cleft system. Some scholars consider that Spanish only has Pseudo-clefts. Other scholars, such as Helfrich (2003: 439) and Metzeltin (2010), distinguish only two main groups of clefts: Cleft sentences (Sp. hendidas) and Pseudo-cleft sentences (Sp. seudohendidas). For Helfrich (2003), the class of Cleft sentences includes examples such as (54) and (56), while Pseudo-clefts coincide with examples such as (55). It 18 Spanish Clefts found in varieties used outside of Spain behave differently, as they can be introduced by the complementizer que (on this issue, see for instance Sedano 1990).
Cleft constructions in a contrastive perspective
27
is important to note that, in this account, the class of Cleft sentences is broader than in the other languages we consider in the present study. In Helfrich 2003, two types of Cleft sentences are distinguished on the basis of the position of the copula: structures such as (56) are labeled clefts with mid-copula (see also Antonia es quien gana, lit. ‘Antonia is who wins’, ‘Antonia is the one who wins’); clefts such as (54) are called clefts with initial copula (cf. Es Antonia quien gana ‘Is [= it is] Antonia who wins’). As can be observed on the basis of the following examples, in German too the picture is quite different from English: (60)
Es ist Elisabeth, der du hättest schreiben sollen. It is Elisabeth REL you have.pst.2sg write should ‘It was Elisabeth you should have written to.’ (Cleft sentence; Engel 1988: 298)
(61)
Was ihn störte, war ihre Gleichgültigkeit. What him bothered was her indifference ‘What disturbed him was her indifference.’ (Pseudo-cleft sentence; Erdmann 1990a: 69)
(62)
geld ist das, was sie wollen. Money is what they want.prs.3pl ‘Money is what they want.’ (Reverse pseudo-cleft; Erdmann 1990a: 69; small capitals are mine)
While the cleft clause of German Pseudo-clefts is opened by a free or complex relative (both involving a wh-form: was ‘what’ in [61] and das, was ‘what’ in [62]), the cleft clause of Cleft sentences is generally opened by a so-called d-pronoun (der, die, das etc.; on this label, cf. for instance Grewendorf and Poletto 1989: 115), which is similar to both a relative and a demonstrative pronoun. Moreover, in contrast with English as well as with the three Romance languages with which we are dealing in this paper, in German a cleft constituent is always in the nominative case; it’s in fact the d-pronoun that bears case-marking (i.e., is declined or preceded by a preposition): (63)
Es war der arme, blutige Christus, dem man nachfolgen sollte. It was the poor, bloody Christ.nom REL.dat one follow should ‘It was the poor, bloody Christ that one had to follow.’ (Kiese 1993: 20)
28
Anna-Maria De Cesare
In German, when a prepositional object or an adjunct based on a prepositional phrase is clefted, the preposition is always placed before the d-pronoun. Consider example (64), as opposed to its English translation: (64)
Zunächst ist es die PUPpenbühne, auf der der sensible Knabe At first is it the puppet stage, on which the sensitive boy seine Träume austobt his dreams lets free course to ‘At first it is on the puppet stage that the sensitive boy lets free course to his dreams.’ (Altmann 2009: 27)
This picture is further complicated by the fact that in German the cleft clause of a Cleft sentence can also start with a different paradigm of forms. Besides the class of d-Pronouns, the cleft clause of Cleft sentences can also be opened by the forms welches, welche, welcher ‘which’, ‘who’. According to Kiese (1993), clefts such as (65) are more formal and thus belong to higher registers of the language. (65)
Es war Jolanthe, welche die Frage gestellt hatte. (Kiese 1993: 52) ‘It was Jolanthe, who asked the question.’
In German, the generic relative form dass corresponding to English that is used only in special cases, such as when the cleft constituent coincides with a temporal expression:¹⁹ (66)
Es ist Elisabeth, der / *dass du hättest schreiben sollen. It is Elisabeth REL / *that you have.pst.2sg write should ‘It is Elisabeth you should have written to.’
(67)
Es war am 13. Dezember 1918, daß er zuerst auf französischen Boden It was on December 13, 1918 that he first on French soil landete landed.3sg ‘It was on December 13, 1918 that he first landed on French soil.’ (Kiese 1993: 24)
19 Clefts on adverbials are not considered to be natural by all native speakers; the following example provided by Motsch (1970: 90) is considered ungrammatical: *es war gestern, dass wir uns trafen ‘it’s yesterday that we met’.
Cleft constructions in a contrastive perspective
29
2.2.2 The status of cleft clauses opened by a wh-form Another taxonomic problem occurs with Italian clefts, such as (68), in which the cleft clause is headed by a wh-form (It. chi ‘who’) and the cleft constituent (il conte ‘the count’) is found after the copula. This form of cleft can also be found in other languages. Below, we provide examples from English, French, Spanish and German: (68)
È il conte chi ha ucciso il maggiordomo Is the count who killed.3sg the butler ‘It’s the count who killed the butler’ (Salvi 1991: 177)
(69)
It was they who fought back during a violent police raid […] (cf. Prince 1978: 898; Biber et al. 1999: 959)
(70)
C’est elle qui viendra. It’s she who will come.3sg ‘She’s the one who will come’
(71)
Ha sido Mercedes quien ha entregado un ramo de flores. Has been Mercedes who delivered a bouquet of flowers ‘It’s Mercedes who delivered the flowers.’ (Metzeltin 2010: 115)
(72)
Es war in dem weniger betretenen Teile des Gartens, wo It was in the lesser known part of the garden where die Rosenschule war. the rose school was ‘It’s in the lesser known part of the garden that the rose school was.’ (Leirbukt 1969: 6)
Differently from (69)-(72), Italian clefts such as (68) have quite a fuzzy status in the literature. While the syntactic structures given in (69) to (72) are almost unanimously considered to be instances of Cleft sentences in English (cf., among many others, Collins 1991; Lambrecht 2001), in French (Lambrecht 2001), in Spanish (Pinedo 2000; Helfrich 2003; Van den Steen 2005; Metzeltin 2010) and in German (Birkner 2008), Italian clefts such as (68) – albeit very similar to the ones given in (69) to (72) – are described as either Cleft sentences (cf. Metzeltin 2010: 111), a type of Pseudo-cleft sentences (for Salvi 1991 these structures are Pseudo-clefts involv-
30
Anna-Maria De Cesare
ing right dislocation of the subject) or Reverse pseudo-cleft sentences (cf. Roggia 2009, who also provides the authentic example è in casa dove lavoro meglio ‘it’s at home that I work best’, p. 19).²⁰ The same uncertain status is associated with the following forms of Italian clefts, based on a complex relative, i.e., on a relative headed by a pronoun (ciò che ‘what’, quello con cui ‘the one with whom’ etc.): (73)
è proprio l’alta inflazione ciò che si teme is precisely the high inflation what one is afraid of ‘It’s precisely high inflation what we are afraid of’ (Berretta 2002: 21)
(74)
È Giovanni quello con cui voglio parlare Is Giovanni the one with whom want.prs.1sg talk ‘It’s Giovanni the one I want to talk to’ (Berretta 2002: 21)
Again, examples such as (73) and (74) are considered to be instances of Cleft sentences (Metzeltin 2010), Pseudo-cleft sentences (Salvi 1991: 178) or Reverse Pseudo-cleft sentences (Berretta 2002: 21; Roggia 2009: 19).²¹
2.3 Classifying clefts with no English counterparts As pointed out in Section 2.2, the traditional taxonomy of Cleft constructions, based on the formal properties of English clefts, is particularly problematic in the cases in which a language has a cleft form in its repertoire that has no (close) structural equivalent in the English language. We will consider three other cases here: the first from Italian, the second from German, the third from Spanish and French.
20 According to Roggia (2009), cleft clauses of (what he considers) reverse clefts opened by dove ‘where’ are possible, albeit rare. This fact could explain why this type of cleft has not been pointed out in the literature prior to his study. 21 Note that similar structures are already discussed in Bolinger (1972: 111): It was John who did it; It was up here where I put it; It will be pretty soon when you have to do it. Bolinger refers to these constructions as inverted clauses.
Cleft constructions in a contrastive perspective
31
2.3.1 Italian implicit Cleft sentences A first type of cleft that is not found in English but is typical of Italian is the socalled “implicit cleft sentence”, illustrated below in (75a) and (76).²² This type of cleft is special as the cleft clause is based on an infinitive verb form (see uccidere ‘kill’ in [75a], as opposed to the tensed verb form rovina ‘ruins’ in [75b] where we have an “explicit cleft sentence”) preceded by a ‘to’ (instead of che ‘that’). The examples provided in (75a) and (76) show that this type of cleft has two linguistic manifestations. The implicit cleft clause can occur at the end or at the beginning of the construction, i.e. after or before the copula and the cleft constituent: (75)
a.
Ho responsabilità politiche, ma non fui io Have.prs.1sg responsibilities political, but neg be.pst.1sg I a uccidere (ICOCP, ANSA) to kill ‘I have political responsibilities, but I was not the killer’
b. “È l'opposizione che ci rovina”. “Is the opposition that us ruins” ‘‘“It’s the opposition that ruins us”.’ (76)
(ICOCP, Adncronos)
A segnare gli aumenti maggiori […] sono pizza, tramezzino, To mark the increases highest […] are pizza, tramezzino, cappuccino e cornetto. (ICOCP, repubblica.it) cappuccino e cornetto ‘To have become more pricey are pizza, sandwiches, cappuccino and cornetto.’
The first type of implicit cleft (illustrated in [75a]) has been quite unanimously assigned to the class of Cleft sentences (cf. Berretta 2002; Roggia 2009), while the second type (76) is associated with two different cleft classes: in a first account, this form belongs to the class of Pseudo-clefts (Berretta 1994, 1995, Gil 2004 and De Cesare 2005 refer to this form as frase pseudoscissa); more recently, based on the observation that this form of cleft is actually reversed in relation to the structure illustrated in (75a), it has been suggested to call examples such as (76) Reverse cleft sentence (It. frase scissa inversa notably by Roggia 2009).
22 Note that in English we find clefts of the type it’s him speaking, with a non-finite verb form in the cleft clause. As can be observed on the basis of the example provided, we do not, however, have to do with clefts involving an infinitive form in the cleft clause.
32
Anna-Maria De Cesare
2.3.2 German Cleft sentences with initial cleft constituent Another problematic case to describe is the German cleft illustrated in (77), in which we find the cleft constituent in initial position – in line with Reverse pseudo-clefts – and a cleft clause opened by a d-pronoun – in line with German Cleft sentences (prototypical examples of German clefts are provided in [60] to [62] above): (77)
SIE [Marilyn Monroe] war es, die sagte: […] SHE [Marilyn Monroe] was it, REL said.3sg ‘SHE [Marilyn Monroe] was it who said: […]’ (Altmann 2009: 21)
The hybrid form of these clefts has led to two distinct taxonomic proposals: in some studies, clefts such as these are called Reverse pseudo-clefts (cf. Doherty 1999: 292; Birkner 2008: 318, 325, 327), while in other studies they are considered to be Cleft sentences (see Kiese 1993; Altmann 2009). This case is informative because it shows quite clearly that the categorization of clefts generally follows from a single criterion, which is on the one hand the position of the cleft constituent in respect to the rest of the sentence (cf. first account of cleft [77]) and on the other hand the form that introduced the cleft clause (cf. second account of cleft [77]).
2.3.3 French and Spanish Reverse pseudo-cleft sentences Related to the problem discussed above, i.e., to the difficulty of classifying clefts that do not have a direct equivalent in the English language, is also the fact that some English cleft types are claimed not to exist in certain languages. One case in point regards the category of Reverse pseudo-clefts, which Lambrecht (2001) and Miller (2006) claim is absent in the family of both French and Spanish clefts. The examples provided by Lambrecht (2001) to support the fact that this structure is only available in English are given in (78) to (81): (78)
CHAMPAGNE is what I like.
(79) * Le champagne est ce que j’aime. the champagne is what I like ‘Champagne is what I like.’ (Lambrecht 2001: 492; note the lack of small capitals on le champagne)
Cleft constructions in a contrastive perspective
(80)
33
I’m the one who bought this coat. (Lambrecht 2001: 492; the pronoun I is stressed)
(81) * YO fui quien compró este abrigo. I was.3sg who bought.3sg this coat ‘I am the one who bought this coat’ (Lambrecht 2001: 492) In Lambrecht’s (2001) view, both French and Spanish use a Cleft sentence where English would use a Reverse pseudo-cleft. Hence the English Reverse pseudoclefts given in (78) and (80) coincide with the following two structures in French and Spanish, respectively: (82)
C’est le champagne que It’s the champagne that ‘It is champagne I like.’
j’aime. I like (Lambrecht 2001: 492)
(83)
Fui YO quien compró este abrigo. Was.3sg I who bought.3sg this coat ‘I am the one who bought this coat.’ (Lambrecht 2001: 492 on the basis of Smits 1989)
However, in our view these claims are in need of a revision (cf. also Wehr in press). As we have seen in Section 2.2.1, there are clear instances of Reverse pseudo-clefts in Spanish as well (cf. Pinedo 2000; Van den Steen 2005; Wehr in press). Consider the following cases, the first with the pronoun yo, excluded by Lambrecht 2001, the second with the demonstrative eso ‘this’ (the latter example, already seen in [56], is from Pinedo 2000): (84)
YO soy el que se va. (Wehr in press) I am the one going ‘I’m the one who’s leaving’
(85)
Eso fue lo que perdió Juan This was.3sg what lost.3sg Juan ‘That’s what Juan lost’ (Pinedo 2000: 131; small capitals are mine)
34
Anna-Maria De Cesare
A similar claim can be made for French. As we can see on the basis of the example provided in (86), there are attested occurrences of what looks like Reverse pseudoclefts in this language as well. What is true is certainly that this form of cleft is rather rare in French, while it is common in English. It should also be noted that the noun phrase le sport is not highlighted here because we are dealing with a written text. However, it is only in the case in which the NP le sport is prosodically highlighted that we are dealing with a Reverse pseudo-cleft (as in le sport est ce que je connais le mieux ‘sports in what I know best’): (86)
Le sport est ce que je connais le mieux. (humanite.fr) The sport is what I know.prs.1sg best ‘Sports is what I know best’
As can be observed in examples (84) to (86), much in line with German Reverse pseudo-cleft sentences (cf. example [87], provided earlier in [62]), in Spanish and in French the cleft clause is opened by a complex pronoun (Sp. lo que, Fr. ce que, G. das, was, all translated by ‘what’) rather than by a free relative pronoun (what), as is commonly found in English: (87)
Geld ist das, was sie wollen. Money is what they want.prs.3pl ‘Money is what they want.’ (Reverse pseudo-cleft; Erdmann 1990a: 69; small capitals are mine)
That said, it seems that there is a difference between Spanish and German, on the one hand, and French, on the other. In French, the syntactic structure given in (86) is more readily interpreted as having a non-focal initial constituent. Hence, the constituent le sport is not associated with a special prosodic contour, i.e. with contrastive accent, and the whole structure cannot be considered to be a cleft. Moreover, when there is the need of contrasting a referent, it seems indeed more likely that French would use a regular Cleft sentence.
3 Cleft constructions: A new taxonomy In Section 2, we have shown that the traditional taxonomy of Cleft constructions is not always easy to apply cross-linguistically and has given rise to different interpretations of the same structure. We believe that the main reason for this difficulty is that clefts have been described from a somewhat anglocentric point
Cleft constructions in a contrastive perspective
35
of view (a claim also made in Berretta 1994: 87 and recently by Valentini 2012: 79, n. 10), on the basis of the somewhat misguided belief that the theoretical categories developed on the basis of one language (in this case English) are valid for and thus transferable to other languages, without major descriptive adjustments. In fact, the accounts of clefts are often offered without taking into account the structural specificities of the languages analyzed, i.e., by overlooking the fact that the main cleft components do not have the same grammatical status in different languages.²³ In light of the classification problems highlighted in the previous paragraph, we believe that a more economical way of classifying the different cleft types available cross-linguistically is to focus first and foremost on the placement of the cleft constituent in the syntactic structure. Additionally, we would also like to claim that it is better to avoid using the two sets of labels commonly employed in the literature on clefts. As we have seen, besides the clear English bias of one of the sets, the traditional labels Cleft sentences (it-clefts), Pseudo-cleft sentences (wh-clefts) and Reverse pseudo-cleft sentences (reverse wh-clefts) are not used to describe the same syntactic structures cross-linguistically and are therefore not optimal to rely on as a starting point in (large-scale) contrastive studies.
3.1 A new taxonomy based on the position of the cleft constituent The criterion that in our view allows the best comparison of the Cleft constructions available cross-linguistically, i.e., that allows us to compare these syntac23 Up to this point, we have not said much on the divergent grammatical interpretations of the sub-parts of the clefts, but have concentrated instead on the interpretation of the structure as a whole. There are, however, numerous discrepancies (both intra- and cross-linguistic) in the description of the grammatical status of the cleft parts, in particular in the description of the cleft clause and the cleft clause introducer. These discrepancies are due on the one hand to different grammatical traditions and on the other hand to the fuzzy status of the cleft components. It is of course beyond the scope of this paper to present a detailed account of the different descriptions provided in the literature for the main cleft components. Suffice it to mention here some proposals related to the interpretation of the cleft clause introducer. The form that in It was mainly his cynical attitude that made him unpopular is described as a relative particle in Smits (1989: 299), a pronoun in Pinedo (2000), while it is considered to be a complementizer in generative accounts (cf. Grewendorf and Poletto 1989: 113). The form where (and how) in It was in the kitchen where John built a ship is called relative adverb in Smits (1989: 300–301), while the corresponding Spanish form donde ‘where’ in Fue en Madrid donde nací ‘It’s in Madrid that he is born’ is considered to be a preposition (cf. Pinedo 2000). Note also that for Smits (1989: 299–300), who in it was I who bought this coat is interpreted as a relative pronoun.
36
Anna-Maria De Cesare
tic structures more easily, is the position that the cleft constituent occupies with respect to the rest of the structure. Such a proposal is of course not new. Following this very criterion, a distinction is sometimes made in the literature between clefts with initial (88) and with final (89) cleft constituent (cf., among others, Berretta 2002 and Roggia 2009 for Italian, as well as Metzeltin 2010: 109 for the main Romance languages): (88)
a.
Questo è quello che This is what ‘That’s what I think’
penso think.prs.1sg (Berretta 2002: 17; small capitals are mine)
b. È questo quello che penso Is this what think.prs.1sg ‘That’s what I think’ (89)
Quello che penso è questo What think.prs.1sg is this ‘What I think is this’
This distinction is based on the position that the cleft constituent (in our examples, the demonstrative pronoun questo ‘this’) occupies with respect to the cleft clause (quello che penso ‘what I think’). In German, this proposal even gave birth to a new set of labels: in Engel (2004: 181), instead of the more common Spaltsatz ‘Cleft sentence’ and Sperrsatz ‘Pseudo-cleft sentence’, clefts are referred to as Linksspaltung ‘left clefting’ and Rechtsspaltung ‘right clefting’, according to the position that the cleft constituent occupies in the structure. Consider the following examples: (90)
Es ist die Michaela, die mir in dieser Geschichte am meisten Leid tut. ‘It is Michaela the person that I feel most sorry for in this story.’ (Linksspaltung ‘left clefting’)
(91)
Wer / Die mir in dieser Geschichte am meisten Leid tut, ist die Michaela. ‘The one who I feel most sorry for in this story is Michaela.’ (Rechtsspaltung ‘right clefting’)
Cleft constructions in a contrastive perspective
37
3.1.1 Cleft constructions with initial, medial and final cleft constituent In line with the new classification system proposed in Calude (2009: 143, 195) for English, as well as with the classification system sometimes used for Spanish (cf. Van den Steen 2005: 278), we propose to distinguish three types of Cleft constructions: with initial, medial and with final cleft constituent. More precisely, the difference between the three types of Cleft constructions is the following: in the first group of Cleft constructions, the cleft constituent (CC in the figure below) opens the construction and precedes both the copula (COP) and the cleft clause (CCL); in the second group of Cleft constructions, the cleft constituent is located between the copula (optionally preceded by an expletive subject in non pro-drop languages, cf. E. it, G. es ‘it’ and Fr. ce ‘it’) and the cleft clause; in the last group, the cleft constituent follows both the cleft clause and the copula (found in reversed order in respect to the other two cleft types) and thus closes the whole structure. The new classification system is outlined in Figure 2:
A. initial cleft constituent CC COP CCL
B. medial cleft constituent COP CC CCL
C. final cleft constituent CCL COP CC
Figure 2: Cleft constructions: A classification system based on the position of the cleft constituent
This classification proposal differs from the account mentioned above in that we consider the clefts given in (88) as two distinct types. In the new classification system, a difference is made between clefts with initial (cf. [88a]) and with medial cleft constituent (cf. [88b]). There are also several important differences between the classification system proposed in Figure 2 and the classic taxonomy of clefts summarized in Figure 1. The most important differences are as follows: (i) the labels referring to the three types of clefts mentioned in Figure 2 are much more transparent and therefore straightforward to apply in cross-linguistic descriptions; (ii) in contrast with what could be suggested in the traditional taxonomy of clefts, in the new
38
Anna-Maria De Cesare
classification proposal there is no hierarchy between the three types of clefts: as can be observed on the basis of their position in Figure 2, the three cleft types are independent from each other; from this it follows that the three types of clefts are considered to be equally important; (iii) in turn, from the previous point it follows that in the new classification system no correlation is suggested between a given cleft type and its derivation pattern and that there are no expectations concerning the reversibility of one cleft type into another; (iv) the classification proposed in Figure 2 is more straightforward in describing the problematic structures discussed in Section 2. We will return to this point below.
3.1.2 Cleft constructions with initial, medial and final cleft constituent: Crosslinguistic evidence
At this point, we would like to make our classification proposal more concrete by providing examples from the five languages in which we are interested. Table 1, which also recalls the abstract formal make-up of the three main types of clefts (cf. line 3), illustrates the paradigm of Cleft constructions in English, German, Italian, French and Spanish: Table 1: Cleft constructions in five European languages cleft constructions with … initial cleft constituent
medial cleft constituent
final cleft constituent
CC-COP-CCL
COP-CC-CCL
CCL-COP-CC
that’s what I think
it’s Stella that helped me
what I think is this
G das ist (es), was ich glaube
es ist Sella, die mir half
was ich glaube, ist das
I
questo è quello che penso
è Stella che mi ha aiutata
quello che penso è questo
F
ça, c’est ce que je pense²⁴
c’est Stella qui m’a aidée
ce que je pense, c’est ça
S
eso es lo que pienso
es Stella la que me ayudó
lo que pienso es eso
A
B
C
E
24 Based on its structural resemblance with Type C clefts (i.e. with the structure ce que je pense, c’est ça ‘what I think is this’), we decided to include this example here. However, we have already observed that it is difficult to interpret this structure as having a contrastive focus without clear prosodic cues (the same holds for other languages).
Cleft constructions in a contrastive perspective
39
As Table 1 demonstrates, there is a certain degree of structural overlap across the five languages. This structural overlap also concerns the paradigm of forms that belong to the three types of clefts. First, it concerns the clefts constructed with a cleft clause opened by a free (or wh-) pronoun (in particular Type C clefts). Consider Table 2, which provides the list of free pronouns (some of which could be considered adverbs) that can open the cleft clause of all three types of clefts (note that these forms basically correspond to interrogative pronouns and, in some cases, also to relative pronouns).²⁵ Table 2: Free relative pronouns in five European languages English
German
Italian
French
Spanish
who
wer
chi
qui
quien
where
wo
dove
où
donde
wieviel
quanto
combien
cuanto
what
was
when
wann
quando
quand
cuando
how
wie
come
comment
como
why
warum
perché
pourquoi
por que
Leaving the realm of prototypical clefts, there is a fair cross-linguistic formal overlap also in the case of clefts constructed with a cleft clause opened by a generic noun. Table 3 provides the list of generic heads that can be used in English, German, Italian, French and Spanish to open the cleft clause of Type C, but also of Type A and B clefts. Following Collins (1991), we propose to consider only the lexical nouns corresponding to the list of free relatives given in Table 2 and recalled in the first column to the left of Table 3 (cf. Collins 1991: 29–31). In Table 3, we adopt this view for the other languages.
25 The empty cells of this table are to be interpreted as a gap in the system of simple wh-forms. It should also be noted that this table does not contain all the possible wh-forms: as well as E. who, we find whom, as well as Sp. quien, the form quienes etc. A slightly more detailed list of whforms is provided in De Cesare et al. (in this volume). Moreover, although semantically related, in Table 2 we have not included the English form all (cf. Calude 2009: 143: a book is all I want; all I want is a book).
40
Anna-Maria De Cesare
Table 3: List of generic lexical heads opening cleft clauses in five European languages English
German
Italian
French
Spanish
who
one / person
Person
persona
personne
persona
where
place
Ort
luogo posto
endroit lieu
sitio lugar
what
thing
Ding Sache
cosa
chose
cosa
when
time moment
Moment
momento moment periodo, volta
momento
how
way
Art Weise
modo
manière façon
manera modo forma
why
reason explanation motive
Grund
ragione motivo
raison motif
razón motivo
Note that we have extended somewhat the list of generic nouns that can open a cleft clause in particular of Type C clefts. In Collins’ (1991) view, the generic noun corresponding to the wh-form who is the one; we believe that other nouns, such as person should also be taken into account. Consider the following authentic example in (92) which corresponds to the forms commonly employed in Italian and French Type C clefts (examples [93] and [94] repeat examples already seen above):²⁶ (92)
I think the person I like best is the step granny.
(93)
la persona che ha aperto la finestra è tuo padre ‘The person who opened the window is your father’
(kiasuparents.com)
26 A detailed discussion of the relationship between free relatives and generic lexical nouns is already provided in Bolinger (1972). In his view (cf. in particular, p. 105), what he calls analytical compounds (the one, which expresses an entity or person; the place, which expresses location; the way, which expresses manner; the time, which expresses time; and the reason, which expresses rationale) are in free variation with synthetic compounds, which are similar to interrogatives (what, who, where, when, why; note that he does not provide an equivalent synthetic compound for manner; in his view, thus, how is only an interrogative pronoun). Interestingly, the list of analytical compounds is the same as the one provided by Collins (1991).
Cleft constructions in a contrastive perspective
(94)
la personne que Nixon a choisie c’est Agnew ‘the person that Nixon chose was Agnew’
41
(Roubaud 2000: 26)
The same is true for other wh-pronouns: in our view, why corresponds not only to the reason but also, in addition to Collins’ (1991) proposal, to the generic lexical nouns the explanation and the motive; when corresponds to the time but also the moment etc. As we highlighted in Section 2 and present in more detail in De Cesare et al. (in this volume), the five languages on which we focus do not always construct the three types of clefts in the same way or use the same structure in the cases where parallel options are available cross-linguistically. There are differences even between languages belonging to the same genetic family (cf. Harries-Delisle 1978; Smits 1989). For instance, as already mentioned, there are important cross-linguistic differences regarding the cleft clause introducers in English and German: while English Type B clefts can involve a “covert” introducer (it’s Stella ø I saw), German Type B clefts cannot do so (*es ist Stella, ø ich sah) and rely instead on a form of cleft clause introducer that has no structural equivalent in English, i.e., a so-called d-Pronoun (es ist Sella, die ich sah). Another cross-linguistic difference regards Type C clefts in the two Germanic as opposed to the two Romance languages we are dealing with (cf. De Cesare et al. in this volume). While English and German strongly favor syntactic structures opened by a simple, free pronoun (mainly E. what, G. was ‘what’), the Romance languages use syntactic structures opened by a complex pronoun (It. quello che, Fr. ce que, Sp. lo que, all three corresponding to ‘what’), as they lack a corresponding generic, simple free pronoun. While the option of using a complex pronoun is not easily available in English, it is possible in German (cf. das, was ‘what’).
3.1.3 Cleft constructions with initial, medial and final cleft constituent: Paradigm of forms At this point it should be highlighted that each paradigm of clefts – i.e., clefts with initial, medial and final cleft constituent – can have different realizations in one and the same language. As displayed in Table 4, the Italian language is particularly adapted to illustrate the wealth of forms that could possibly be found in this language for each class of clefts. Each class of clefts, i.e., each of the three Cleft construction types (A-B-C), are based on the same structure: they are constructed with a cleft clause headed by a generic (relative) pronoun (It. che ‘that’),
42
Anna-Maria De Cesare
a free relative pronoun (It. chi ‘who’), a complex pronoun (quello che ‘what’) or with an implicit verb form opened by a ‘to’:²⁷ Table 4: Subtypes of Cleft constructions with initial, medial and final cleft constituent in Italian cleft constructions Type A (with initial cleft constituent) A-1 Stella è che ama la linguistica
A-3 Stella è quella che ama la linguistica
A-2 Stella è chi ama la linguistica
A-4 Stella è ad amare la linguistica
Type B (with medial cleft constituent) B-1 È Stella che ama la linguistica
B-3 È Stella quella che ama la linguistica
B-2 È Stella chi ama la linguistica
B-4 È Stella ad amare la linguistica
Type C (with final cleft constituent) C-1 Che ama la linguistica è Stella
C-3 Quella che ama la linguistica è Stella
C-2 Chi ama la linguistica è Stella
C-4 Ad amare la linguistica è Stella
Several observations ought to be made here. The first one is, again, that the three cleft types identified in Table 4 correspond only roughly to the three cleft types traditionally identified (cf. Figure 1) and should therefore not be equated in block. In other words, Cleft constructions of Types A, B and C do not necessarily correspond to what is traditionally subsumed in the classes of Reverse pseudo-cleft sentences (Reverse wh-clefts), Cleft sentences (or it-clefts) and Pseudo-cleft sentences (or wh-clefts). Thus, in contrast to what is traditionally considered to be a Pseudo-cleft in Italian, Cleft constructions of Type C do not only include syntactic structures opened by a free relative (cf. C-2) and by a relative headed by a complex pronoun (C-3), but also by other forms (C-1 and C-4).²⁸
27 In order to ease the reading of the data contained in Table 4, we do not provide the glosses and translations of the examples. The monoclausal counterpart of all these clefts is the following: Stella ama la linguistica ‘Stella likes linguistics’. 28 As mentioned earlier, there have been some proposals in the literature on Italian to extend the domain of application of both the classes of Cleft sentences and Pseudo-cleft sentences. Such a proposal has been made for instance by Monica Berretta (cf. Berretta 1994, 1995, 2002), who suggests that all the Clefts Constructions with medial cleft constituent should be called Cleft sentences and all the clefts with final cleft constituent should be labeled Pseudo-cleft sentences (“if the focus introduced by essere [‘to be’] precedes the pseudo-relative we will talk about clefts, if the order is reversed of Pseudo-cleft”, Berretta 2002: 16; the translation is ours). A similar proposal is found in Gil (2004) for the class of Pseudo-clefts: in his view, this category can be
Cleft constructions in a contrastive perspective
43
It should of course also be observed that this new taxonomic proposal is straightforward to apply and can easily solve the problematic cases highlighted in Section 2. In this new taxonomy, clefts like (68) and (73), repeated below as (95) and (96), are both instances of Type B clefts, as the cleft constituent is located between the copula and the cleft clause. Although in the literature these two types of clefts have been assigned to the class of Cleft sentences, to the class of Pseudo-cleft sentences as well as to the class of Reverse pseudo-cleft sentences, in our proposal these clefts are quite clearly instances of Type B-2 and Type B-3, respectively: (95)
È il conte chi ha ucciso il maggiordomo Is the count who killed.3sg the butler ‘It’s the count who killed the butler’ (Salvi 1991: 177)
(96)
è proprio l’alta inflazione ciò che si teme is precisely the high inflation what one is afraid of ‘It’s precisely high inflation what we are afraid of’ (Berretta 2002: 21)
Another observation is that all the examples provided in Table 4 have to be interpreted as potential realizations of A-B-C type clefts; this table does not mean to convey the idea that all these clefts are attested and that they have the same distribution and frequency of use. The latest results of corpus-based researches (cf. also De Cesare et al. in this volume) show very clearly that these cleft types and subtypes are strongly register and genre-specific, i.e., that their distribution varies according to textual as well as socio-linguistic parameters. It is well known, for instance, that there are major differences in the distribution of clefts in spoken vs. written varieties of contemporary Italian, specifically in formal vs. informal registers of the language. Let us illustrate this point on the basis of Type C-clefts. Implicit clefts of Type C-4 are typical of written / formal language varieties (Roggia 2009; also see De Cesare 2014), while Type C-1 clefts are used chiefly in
extended so as to cover not only the constructions opened by a free relative (chi ‘who’) and by a relative headed by a pronoun (quello che ‘the one that’), but also the constructions opened by an implicit cleft clause (our Type C-4). In this paper, we make a different suggestion: the labels Cleft sentences and Pseudo-cleft sentences should be maintained only for the traditional and prototypical forms of Cleft sentences (i.e. B-1 and B-4) and Pseudo-cleft sentences (i.e. C-2 and C-3). In the other cases, these labels should be avoided, in particular for languages that differ significantly from English.
44
Anna-Maria De Cesare
colloquial / substandard varieties of Italian (cf. Berretta 1996: 120). The colloquial nature of Type C-1 clefts is evident in the example provided below, taken from an informal conversation, in which we can observe several oral features (cf. the constructio ad sensum, i.e., the notional agreement between the plural verb form parlano and the singular noun la gente): (97)
A me che fa paura è la gente che To me that makes me scared is the people.SG that parlano piano speak.PL low ‘What scares me is people that speak in a low tone of voice’ (Berretta 1996: 120)
4 Concluding remarks In light of important taxonomic shortcomings resulting from the application of the traditional taxonomy of clefts to five European languages, this paper proposes a new classification system of Cleft constructions based on a single factor, which identifies the linear position of the cleft constituent within the syntactic structure. In contrast to the classic tripartite taxonomy of clefts, which is primarily rooted on the formal properties of both the cleft construction introducer as well as the cleft clause introducer, the new tripartite classification of clefts is based on the location of the cleft constituent in relation to both the cleft constituent and the cleft clause. The main advantage of the new classification system of clefts we proposed is that it is straightforward to apply to data from different languages and allows the classificatory problems listed in the first part of the paper to be easily solved. We believe that this new classification system of clefts offers a very good basis for contrastive studies which aim to describe the frequency, forms and functions of clefts from a cross-linguistic perspective. We also believe that this classification is especially useful when several languages are taken into account at once and when the comparison involves languages that differ structurally, not only from English but also from each other. The papers included in the first part of the volume partly rest on this new classification system to describe the cross-linguistic similarities and differences in the frequency, forms and functions of clefts. For an overview of the distribution of Type B and C clefts in a corpus of online news articles, see in particular De Cesare et al. (in this volume).
Cleft constructions in a contrastive perspective
45
There are of course numerous open questions to address in future research and several pieces of data that ought to be further developed. For instance, while the literature has mainly shown that different types of clefts (in particular the traditionally labeled it-cleft and wh-cleft) encode different syntactic and semantic properties,²⁹ we need to demonstrate more thoroughly that there are significant differences in the ways syntactic and semantic features are encoded within the same paradigm of Cleft construction types, i.e., within Type A, Type B and Type C clefts. Moreover, a more fine-grained account of the functional differences and similarities between Type A, Type B and Type C clefts must be provided. Specifically, while the differences between Type A / B clefts, on the one hand, and Type C clefts, on the other hand, have already been described in numerous studies – and so we thus moved away long ago from the belief that Cleft sentences and Pseudo-cleft sentences are functionally equivalent; crucial studies in this respect are certainly Prince (1978) and Sornicola (1988) – the differences between Type A and Type B clefts are not so evident. In our view, the main evidence showing that there is a difference between Type A and B clefts is the fact that many languages (cf. English, Spanish, German) have both types of clefts in their repertoire and that in these languages, the cleft constituent of these two types of clefts does not have the same formal and pragmatic properties (cf. De Cesare 2011 on German). This, of course, presupposes that we have solved another difficult issue, namely the distinction between Type A clefts and canonical predicative sentences.³⁰
References Agar Marco, Rocío. This vol. Pseudo-cleft Sentences. Italian-Spanish in contrast. Akmajian, Adrian. 1979. Aspects of the Grammar of Focus in English. New York: Garland Publishing. Altmann, Hans. 2009. Cleft- und Pseudocleft-Sätze (Spalt- und Sperrsätze) im Deutschen. In Rita Brdar-Szabó, Elisabeth Knipf-Komlósi & Attila Péteri (eds.), An der Grenze zwischen Grammatik und Pragmatik, 13–34. Bern: Peter Lang. Atayan, Vahram & Ursula Wienen. This vol. Inferential cleft constructions in translation. French c’est que in political texts. Ball, Catherine N. 1977. Th-clefts. Pennsylvania Review of Linguistics 2. 57–69. Baranzini, Laura. This vol. Pseudo-cleft Sentences. Italian-French in contrast.
29 Cf. Prince (1978: 884–885) and later, for instance, Biber et al. (1999: 160). 30 This problem is particularly acute in written texts, as prosodic cues are missing and the intonational contour of syntactic structures ought to be reconstructed on the basis of the context; on these issues, cf. Garassino 2014, as well as Wehr in press.
46
Anna-Maria De Cesare
Berretta, Monica. 1994. Ordini marcati dei costituenti di frase in italiano. Vox romanica 53. 79–105. Berretta, Monica. 1995. Ordini marcati dei costituenti maggiori di frase: una rassegna. Linguistica e Filologia 1. 125–170. Berretta, Monica. 1996. Come inseriamo elementi nuovi nel discorso/III: Che mi fa paura è la nebbia. Italiano & Oltre XI. 116–122. Berretta, Monica. 2002. Quello che voglio dire è che: le scisse da strutture topicalizzanti a connettivi testuali. In Gian Luigi Beccaria & Carla Marello (eds.), La parola al testo. Scritti per Bice Mortara Garavelli, 15–31. Torino: Edizioni dell’Orso. Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad & Edward Finegan. 1999. Longman Grammar of Spoken and Written English. London: Longman. Birkner, Karin. 2008. Relativ(satz)konstruktionen im gesprochenen Deutsch. Syntaktische, prosodische, semantische und pragmatische Aspekte. Berlin & New York: Mouton de Gruyter. Bolinger, Dwight. 1972. A look at equations and cleft sentences. In Evelyn Scherabon Firchow, Kaaren Grimstad, Nils Hasselmo & Wayne A. O’Neil (eds.), Studies for Einar Haugen. Presented by friends and colleagues, 96–114. The Hague & Paris: Mouton. Calude, Andreea S. 2009. Cleft Constructions in Spoken English. Saarbrücken: VDM Verlag Dr. Müller. Collins, Peter C. 1991. Cleft and Pseudo-cleft Constructions in English. London: Routledge. De Cesare, Anna-Maria. 2005. La frase pseudoscissa in italiano contemporaneo. Aspetti semantici, pragmatici e testuali. Studi di grammatica italiana XXIV. 293–322. De Cesare, Anna-Maria. 2011. L’ordine dei costituenti in italiano contemporaneo e in prospettiva contrastiva con il tedesco. Tra sintassi, pragmatica e tipologia linguistica. Habilitationsschrift, Università di Basilea, pp. 430. De Cesare, Anna-Maria. 2014. Les phrases clivées en italien contemporain. Quelle caractérisation dans les volumes de référence de l’italien et quelle description à partir de corpus?. In Georgia Veldre-Gerner & Sylvia Thiele (eds.), Sprachen und Normen im Wandel, [Romanische Sprachen und ihre Didaktik], 201–216. De Cesare, Anna-Maria, Davide Garassino, Rocío Agar Marco & Laura Baranzini. This vol. Form and frequency of Italian cleft constructions in a corpus of electronic news. A contrastive perspective with French, Spanish, German and English. Declerck, Renaat. 1984. The pragmatics of it-clefts and wh-clefts. Lingua 64. 251–289. Doherty, Monika. 1999. Clefts in translation between English and German. Target 11. 289–315. Dufter, Andreas. 2010. El que galicado: distribución y descripción gramatical. In Alfonso Zamorano Aguilar & Carsten Sinner (eds.), La ‘excepción’ en la gramática española. Perspectivas de análisis, 253–278. Madrid & Frankfurt am Main: Iberoamericana Vervuert. Engel, Ulrich. 2004 [1988]. Deutsche Grammatik. München: Iudicium Verlag. Erdmann, Peter. 1990a. Fokuskonstruktionen im Deutschen und Englischen. In Claus Gnutzmann (ed.), Kontrastive Linguistik, 69–83. Bern: Peter Lang. Erdmann, Peter. 1990b. Discourse and Grammar. Focussing and Defocussing in English. Tübingen: Niemeyer. Fornaciari, Raffaello. 1881. Sintassi italiana dell’uso moderno. Firenze: Sansoni. Garassino, Davide. 2014. Reverse Pseudo-cleft sentences in Italian and English. A contrastive analysis. In Iørn Korzen, Angela Ferrari & Anna-Maria De Cesare (eds.), Tra romanistica e germanistica: lingua, testo, cognizione e cultura / Between Romance and Germanic: language, text, cognition and culture, 55–74. Bern: Peter Lang.
Cleft constructions in a contrastive perspective
47
Gast, Volker & Daniel Wiechmann. 2012. W(h)-Clefts im Deutschen und Englischen. Eine quantitative Untersuchung auf Grundlage des Europarl-Korpus. In Lutz Gunkel & Gisela Zifonun (eds.), Jahrbuch des IDS 2011, 333–362. Berlin & New York: Mouton de Gruyter. Gast, Volker & Natalia Levshina. This vol. Motivating w(h)-clefts in English and German: A hypothesis-driven parallel corpus study. Geluykens, Ronald. 1988. Five types of clefting in English discourse. Linguistics 26. 823–841. Gil, Alberto. 2004. Textstrukturelle Dimensionen der Satzspaltung im Italienischen. In Alberto Gil, Dietmar Osthus & Claudia Polzin-Hausmann (eds.), Romanische Sprachwissenschaft. Zeugnisse für Vielfalt und Profil eines Faches. Festschrift für Christian Schmitt zum 60. Geburtstag, vol. 2, 369–85. Bern: Peter Lang. Grewendorf, Günther & Cecilia Poletto. 1989. La costruzione scissa: un’analisi contrastiva. Rivista di Grammatica Generativa 14. 105–142. Halliday, Michael A. K. 1967. Notes on transitivity and Theme in English. Part II. Journal of Linguistics 3. 199–244. Harries-Delisle, Helga. 1978. Contrastive emphasis and cleft sentences. In Joseph H. Greenberg (ed.), Universals of Human Language, vol. 4, Syntax, 419–486. Stanford: Stanford University Press. Hedberg, Nancy. 1988. The discourse function of cleft sentences in spoken English. Linguistic Society of America Meeting, Ms. New Orleans: Louisiana. Helfrich, Uta. 2003. Hendidas y seudo-hendidas. Un análisis empírico-diacrónico. In Fernando Sánchez Miret (ed.), Actas del XXIII Congreso Internacional de Lingüística y Filología Románica (Salamanca, 24–30 septiembre 2001), vol. 2 (1), 439–452. Berlin & New York: Mouton de Gruyter. Horn, Laurence R. 1981. Exhaustiveness and the semantics of clefts. In Victoria Burke & James Pustejovsky (eds.), Papers from the 11th annual Meeting of NELS. 124–142. Jespersen, Otto. 1937. Analytic Syntax. London: Allen and Unwin. Kiese, Jörn. 1993. Fokussierende Sätze im Deutschen und Englischen. Bern: Peter Lang. Lambrecht, Knud. 2001. A framework for the analysis of cleft constructions. Linguistics 39. 463–561. Leirbukt, Oddleif. 1969. Zum deutschen Konstruktionstyp Ich bin es, der es getan hat. Språk og Språkundervisning 5 (= Sprache und Sprachunterricht). 2–16. Metzeltin, Michael. 2010. Erklärende Grammatik der romanischen Sprachen. Wien: Praesens. Mikkelsen, Line. 2007. On so-called truncated clefts. In Ljudmila Geist & Björn Rothstein (eds.), Kopulaverben und Kopulasätze: Intersprachliche und Intrasprachliche Aspekte, 47–68. Tübingen: Niemeyer. Miller, Jim. 2006. Focus in the languages of Europe. In Giuliano Bernini & Marcia L. Schwartz (eds.), Pragmatic Organization of Discourse in the Languages of Europe, 121–214. Berlin & New York: Mouton de Gruyter. Motsch, Wolfgang 1970. Ein Typ von Emphasesätzen. In Hugo Steger (ed.), Vorschläge für eine strukturale Grammatik des Deutschen, 88–108. Darmstadt: Wissenschafliche Buchgesellschaft. Panunzi, Alessandro. 2009. Strutture scisse e pseudoscisse: valori d’uso del verbo essere e articolazione dell’informazione nell’italiano parlato. In Angela Ferrari (ed.), Atti del X Congresso SILFI. Sintassi storica e sincronica, Subordinazione, coordinazione, giustapposizione (Basilea, 30 giugno-3 luglio 2008), vol. 2, 1121–1137. Firenze: Franco Cesati. Panunzi, Alessandro. 2010. La variazione semantica del verbo essere nell’italiano parlato. Uno studio su corpus. Firenze: Firenze University Press.
48
Anna-Maria De Cesare
Pinedo, Alicia 2000. English clefts as discourse-pragmatic equivalents to Spanish postverbal subjects. Estudios Ingleses de la Universidad Complutense 8. 127–151. Prince, Ellen F. 1978. A Comparison of WH-clefts and IT-clefts in discourse. Language 54. 883–906. Roggia, Carlo Enrico. 2008. Frasi scisse in italiano e francese orale: evidenze dal C-ORAL-ROM. Cuadernos de filología italiana 15. 9–29. Roggia, Carlo Enrico. 2009. Le frasi scisse in italiano. Struttura informativa e funzioni discorsive. Genève: Slatkine. Roubaud, Marie-Noëlle. 2000. Les constructions pseudo-clivées en français contemporain. Paris: Honoré Champion. Salvi, Giampaolo. 1991. Le frasi copulative. In Giampaolo Salvi, Lorenzo Renzi & Anna Cardinaletti (eds.), Grande grammatica italiana di consultazione, vol. 2, 163–189. Bologna: il Mulino. Sedano, Mercedes. 1990. Hendidas y otras construcciones con ser en el habla de Caracas. Caracas: Universidad Central de Venezuela. Smits, Rik J. C. 1989. Eurogrammar: The Relative and Cleft Constructions of the Germanic and Romance Languages. Dordrecht: Foris. Sornicola, Rosanna. 1988. It-clefts and wh-clefts: Two awkward sentence types. Journal of Linguistics 24. 343–379. Valentini, Ada. 2012. Per una tipologia della struttura informativa: il caso delle frasi scisse in un dialetto italo-romanzo. Linguistica e filologia 32. 75–117. Van den Steen, Katleen. 2005. Cleft constructions in French and Spanish. In Nicole Delbecque, Johan van der Auwera & Dirk Geeraerts (eds.), Perspectives on Variation: Sociolinguistic, Historical, Comparative, 275–290. Berlin & New York: Mouton de Gruyter. Wehr, Barbara. In press. Cleft constructions from a typological perspective. In Sabine Diao-Klaeger & Britta Thörle (eds.), Linguistique interactionnelle contrastive. Grammaire et interaction dans les langues romanes. Tübingen: Stauffenburg.
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini*
Form and frequency of Italian Cleft constructions in a corpus of electronic news A contrastive perspective with French, Spanish, German and English
1 Introduction The aim of this paper is to assess the form and frequency of Cleft constructions (which traditionally subsume at least the classes of Cleft sentences, Pseudocleft sentences and Reverse pseudo-cleft sentences; Lambrecht 2001) in Italian by taking a contrastive perspective with four other main European languages: two Romance – French and Spanish – and two Germanic languages – German and English.¹ As our focus of interest is written contemporary Italian, priority will be given to the description and explanation of facts related to this language. The other languages primarily serve as a benchmark to assess the frequency of Cleft constructions in Italian. This paper is intended to be a further step towards a better understanding of the frequency, form and function of Type B and Type C clefts used in contemporary written Italian (on the terminology used in this paper, cf. De Cesare in this volume).² * The research presented in this paper has been funded by the Swiss National Science Foundation (Project PP00P1-133716/1, Italian Constituent Order in a Contrastive Perspective, in short ICOCP). The order of presentation of the authors is motivated as follows: the paper has been written by A.-M. De Cesare, PI of the SNSF project; Davide Garassino has translated and glossed most of the examples; overall, the results presented here are based on the research of all the members of the ICOCP research project mentioned in the title. Collectively, we would like to thank Margarita Borreguero Zuloaga for her insights and helpful comments on an earlier draft of this paper. 1 The contrastive analysis between Italian and the other languages will be further developed in the following single-authored articles of the first part of the volume, each of which is devoted to a particular Cleft construction type. More references about the languages we take into account here, i.e. Italian, French, Spanish, English and German, will also be provided in the other chapters included in this volume. 2 In this paper, we focus only on these forms of clefts because they are amongst the most frequent and standard, and because their cross-linguistic distribution in different language varietAnna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini, University of Basle
50
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
In the linguistic literature, the cross-linguistic frequency of Cleft constructions is rather unclear, on the one hand because the frequency of these forms is interpreted in different ways (and often quite impressionistically and in absolute terms, i.e., with no real reference point in mind, which would be crucial in assessing something as being “very frequent”, “frequent”, “rare” etc.), and on the other hand because the data available is not fully comparable. Moreover, we currently have more data on Cleft sentences proper than on the other types of clefts. So far, checking the claims about the frequency of clefts on the basis of the great wealth of studies on Cleft constructions has not been possible. This is due to the fact that the frequency counts found in the literature are not comparable, as they are based on different corpora: some studies focus on oral communication (as is the case of Scarano 2003; Panunzi 2009 and Roggia 2008 for Italian; Gómez González 2007 for English; Birkner 2008 for German; Sedano 1994 for Spanish spoken in Caracas); others deal with translated texts (cf. Dufter 2009 on English clefts and their translation equivalents in Italian, French, Spanish, Portuguese and German); other studies are based on a particular geographic variety and/or use rather old corpora (Collins 1991 is based on a corpus of British written texts dating back to 1961). Moreover, a cross-linguistic assessment of the frequency of clefts is difficult because some empirical data is missing (we do not have counts for contemporary written German) or because the constructions have not been searched for in the same way in different languages (regardless of the formal differences between the clefts). It is generally assumed that clefts (in particular Cleft sentences) are a diagnostic feature of neo-standard Italian. This is a new variety of contemporary Italian (Sabatini 1985; Berruto 1987) in which it is also believed that these special syntactic forms are on the rise, while being considered absolutely standard in French, i.e., as being a “vital construction found with great frequency in spoken and written French” (Katz 2000: 253³), and restricted to certain registers in German (Altmann 2009) or even in English, where they are associated with formal and higher speech registers (Collins 1991: 185). Cleft sentences proper are thus often claimed to occur more frequently in French than in both (most of) the other main Romance languages (cf. Dufter 2008) and English (Katz 2000⁴). By contrast, the
ies and text types is still not clear. In light of this fact, we do not take into account other forms of clefts, such as Inferential clefts (discussed by Atayan and Wienen in this volume). More details on the cleft forms not taken into account in this study are provided in footnote 16. 3 Cf. also, among others, Ahlqvist (2002: 278) in his discussion of Irish and other (non-)IndoEuropean languages: “French […] makes abundant use of Cleft sentences”. 4 Katz (2000: 254) does not provide specific numbers and corpus counts for the following claims: “in contrast to English, the use of the French cleft is as widespread in the spoken language as it is
Form and frequency of Italian Cleft constructions in a corpus of electronic news
51
typological literature gives us a somewhat different picture. Here, Cleft constructions seem to enjoy a special status in English, in particular as opposed to Italian or even to French, because it is claimed that only English has and frequently uses three different forms of clefts (see Miller 2006: 205). However, as we will see, and will discuss in the conclusion of this paper, this view is in need of revision, especially as far as the Italian language is concerned. The present study aims to empirically verifying the most common assumptions on the frequency of clefts in Italian and in the four other languages on the basis of a comparable and multilingual corpus of written texts. Our assessment is based on a search in a new, self-assembled corpus amounting to approximately two million words of news items collected from the Internet during the last quarter of 2011 and the beginning of 2012.⁵ In addition to assessing the frequency of Cleft constructions in Italian by adopting a contrastive perspective with French, Spanish, German and English, we describe the paradigm of forms found in our corpus. Here, too, we aim to provide a clearer picture of the types of clefts that are used in (journalistic) writing. So far, many works on these syntactic structures (especially, it seems, the ones that are cross-linguistic in nature) offer either invented examples or occurrences found in a heterogeneous corpus of texts.⁶ The following section (Section 2) describes the corpus of electronic news – the Italian Constituent Order in a Contrastive Perspective (ICOCP) corpus – which was assembled for the purpose of this study and provides the list of strings searched in the corpus to find the constructions in which we are interested. We focus on two main types of Cleft constructions (on the terminology used in this paper, see De Cesare in this volume): with the cleft constituent in medial position, after the copula (It. è la moglie che decide, E. it’s the wife that decides), and with the cleft constituent in final position (It. Quello che voglio dire è che sono contenta di vederti, E. What I want to say is that I am happy to see you), which correspond to Type B and C of the taxonomy proposed in De Cesare in this volume. Section 3 describes the variety of Cleft constructions found in our reference corpus. In
in its written form”; “The c’est-cleft is found with greater frequency than the it-cleft, especially in the spoken language” (Katz 2000: 258). 5 We would like to thank Federico Aboaf for his help in searching the ICOCP work-corpus for Italian and German Cleft constructions, as well as Boris Bouquet and Michela Puopolo for their help in collecting the texts in Italian, French, German and English. The Spanish part of the corpus has been collected by Rocío Agar Marco for her PhD thesis. 6 Cf., among others, Sornicola (1988) on clefts in English and in three Romance languages (Italian, French and Spanish), D’Achille, Proietti, and Viviani (2005), on clefts in Italian and English. See also Lambrecht (2001) on clefts in – among others – English, French, Italian, Spanish, (Brazilian) Portuguese and German and Smits (1989) on clefts in the Romance and Germanic languages.
52
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
Section 4 we assess the frequency of Cleft constructions in written (journalistic) contemporary Italian by discussing the data on this language in a contrastive perspective with the other languages taken into account in this study, revise some of the standard assumptions on the frequency of clefts in Italian and in the other languages and discuss the status of clefts in two geographic varieties of Italian (written Italian used in Italy and Switzerland). Finally, Section 5 provides new insights for the field of language typology and points to open questions.
2 ICOCP Corpus The corpus created for the purpose of this study is called ICOCP (Italian Constituent Order in a Contrastive Perspective). In this paragraph, we describe its size, design and main properties and justify some of the choices made in its construction. Before turning to the properties of the ICOCP corpus, it is necessary to explain why we decided to create a new corpus of texts instead of relying on already existing written data collections. There are two main reasons for this. The first is that there is currently no real comparable corpus or group of corpora of non-translated texts in the languages in which we are interested. There are, of course, corpora of written texts for each of the languages we are taking into consideration, but these corpora cannot be considered as being fully comparable.⁷ For instance, the French corpus FRANTEXT (http://www.frantext.fr/), which is currently the only French corpus to be based on written contemporary texts, includes 80% of literary texts. A study like ours, which aims first and foremost to assess the frequency of Clefts constructions in functional, i.e., nonliterary texts, would thus either have to renounce to a comparison between Italian and French or else, as we did, to assemble a new corpus. Another drawback of the independent corpora readily available is that they have not been collected over the same period of time. The content of the text is thus not strictly comparable. The other reason why we decided to create the ICOCP corpus for our study is that the second type of data collections that are in a sense comparable, i.e., the so-called translation corpora that are made of original as well as translated 7 It is beyond the scope of this paper to discuss the advantages and disadvantages of using the corpora already available. One point that should be mentioned here, though, is that some corpora (such as the Italian CORIS) cannot be searched in their integrality and/or have a user interface that is not ideal to find Cleft constructions. In addition, other corpora, such as NUNC (Newsgroup UseNet Corpora, a collection of Newsgroups texts in Italian and other languages, available at bmanuel.org), were still in progress when we started this research.
Form and frequency of Italian Cleft constructions in a corpus of electronic news
53
texts, are also problematic. One of the most important corpus currently available, which is used in several recent contrastive studies (on clefts, see Dufter 2009; Gast and Wiechmann 2012; Gast and Levshina in this volume as well as Korzen in this volume), is the EUROPARL parallel corpus, a collection of texts extracted from the proceedings of the European Parliament, which includes 21 European languages (http://www.statmt.org/europarl/). Now, as has been shown in several studies, and specifically for Cleft sentences in Italian translations from English by Pavesi (2005) on the basis of film dubbing, as well as in Italian translations from French journalistic texts by Brianti (in this volume), translations differ from original texts in that they show a higher frequency of linguistic features from the source text (these features can be lexical or grammatical). Moreover, as pointed out by Brianti (in this volume), the EUROPARL corpus is problematic to use because it is not always easy to determine the source text. A large amount of texts presented as originals are in fact the result of translation. This is the case in particular with the texts in English, French and German, which are three so-called “relay” languages, i.e., languages in which a text is translated (from another, minor language) before being translated into the other remaining languages of the EU. Again, this is an important drawback for a study like ours, which aims to assess the frequency of Cleft constructions in five languages as precisely as possible.
2.1 Corpus design The ICOCP corpus is a collection of full-length (there are no text samples), authentic (i.e., non-translated) and commonly occurring texts. The main properties of the ICOCP corpus are listed in Table 1 below and will be further discussed in the subparagraphs to follow: Table 1: Main properties of the ICOCP corpus 1. 2. 3. 4. 5.
Multilingual corpus of ± 2 million words Written texts Electronic texts News items Comparable texts
Before describing each of these properties in more detail, it is important to note that the texts produced by the mass media industry are one of the most commonly occurring forms of written texts (and, according to the latest Audipress surveys, online news articles are on the rise in terms of readership). This claim can be
54
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
confirmed for Italian by the design of one of the main reference corpora – the CORpus di Italiano Scritto (CORIS), which was constructed to be a balanced and representative corpus of present-day written Italian. In the CORIS corpus, news items – of different sorts – make up 38% of the whole corpus and newspaper articles amount to almost 30%.⁸ The other main forms of written texts and their weight in the CORIS corpus are the following: narrative 25%, academic writing 12%, administrative and legal writing 10%. It is thus clear that the texts produced by the written mass media industry are what people are most exposed to, and can be considered to be the most commonly occurring macrovariety in Italian. Moreover, the texts collected in the ICOCP corpus are highly representative of present-day (neo-)standard, written Italian (as well as French, Spanish, German and English). Specifically, the language used in Italian newspapers is considered to be one of the most representative cases of what has come to be called italiano neo-standard ‘neo-standard Italian’ (by Berruto 1987) or italiano dell’uso medio, ‘medium [i.e. between formal and informal] variety of Italian’ (by Sabatini 1980, 1985). It is a new variety of written (and spoken) Italian which differs in a number of important grammatical and lexical features from the (former) standard, literary-based language. One of the diagnostic features of this variety of contemporary Italian is the very use of Cleft sentences (Sabatini 1985). There are, of course, differences between the news produced by traditional media, i.e., news items published on paper, and news produced by more recent media, which are mainly spread online. In the next subsections of the paper, we will also point out some of the differences between these news items (cf. Bonomi 2002 on this issue).
2.1.1 Multilingual corpus and corpus size The ICOCP corpus is a collection of texts in five main European languages: Italian, French, Spanish, German and English. The general design and the corpus size are presented in Table 2, below.
8 The size of the news subcorpus of the CORIS, as well as of the other subcorpora, has been decided on the basis of text circulation and distribution. For more information, see the data provided on the pages of the corpus website (http://corpora.dslo.unibo.it/coris_itaProgett.html).
Form and frequency of Italian Cleft constructions in a corpus of electronic news
55
Table 2: Corpus size Total words ICOCP_Italian ICOCP_French ICOCP_Spanish ICOCP_German ICOCP_English
± 600,000 ± 400,000 ± 350,000 ± 350,000 ± 425,000
ICOCP_Total
± 2 million
It should be noted that the corpus includes texts in one language produced in different geographic locations: for instance, the texts in Italian have been produced in Italy and in Switzerland; the texts in French in France, Switzerland and Belgium; the texts in German in Germany and in Switzerland; and the texts in English in the UK, in the US as well as in Switzerland. By contrast, the texts in Spanish were only produced in Spain. It is beyond the scope of the ICOCP corpus to serve as a tool to measure the impact of the diatopic parameter in the distribution of clefts. However, our frequency counts can hint at the fact that the location of text production may play a role in the frequency of Cleft constructions (see the results on the Italian used in Italy and in Switzerland, as well as on the English used in the UK and in the United States of America; we will return to this issue in Section 4.4).
2.1.2 Written texts The news items included in the ICOCP corpus are written texts collected during the last quarter of 2011 for Italian, French, English and German, and during the first months of 2012 for Spanish. Although we are dealing with written texts, news items include an important amount of direct, reported speech. This is due to the obvious fact that news items are generally based on claims made by sources that are not part of the newspaper, but that are reported to the journalists. In this respect, interesting observations can be made in relation to the use of clefts. As discussed, for instance, in De Cesare and Baranzini (2011: 279) and De Cesare (2012, and 2014b), Italian Inverted cleft sentences (which do not exist in the other languages taken into account here) typically appear before or after new information quoted by the journalist in the form of direct speech, and have the purpose of identifying the source of this information. Here is one example found in a news release:
56
(1-I)
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
“Coloro che ci hanno attaccato l’11 settembre, volevano scavare un fossato tra gli Stati Uniti ed il resto del mondo. Hanno fallito”. A scriverlo è il presidente americano Barack Obama” (adnkronos.com) ‘“The ones that attacked us on September 11th wanted to create a gap between the United States and the rest of the world. They failed”. It is the American president Barack Obama who wrote this.’ A scriver-lo è il presidente americano Barack To write.INF-it is the president American Barack
Obama Obama
Second, these texts (some more than others, for instance news releases and news items from the free press more than the news pieces found on the free electronic pages of the main daily papers) are produced relatively fast, a fact that could lead us to consider at least some of them as a kind of “semi-spontaneous” writing. The production speed of these texts can be traced to the fact that they include different types of mistakes (grammatical, typographical etc.) that do not occur as often in the corresponding print press.⁹ It is also important to note that these mistakes are produced by professional journalists.¹⁰ Here are some examples from our corpus:¹¹ (2-I)
Saranno il Capo dello Stato Giorgio Napolitano, sua moglie Clio e il Presidente del Consiglio Silvio Berlusconi ha [instead of a] rendere omaggio, con un breve saluto, a Benedetto XVI […] (repubblica.it) ‘It will be the President Giorgio Napolitano, his wife Clio, and the Prime Minister Silvio Berlusconi to pay tribute, with a brief greeting, to Benedict XVI.’
9 There are some differences here as well: overall, the articles published on the free pages of the French Le Monde and the American New York Times contain less mistakes than the ones published on the websites of the Italian la Repubblica and Corriere della Sera. 10 These mistakes are typical of other forms of Internet-based spontaneous writing, such as blogs, chats etc., which are generally not produced by professionals. 11 For more examples based on the Italian press, and more generally on these questions, cf. De Cesare and Baranzini (2011) on news releases. In this paper, we will not correct any of the texts provided as examples, but reproduce them as we found them online, including the typos etc. Note that we also reproduce the texts with their typographical markings. The only changes that we have made concern the hyperlinks and the boldface character that some of these texts might include: we have elimitated both of these markings. Boldface marking has been used in this paper to highlight the Cleft construction in long examples.
Form and frequency of Italian Cleft constructions in a corpus of electronic news
57
(3-I)
Per quanto riguarda il risultato sportivo, il secondo derby della stagione si è chiuso con la vittoria del Lugano ai rigori. E dire è che [che è] stato proprio l’Ambrì a impressionare nella prima parte del match (cdt.ch) ‘As for the sportive result, the second derby of the season ended in favor of Lugano that won on penalties. And this even if it was Ambrì that dominated in the first half of the game.’
(4-I)
Confermano la necessità di un approccio differenziato paese per pèese [paese] delle politiche economiche (tmnews.it) ‘They confirm the need for a differentiated, country-specific approach to economic policies.’
(5-I)
BBattisti [Battisti]: chiedo perdono ma non pentito (ansa.it) ‘Battisti: I am asking for forgiveness but I don’t regret what I did.’
(6-E) Keibler – who started seeing the eternal bachelor earlier this fall – has reportedly boosted her number of paid public apearances thanks to her relationship with the silver fox (amny.com) (7-E)
Minnesota rep and presidential hopeful Michele Bachmann stopped by “Late Night with Jimmy Fallon” Monday, and house band The Roots accompanyed her entrance with a cover of Fishbone’s “Lyin’ Ass Bi----” (amny.com)
(8-E) Yet, perhaps tthe most fitting purchaser for narrative purposes would be another person seeking a truly memorable gift for their beloved (amny. com) (9-E) The somewhat odd premise behind the contest irked some New Yorkers ,, who said that plastic surgery isn’t an operation we that should be casually thrown around in something like a casino contest (amny.com) The fact that the news items collected in the ICOCP corpus present some features of spoken communication is particularly interesting for the research on Cleft constructions, because it allows us to discuss the claim that at least some forms of clefts are more typical of oral than of written communication.
58
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
2.1.3 Electronic texts The ICOCP corpus exclusively includes news items from the Internet. This, of course, does not mean that the articles have not been published later or before in one form or the other in print (i.e., in paper editions). The dynamics between online and print publishing is complex and does not concern us here. On the basis of our observation of the Italian media, it suffices to say that the articles published on the Internet seem to be (at least today) written before the pieces published the next day in the printed edition.¹² Overall, as it was partly already mentioned, the news pieces written for the Internet, generally by a specialized group of journalists, can be assumed to have been written fast, i.e., to include less editing.
2.1.4 News items The texts included in the ICOCP corpus are online news items and thus present basic differences with respect to the traditional print media. In addition to the differences mentioned earlier (respect of orthographic norms etc.), we can mention the fact that online news items are typically multimedial, hypertextual and updated frequently (several times a day) in real-time.¹³ Because they are published on the Internet, online news texts are also spread more widely as they can be read by both national and international audiences. It should be noted that one of the disadvantages of working with news items in Italian and in the other languages is the fact that these are texts that can include invisible translations from other languages, mainly from English. In general, though, this is a problem that is also true for paper articles and even for other text types (scientific etc.). The news items collected in the ICOCP corpus are of four different types (in the remainder of the paper, we will refer to these four categories as I, II, III and IV). I. Most of the news items included in our corpus feature pieces that are generally also found in the print editions of the corresponding newspapers, with only minor changes, if any. These news items have been collected mainly in free online versions of national newspapers. The websites chosen as sources for our data col-
12 The current situation thus differs from the one described in Bonomi et al. (2002), i.e., more than ten years ago, when the articles published on the Internet were first published on paper and then, unchanged, on the Internet. 13 On this issue, see for instance Lepri (2011: 29–30).
Form and frequency of Italian Cleft constructions in a corpus of electronic news
59
lection include the most visited and read ones, as can be observed on the basis of Table 3: Table 3. Sources of news items of category I ICOCP_Italian¹⁴
la Repubblica (repubblica.it), Corriere della Sera (corriere.it), il Sole 24 Ore (ilsole24ore.com), la Stampa (lastampa.it), Corriere del Ticino (cdt.ch), Swissinfo (swissinfo.ch), Zenit (zenit.org)
ICOCP_French
le Figaro (lefigaro.fr), le Monde (lemonde.fr), le Temps (letemps.ch), le Courrier (lecourrier.ch), La Libre (lalibre.be), Swissinfo (swissinfo.ch)
ICOCP_Spanish
El País (elpais.com), El Mundo (elmundo.es)
ICOCP_German
Frankfurter Allgemeine (faz.net), Tagesspiegel (tagesspiegel.de), NZZ (nzz. ch), Swissinfo (swissinfo.ch)
ICOCP_English
The Guardian (guardian.co.uk), The New York Times (nytimes.com), Swissinfo (swissinfo.ch)
It should be noted that the subcorpus of Spanish news is less diverse than the other language subcorpora in terms of the number of sources taken into account. As shown in Table 3, the Spanish articles that we collected only belong to two electronic national newspapers. This, of course, constitutes a hindrance in the comparability of the data. However, we also believe that this is only a minor drawback and that comparability of the Spanish data with the data from the other languages is still possible (as the comparisons are always made on the basis of normalized frequencies). II. The second category of news items included in the ICOCP corpus, which was only collected for Italian and which accounts for a rather small part of our Italian subcorpus, are more innovative news articles. These are news items that are not found in the corresponding print editions of the online free newspapers and are thus specific to online journalistic writing. Included here are two types of news. (i) News collected from pages called ora per ora ‘hour-by-hour’, which consists of articles belonging to a category that lies somewhere between the articles described in I and news releases (described in III); the news items found in the pages headed ora per ora differ from the other two in both size and degree of elaboration. (ii) News published on the pages headed Diretta ‘Live’ (found on repubblica.it), a term that is commonly used to refer to live reporting. What makes this second type of texts special is that, in contrast to traditional live reporting (cf.
14 For Italian, cf. the surveys of Audipress (http://www.audipress.it/).
60
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
sports), they are written and they report on subjects that go beyond the realm of sports: they can address political or social issues. The aim of this new form of journalistic writing is to provide frequently updated information on major topics of interest. These news items are generally very brief (two to five lines) and are updated frequently (sometimes even every couple of minutes). One segment of the Diretta can span over 24 hours. Note that because this text type is only included in our corpus of Italian news, the results presented in this paper will disregard the occurrences of clefts found in this subsection of the corpus. III. Another small group of news collected in the ICOCP corpus is made of news releases. When possible, i.e. available, these items have been collected directly from the websites of the news agencies (for Italian, for instance from ANSA, Adnkronos, AGI, Italpress, TMNews). When the agencies themselves do not allow access to (a selection of) news releases, we have collected them from other websites: for instance, the news releases produced by the Swiss Agency ATS, which were not available on the website of the agency at the moment of the corpus creation, have been collected via the website of Swissinfo (an online portal of Swiss news). Similarly, news releases produced by the Associated Press have been collected from the site of the New York Times. IV. The last category of news included in the ICOCP corpus belongs to the socalled free press. For Italian, we collected news from the online version of the daily Leggo, for French from the Swiss daily 20 Minutes, for German from the correspondent Swiss version 20 Minuten, and for English from the free daily AM New York. Thus, in line with the other news pieces, these items were collected from the online versions of these recent news providers and not from the print edition. As has been pointed out (cf. Dardano and Frenguelli 2008: 87–88), a great number of articles published in the free daily press are unchanged news releases (i.e., texts that do not differ from the original texts produced by the news agency) or news releases published with minimal editing. The choice of collecting only electronic news items for the ICOCP corpus can be justified as follows: (i) from a methodological point of view, we wanted to construct a multilingual corpus as homogeneous, i.e., as comparable as possible; (ii) from a practical point of view, news items are easy to reproduce; thus, creating a corpus of electronic news items is relatively easy, especially compared with the creation of a corpus of printed news items (which are either not free or difficult to gain access to); (iii) from a descriptive point of view, we are interested in investigating a type of written journalism that is relatively new and unknown (in
Form and frequency of Italian Cleft constructions in a corpus of electronic news
61
particular from a linguistic point of view¹⁵), while at the same time on the rise in terms of readership.
2.1.5 Comparable texts As much as possible, the texts included in the ICOCP corpus have been collected from the same sections of the news. The following sections of the online “papers” have been considered: national, regional (“cronaca”) and international news; politics, sports, economics and business, science and technology, culture. Moreover, as already mentioned, these texts have been collected over the same period of time, roughly from the beginning of September to the end of December 2011 for Italian, French, German and English, and at the beginning of 2012 for Spanish. As a result, the ICOCP corpus is relatively homogeneous in terms of the topics reported on (cf. for instance the death of Steve Jobs, which occurred on October 5, 2011).
2.2 ICOCP corpus sections and subsections In order to have a better idea of the size of the ICOCP corpus and of its internal composition, Table 4 provides the number of news items (roughly of articles) included in each of the four news categories described in Section 2.1.4. As we can see, there are some important differences between the five subcorpora: one of these differences is that the Spanish subcorpus does not include articles from the free press, nor does it include news releases. Consequently, the Spanish data is only partly comparable to the data of the other languages. In addition, as previously mentioned, the French, Spanish, German and English subcorpora do not include the collection of news items which result from a more innovative form of journalism (news category II). In assessing the frequency of clefts in Italian and
15 As mentioned previously, online news items do not seem to differ greatly from the traditional paper press: a large amount of electronic articles are either identical to the print editions or differ very minimally. This observation is based on a non-systematic and rather informal comparison between electronic and print press of both pay press (la Repubblica, il Corriere della Sera, Le Monde) and free press (Leggo). Interestingly, news pieces from English-based websites (cf. the New York Times) sometimes explicitly end by stating that they coincide with the paper edition. This is of course not to say that there are no differences between electronic and print papers. For a description of the linguistic features found in both types of papers, see Bonomi (2002), in particular Bonomi et al. (2002), and Bonomi (2003), as well as Gualdo (2007) for Italian.
62
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
in the other four languages, we will have to take these differences into account. We will not include clefts found in the Italian subcorpus II. Table 4: News items distribution in the ICOCP corpus Italian
French
Spanish
German
English
Total
I
942
603
541
510
332
2928
II
50
–
–
–
–
50
III
502
82
–
93
121
798
IV
157
144
–
80
362
743
Total
1651
829
541
683
815
4519
Table 5 presents the number of words included in each of the four subsections of the ICOCP corpus for the five languages and provides the percentage of words included in the first category. The variation in size has not been planned this way; in a certain way it is random. This difference can of course be significant if we want to take into consideration, as a separate and autonomous factor, the role played by different types of news in the distribution of clefts across the languages examined. The counts of clefts already available illustrate that this factor is indeed significant. The results described in Collins (1991: 187) for English, for instance, show that there are roughly two times less Cleft constructions with medial cleft constituents in press reportage than in press editorials (27 vs. 70 occurrences in 100,000 words). Table 5: ICOCP subsection sizes (in words) Italian
French
Spanish
German
English
I
395,500
325,000
350,000
300,000
264,000
%
65
80
100
85
62
II
82,000
–
–
–
–
III
60,000
30,000
–
20,000
53,000
IV
70,000
45,000
–
28,000
104,000
Total
± 600,000
± 400,000
± 350,000
± 350,000
± 425,000
Form and frequency of Italian Cleft constructions in a corpus of electronic news
63
2.3 Corpus query As already acknowledged on several occasions, in this study we are only interested in Cleft constructions with medial and final cleft constituents, i.e., with clefts of Type B and C in the new taxonomic proposal made in De Cesare (in this volume). These syntactic structures can be illustrated on the basis of the following two invented examples, respectively: it’s Stella who invited Eliana and (the person) who invited Eliana is Stella.¹⁶ All the syntactic and illocutionary manifestations of these forms of Cleft constructions were taken into account, i.e., their occurrences in main and subordinate clauses, as well as in assertions, questions etc. It should be noted that we also took into account a type of cleft that is not always considered to be a real instance of a cleft: the temporal cleft; this cleft is illustrated below for the languages in which we are interested: (10-I) È la seconda volta in tre mesi che il tribunale è colpito da una deflagrazione (ansa.it) ‘It is the second time in three months that the courthouse is hit by a blast.’ (11-F) Ce n’est pas la première fois qu’Elio Di Rupo dramatise la situation politique (lalibre.be) ‘It is not the first time that Elio Di Rupo dramatizes the political situation.’ (12-S) Es la primera vez que la madre y su entorno son tan claros en sus acusaciones (elpais.com) ‘It is the first time that the mother and those around her are so clear in their allegations.’ (13-G) Es war das erste Mal, dass er öffentlich in Erscheinung trat (swissinfo.ch) ‘It was the first time that he made a public appearance.’
16 Here is the list of Cleft constructions and cleft look-alikes that have not been taken into account in this study: Inferential cleft constructions, pseudo-conditional Cleft constructions, special interrogative Cleft constructions, Cleft constructions with no copula (“nominal Cleft constructions”), (potential) Cleft constructions with verbs different than essere / être / ser / sein / be, (potential) Cleft constructions with no subordinate clause (“Reduced clefts”), Cleft constructions with subordinate clauses different than infinitive complement clauses, some specificational and pseudo-specificational Cleft constructions, (potential) Clefts constructions with initial cleft constituent, uncertain Cleft constructions.
64
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
(14-E) It was the last time [ø] she would bow to leadership pressure (nytimes.com) Overall, these forms of clefts are relatively marginal in most of the languages: in the Spanish corpus, they amount to 3% of the total number of clefts with a medial cleft constituent, in the Italian and French subcorpora to roughly 5% (in each language subcorpus, we found 12, 10 and 12 tokens, respectively), and in English to 6% (6 occ.). These temporal clefts are more frequent in the German subcorpus, as they amount to 11% of the corpus of Type-B clefts (5 occ.). The strings of words searched for semi-manually in the ICOCP corpus to find the occurrences of Cleft constructions with medial and with final cleft constituents are given below. Table 6 provides the list of strings on the basis of which we will assess the frequency of Cleft sentences proper; Table 7 presents the list of strings on the basis of which we will assess the frequency of the “traditional” Pseudo-clefts; Table 8 gives the list of strings on the basis of which we will assess the frequency of the other types of Pseudo-clefts; Table 9 provides the list of strings on the basis of which we can assess the frequency of additional specificational constructions, not considered unanimously in the literature as being instances of Cleft constructions (for a discussion of these forms, cf. De Cesare in this volume as well as Agar Marco also in this volume). Unless specified in the tables, all the listed forms have been searched using insensitive case.¹⁷ With few exceptions, all the data have been double-checked. Table 6: List of keywords 1 (Cleft sentences proper) Italian
French
English
German
ce
it
es
essere¹⁸ ‘to be’ a ‘to’ che, cui, quale ‘who, what, which’
17 In order to have a clear overview of the forms that were searched for in our corpus, Tables 6 to 9 provide these forms in the masculine singular. More information on these forms is given in the footnotes below. 18 Since the ICOCP corpus is not tagged for parts of speech, we had to look for all the most frequent forms of the copula essere ‘to be’ (with different accent markings: è / é / e’ / e’ ‘is’), namely all the forms of the indicative present, future, past (i.e. Italian imperfetto and passato remoto), the subjunctive and non-finite verbs (infinitive and gerundive).
Form and frequency of Italian Cleft constructions in a corpus of electronic news
65
Table 7: List of keywords 2 (wh-forms) Italian
French
Spanish¹⁹
English
German
chi
qui / que
quien
who
wer
dove
où
donde
where
wo
quanto
combien
cuanto
wieviel what
was / wo-²⁰
quando
quand
cuando
when
wann
come
comment
como
how
wie²¹
perché
pourquoi
por [*²²] que
why
warum, weswegen, wieso
19 Spanish clefts with the relative que, ‘that’ (also called “que galicado”, cf. Dufter 2010), have not been searched in the corpus. One example found by chance is the following: Y es por eso que los tres documentales […] se han podido ver en la cadena Intereconomía TV ‘it is for this reason that the three documentaries [...] could appear in the channel Intereconomía TV’ (elpais.com). It should be noted that this instance of the expression es por eso que, ‘it is for this reason that’, is the only one found in the Spanish subcorpus. In the variety of Spanish used in Spain, the standard form of cleft is based on a cleft clause opened by a complex relative pronoun: y es por eso por lo que… ‘it is for this reason that...’. On Spanish clefts, cf. also Guitart 2013. 20 Note that words with a hyphen indicate the search of an incomplete word. In the case of wo-, we were looking for the forms wohin ‘where...to’, woher ‘where...from’, wonach ‘what...for/of’, womit ‘with which’, wodurch ‘how/where...through/through which’, etc. 21 Due to time constraints, the search was only possible in the NZZ (Politik, Wirtschaft and Kultur) and in Swissinfo (for a total of 95,000 words). 22 In this case, the equivalent of Italian perché ‘because, why’, in Spanish would not be por que, but por lo que (with all its possible forms: por el que, por la que etc). See also Moreno Cabrera (1999: 4273) for the cases in which the relative clause is preceded by a preposition.
66
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
Table 8: List of keywords 3 (th-forms: pronominal NPs) Italian
French
Spanish
quello, ciò²³ ‘that’
ce ‘this, that, it’ el ‘the’²⁴ lequel ‘who, whom, which’ auquel ‘to whom, to which’ duquel ‘whose, of which’ quoi ‘what’, dont ‘whose, of which’, à ‘to’
English
German
all
welcher ‘who / [the one] that’, derjenige ‘who, the one’
Table 9: List of keywords 4 (th-forms: lexical NPs) Italian
French
Spanish
English
German
cosa
chose
cosa
thing
Ding, Sache
luogo, posto
endroit
sitio, lugar
place
Ort
persona
personne
persona
one
Person
momento, periodo, moment volta
momento
time
Moment
modo
manière
manera, modo, forma
way
Art, Weise
ragione, motivo
raison
razón, motivo
reason
Grund
único, primero, último, otro, uno, mayoría, minoría
only, first, last, other, another, most, least, one
einziger, erster, letzter, anderer, meisten, wenigsten, einer
unico, solo, primo, seul, premier, ultimo, altro, più, dernier, autre, meno, uno plus, moins, un
23 After these forms, we could find a wide array of prepositions (su ‘up, above’, con ‘with’, per ‘for’, a ‘to’, in ‘in’, davanti a ‘in front of’, dietro a ‘behind’, sotto ‘down, under’, sopra ‘above’, contro ‘against’, da ‘from’, verso ‘towards’, secondo ‘according to’, senza ‘without’, durante ‘during’, attraverso ‘through’), possibly followed by the article il ‘the’. 24 I.e., el que, la que, lo que, los que, las que, al que.
Form and frequency of Italian Cleft constructions in a corpus of electronic news
67
3 Forms of the Cleft constructions in the ICOCP corpus 3.1 Cleft constructions with medial cleft constituent The Cleft constructions with medial cleft constituent found in the ICOCP corpus and included in our count are syntactic structures in which the cleft constituent occurs after the copula and is followed by either a subordinate clause (Type B-I) or a noun phrase + a subordinate clause (Type B-II).²⁵ In Type B-I clefts, the subordinate clause can be either a relative or a pseudorelative (included a complement clause) and be opened by one of the following forms: – opaque (generic) or transparent relative pronoun (cf. German), possibly coinciding with a wh-form: It. che / a ‘that, to’ (208 occ.); Fr. que / qui ‘that, who’ (255 occ.); Sp. quien / quienes ‘who’, donde ‘where’, cuando ‘when’ (31 occ.); G. der / die / das (in the nominative or in other case marking forms, 30 occ.), ‘who, which’, dass ‘that’ (7 occ.), als ‘when’ (1 occ.); E. that (65 occ.) and who (23 occ.) / which (3 occ.) – zero: English Cleft sentences can lack an overt complementizer (9 occ.; cf. [31-E]). A cleft of this type can occasionally be found in the other languages as well, but it is somewhat rare (the only French example found in the corpus is given in [20-F]; there is no example in the Italian, Spanish and German corpus data). In the Romance languages, and perhaps also in German, this form of cleft is typical of unplanned, spontaneous oral communication; here is one example found in a corpus of spoken Italian: (15-I) SAN: // se vien una cartella di sessanta milioni/ io [///] è questo / [ø] le voglio dire / io/ e voi non la dovete pagare […] // (Lablita, ex. from Scarano 2003: 187) ‘// If you receive a sixty millions collection notice/ I [///] it is this / [ø] I want to tell you / I/ and you do not have to pay it […] //’
25 This second type of cleft is not unanimously considered as such. On this issue, cf. De Cesare (in this volume) and Wehr (In press). Also note that the two groups of clefts considered here, i.e. Type B-I and B-II clefts, do not correspond to Type B-1 and B-2 clefts proposed in Table 4 in De Cesare (in this volume). In this paper, we divide Type B clefts in two larger groups, while De Cesare (in this volume) proposes a finer subdivision of the class. The same applies to Type C clefts, described in § 3.2.
68
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
Representative examples of Type B-I clefts found in our corpus include the following: (16-I) “È l’opposizione che ci rovina” “Is the opposition that us ruins” ‘“It is the opposition that is destroying us.”’ (17-I) Non fui io a uccidere Not was I to kill.INF ‘It was not me who killed.’
(adnkronos.com)
(ansa.it)
(18-F) C’est à elle que reviendra la lourde tâche It is to her that will come.3sg the hard task d’ alerter la Commission européenne et le Fonds (lemonde.fr) of warn.INF the Commission European and the Fund ‘It is her [the BCE] that will have the hard task of alerting the European Commission and the Fund.’ (19-F) C’est le pouvoir qui les dope? It is the power who them stimulates? ‘Is it power that stimulates them?’
(lalibre.be)
(20-F) C’est un garçon de 12 ans [ø] a été retiré des décombres (lemonde.fr) It is a boy of 12 years ø was rescued from the debris ‘It was a 12-year-old boy that was rescued from the debris.’ (21-S) Es la arqueóloga Gemma Menéndez quien se encarga Is the archeologist Gemma Menéndez who REFL puts in charge de su estudio (elmundo.es) of its study ‘It is the archeologist Gemma Menéndez who is in charge of studying it [i.e., a tomb].’ (22-S) Fueron precisamente estas empresas quienes nos decían que Were precisely these companies who.PL us told that les gustaría tener su propia apli (elpais.com) them would like have.INF its own app ‘It was precisely these companies that told us they would like to have their own application.’
Form and frequency of Italian Cleft constructions in a corpus of electronic news
(23-S) Fue
allí
donde Galán dio con el tesoro
69
que
buscaba (elpais.com) Was there where Galán found the treasure that searched.3SG ‘It was there that Galán found the treasure that he was looking for.’
(24-S) No será hasta final de año cuando podamos enfrentar-nos Not will be until end of year when can.1PL face-us a la mayor amenaza jamás vista ante el legendario Jefe Maestro (elmundo.es) to the biggest menace ever seen before the legendary Jefe Maestro ‘It will not be until the end of the year that we will be able to face the biggest threat ever seen in the presence of the legendary Jefe Maestro.’ (25-G) Unter den heutigen Jugendlichen seien es wieder eher die Verlierer, Amongst today’s young people are it again rather the losers die sich der rechtsextremen Szene zuwendeten (faz.net) who.PL REFL the extreme right scene turned to ‘Amongst today’s youth, it is again the losers that turned to the extreme right.’ (26-G) Es war das erste Mal, dass er öffentlich in Erscheinung trat (swissinfo.ch) It was the first time that he publicly into appearance stepped ‘It was the first time that he made a public appearance.’ (27-G) Es ist jetzt gut ein halbes Jahr her, als Philipp Rösler It is now already a half year ADV when Philipp Rösler die etwas blutige Geschichte erzählt hat (tagesspiegel.de) the somewhat bloody story has told ‘It has now been half a year since Philipp Rösler told the bloody story.’ (28-E) It is not just law-abiding British society that is being polarised (guardian.co.uk) (29-E) “[…] it was the wife who had decided on something like that” (nytimes.com)
70
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
(30-E) “Whether it was patterns of patronage or his own inclination which later impelled him towards a smaller scale and lighter touch, this View of the Rialto demonstrates that in 1768 he could impress as well as delight” (guardian.co.uk) (31-E) it was the last time [ø] she would bow to leadership pressure (nytimes. com) In Type B-II clefts, the cleft constituent is followed by a noun phrase headed by a pronoun or a noun (the subordinate clause following the noun can be a relative or a completive clause). In our corpus, the head of the noun phrase has the following forms: It. (8 occ. total) quello che (quella che etc.) ‘what, which, the one who’, il primo ‘the first one’, etc.; Fr. (1 occ. total) ce qui ‘which, what’; Sp. (47 occ. total) el que (la que, etc.) ‘what, which, the one who’ (41 occ.), la primera ‘the first one’, la manera ‘the way’, etc. (6 occ.); G. (8 occ. total) der einzige ‘the only one’, der Ort ‘the place’, etc. Representative examples from the corpus include the following: (32-I) È dunque un imputato come tutti gli altri quello che si Is so an accused like all the others the one who REFL presenta nel tardo pomeriggio al palazzo di giustizia di Milano (lastampa.it) shows in the late afternoon at the courthouse of Milan ‘It is an ordinary accused that showed up in the late afternoon at the courthouse of Milan.’ (33-I) È lui il primo tra tutti che deve pubblicamente Is he the first one between all that must publicly spiegare che cosa è realmente accaduto allora (leggo.it) explain what has really happened then ‘It is he the first person who has to publicly explain what really happened then.’ (34-I) “È questo il momento nel quale [...] sarebbe
delittuoso divider-si” (corriere.it) Is this the moment in which […] would be criminal to separate-REFL ‘It is at this moment that it would be foolish to separate.’
Form and frequency of Italian Cleft constructions in a corpus of electronic news
71
(35-F) C’est surtout la rencontre avec Alexandre Astier [...] ce qui It is above all the encounter with Alexandre Astier which fait son génie (swissinfo.ch) makes his genius ‘It is above all the encounter with Alexandre Astier which makes him [i.e., a character played on a TV show] special.’ (36-S) Será el Ministerio de Fomento el que los Will be the Ministry of Development the one that them establezca de nuevo (elmundo.es) will establish again ‘It will be the Ministry of Development who will establish them again.’ (37-S) En España será Samsung la primera en vender In Spain will be Samsung the first one to sell.INF tabletas adaptadas (elpais.com) tablets adapted ‘In Spain, it is Samsung the first company that is going to sell modified tablets.’ (38-G) Nach mehr als einem halben Jahrhundert wäre Griechenland After more than half a Century would be Greece zudem der erste souveräne Staat aus dem Kreis der arrivierten moreover the first sovereign state from the circle of the europäischen Volkswirtschaften, der Konkurs European national economies that made it who bankrupt ginge (tagesspiegel.de) would go ‘[…] it would be Greece the first sovereign country among the European national economies […] that would go bankrupt.’ Interestingly, there is only one such example in the French and none in the English sections of the ICOCP corpus. Moreover, as can be observed, there are only few cases in Italian and German. This can be explained by a compensation principle, on which we will come back below (on this principle also see Brianti, Korzen, as well as Gast & Levshina, all in this volume).
72
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
3.2 Cleft constructions with final cleft constituent The Cleft constructions with final cleft constituents (i.e. Type C clefts) found in the ICOCP corpus and included in our count are syntactic structures in which the cleft constituent is preceded by the copula and by one of the two following forms: a free relative clause (Type C-I) and a noun phrase followed by a subordinate clause (Type C-II). In Type C-I, the subordinate clause can have the form of a relative or a pseudo-relative. According to the ways the initial subordinate clause is opened, we further distinguish the following two types of clefts: A:
Type C-I clefts opened by a free / fused relative: It. dove ‘where’ (1 occ.), chi ‘who’ (1 occ.); Sp. quien/quienes ‘who’ (9 occ.), donde ‘where’ (1 occ.); G. was ‘what’ (13 occ.); E. what (61 occ.), who (1 occ.), where (2 occ.).
B:
Type C-I clefts opened by another type of free relative: It. a ‘to’ (only followed by the infinitive and only available in the cases in which the cleft constituent coincides with the subject of the corresponding monoclausal structure; there are 123 occ. of this type of clefts in our corpus; cf. example [41-I]); E. all (10 occ.).
Here are some examples from our corpus: (39-I) Dove Monti ha rivelato sapienza politica oltre ogni Where Monti has revealed cunning politic beyond every previsione, è stato nell’affrontare i dossier esplosivi Ici (la stampa.it) expectation has been in the addressing the dossiers explosive Ici ‘Where Monti revealed unexpected political cunning was in managing the explosive dossiers Ici.’ (40-I) Chi lo ha messo lì è il presidente della Repubblica (leggo.it) Who him has put there is the president of the Republic ‘Who put him there is the President of the Republic.’ (41-I) A chiamare i soccorsi è stata To call.INF the help has been ‘It was a woman that called for help.’
una donna a woman
(ansa.it)
Form and frequency of Italian Cleft constructions in a corpus of electronic news
73
(42-S) Quienes mejor se están adaptando a esta nueva realidad y Who.PL better REFL are adapting to this new reality and se muestran algo más avanzadas son Disney y REFL show.3PL a little more up-to-date are Disney and Warner (elmundo.es) Warner ‘The ones adapting better to this new reality and appear more up-to-date are Disney and Warner.’ (43-S) Donde está el verdadero fraude no es en ese grupo de Where is the true fraud not is in this group of defraudadores (elpais.com) impostors ‘Where the real fraud lies is not in this group of impostors.’ (44-G) Was bleibt, ist der Versuch der Großen, die Kleinen aus What remains is the attempt of the big the little from der Markt zu drängen (faz.net) the market to push.INF ‘What is left is the attempt of the big to drive the little ones away from the market.’ (45-E) What I’m describing here is pain (guardian.co.uk) (46-E) Where they would lose seats would be in moderate districts (nytimes.com) (47-E) “[…] who has had the largest effect on the whole planet without us really paying attention is the entire banking industry, and their disregard for the people that they’re supposed to be working for” (guardian.co.uk) (48-E) All I wanted was to get my life back (guardian.co.uk) In Type C-II, the relative clauses opening the cleft are headed by a noun phrase realized in the form either of a pronoun or of a noun (cf. the papers by De Cesare; Agar Marco as well as Baranzini in this volume for a discussion of data from Italian, French and Spanish). According to the form that opens the cleft, we further distinguish two sub-groups of Type C-II clefts:
74
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
A:
Type C-II clefts opened by a (complex) pronoun, such as It. (tutto) quello che, (tutto) ciò che ‘(all) that’, colui che ‘who’ (33 occ.); Fr. (tout) ce qui / que ‘(all) that’, ceux qui ‘who / the ones that’ (36 occ.); Sp. (todo) lo que ‘all that / what’, etc. (76 occ.); G. diejenige, die ‘the one that’ (1 occ.)
B:
Type C-II clefts opened by a generic noun (possibly modified by a small group of adjectives and numerals; cf. the keywords provided in Table 9): It. la cosa che ‘the thing that’, la persona che ‘the person who’, il primo (+ N) ‘the first (+ N’), etc. (28 occ.); Fr. la chose qui ‘the thing that’, le seul qui ‘the only one who’, le premier qui ‘the first to’, etc. (8 occ.); Sp. la persona que ‘the person who’, la primera que ‘the first one to’, etc. (36 occ.); der Einzige ‘the only one’, etc. (2 occ.); E. the thing, the place, the one / person, the way, the reason etc. (40 occ.). Representative examples from our corpus of news items are given below:
(49-I) “Quello che rileva ai fini della decisione [...] è soltanto “What is important to the ends of the decision […] is only la mancanza di prova di colpevolezza degli imputati” (leggo.it) the lack of evidence of guilt of the accused” ‘What is important in view of the decision is only the lack of evidence of guilt of the accused.’ (50-I) Colui che
abusò
di lui
all’inizio
era John Stockdale (zenit.org/it) The one who abused of him at the beginning was John Stockdale ‘The one who sexually abused him in the beginning was John Stockdale.’
(51-I) “Ciò che conta prima di tutto [...] è la credibilità della What matters first of all is the credibility of the nostra politica economica e la nostra strategia determinata our policy economic and our strategy intended per la riduzione delle spese” (leggo.it) for the reduction of the expenses ‘“What matters most is the credibility of our economic policy and our strategy to reduce expenses.”’ (52-I) “Tutto ciò che abbiamo prodotto è una T-shirt” All that have produced.1PL is a T-shirt ‘All we have produced is a T-shirt.’
(repubblica.it)
Form and frequency of Italian Cleft constructions in a corpus of electronic news
(53-I) Tutto quello che
il proprio computer (repubblica.it) All that was required was turn on.INF the own computer ‘All that was required was turning on one’s own PC.’
(54-I) “E
l’altra cosa
era richiesto
che
era
75
avevo
accendere
era
la mia reputazione […]”. (repubblica.it) “And the other thing that had.1sg was my reputation” ‘“And the other thing I had was my reputation’ ”
(55-I) L’ultimo ad utilizzar-la era stato The last one to use.INF-it had been l’ex sottosegretario alla presidenza del Consiglio
Paolo Bonaiuti (corriere.it) the ex-Undersecretary to the Prime Ministry Paolo Boaniuti ‘The last one to use it was the former Prime Ministry Undersecretary Paolo Bonaiuti.’
(56-I) L’unico a poter impedire / rimandare The only one to be able to prevent / delay la gioia infinita del tedeschino di fare il bis a 24 anni the joy infinite of the little German of repeat.INF at 24 years è il bell’inglese della McLaren (leggo.it) is the handsome English of the McLaren ‘The only one who can prevent / delay the little German’s joy of winning the Championship for the second time is the handsome Englishman of McLaren.’ (57-F) Ce qui est sûr, c’est que la révolution égyptienne n’ aura What is certain it is that the revolution Egyptian not will have vraiment mérité ce nom que si la chute de Moubarak really deserved this name ADV if the fall of Moubarak marque l’avènement d’un nouvel ordre politique [...] (letemps.ch) marks the beginning of a new order politic ‘What is certain is that the Egyptian revolution will only deserve its name if the fall of Moubarak marks the beginning of a new political order.’
76
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
(58-F) Ce que l’on veut, c’est redonner la parole What one wants it is give.INF the floor aux citoyens et aux experts pour que se tienne to the citizens and to the experts so that REFL takes place un vrai débat public (lefigaro.fr) a true debate public ‘What we want is to involve citizens and experts so that we can have a truly public debate.’ (59-F) Ceux qui
payent la garantie […], ce sont
les contribuables (lemonde.fr) Those who pay the insurance […] these are the taxpayers ‘The ones who pay the insurance are the taxpayers.’
(60-F) Tout ce que je sais c’est qu’ il fallait passer par ce All that I know it is that it had to go through this soulèvement pour arriver […] à la démocratie (swissinfo.ch) turmoil to reach.INF […] to the democracy ‘All I know is that we had to go through this turmoil to have democracy.’ (61-F) La première chose que font les autorités françaises, c’est de me The first thing that do the authorities French it is to me rentrer dedans (lemonde.fr) lead into.INF ‘The first thing that French institutions do is to attack me.’ (62-S) Lo que ya está abriendo la brecha es el reparto (elpais.com) What already is creating division is the distribution ‘What is already creating division is the distribution.’ (63-S) Todo lo que quiero es juntar-les All what want.prs1SG is to.bring-them.together ‘All I want is to bring them together.’
(elpais.com)
(64-S) El primero que ha estado luchando para que aumentaran The first one that has been fighting in order to increase.3PL he sido yo (elpais.com) have been I ‘The first one who fought to increase them [i.e., exercises for Formula One drivers] was me.’
Form and frequency of Italian Cleft constructions in a corpus of electronic news
77
(65-G) Diejenige, die Vettel an der Rennstrecke […] am meisten Arbeit The one who Vettel from the race circuit […] the most work abnimmt, ist Britta Roeske (tagesspiegel.de) takes away is Britta Roeske ‘The one who relieves Vettel from most of the work on the race circuit is Britta Roeske.’ (66-G) Der Einzige, der durch trotzige Abwesenheit auffällt, The only one who because of stubborn absence stands out Hollys Halbbruder Hollys Halbbruder ‘The only one that stands out because of his stubborn absence Halbbruder.’
ist is (nzz.ch) is Hollys
(67-E) “[…] the reason they were purchased by Sovereign was that they all offer something in the local community that is special […] ” (guardian.co.uk) (68-E) “The thing you have to remember is that the dangerous bugs are inside you” (guardian.co.uk) (69-E) The first thing she’d do is give raises
(usatoday.com)
4 Frequency assessment of Cleft constructions and data analysis In order to assess the frequency of Cleft constructions in written contemporary Italian, we will now provide the results yielded by our corpus search for the five European languages.²⁶ Columns B and C in Tables 10 and 11 give the absolute 26 Included in our data are also ambiguous cases of clefts. In our corpus, there are occurrences that can be interpreted in at least two different ways. We found cases in which a structure could be interpreted either as a Cleft construction proper or as a normal predicative construction (as in the following invented example chi ha riconosciuto Mario è un bambino ‘who recognized Mario is a child’), or as presentational structures followed by a restrictive or an appositive relative clause (è un’amica di Mario che ho conosciuto ieri ‘it is a friend of Mario’s that / whom I met yesterday’). It is also important to note that our data has a certain margin of error, which we cannot assess in terms of frequency. However, all our data has been checked by at least two members of the ICOCP research group and the cleft candidates have been thoroughly discussed by all the authors of this paper.
78
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
and normalized frequencies of Cleft constructions with medial and final cleft constituents, respectively: Table 10: Absolute frequency of Cleft constructions in the ICOCP corpus²⁷ Absolute frequency B
C
TOT
I²⁸
216
186
402
F
256
44
300
S
78
122
200²⁹
E
100
114
214
G
46
16
62
Table 11: Normalized frequency of Cleft constructions in the ICOCP corpus Frequency per 100,000 words B
C
TOT
I
42
36
78
F
64
11
75
S
22
35
57
E
24
27
51
G
13
5
18
27 In order to assess whether the different proportion of Cleft constructions observed within the five languages is statistically significant, we performed a Pearson’s Chi-square test: the result (χ2 = 143.00, df = 4, p < 0.0001) shows a very high level of significance (in this paper the significance level is chosen to be 0.01). Moreover, a post-hoc test for testing the strength of association revealed medium association between rows and columns (Cramér’s V = 0.348). 28 As stated earlier, for reasons of comparability, we do not take into account the clefts found in a category of news that is only collected in the Italian subcorpus (cf. Repubblica “Diretta” and “Ora per ora”, a corpus amounting to approximately 82,000 words; see § 2.1.4). However, the occurrences found in these texts could be discussed in the contributions following this paper. 29 This data does not include the clefts with the relative que. However, these clefts are considered to be fairly marginal in the Spanish variety used in Spain (as opposed to the Spanish used in South America; on this issue, cf. Sedano 1990 and Dufter 2010).
Form and frequency of Italian Cleft constructions in a corpus of electronic news
79
4.1 Overall frequency of Cleft constructions We will begin the discussion by focusing on the overall results provided in the column on the right hand side of Table 11 (from now on, we will of course only rely on normalized frequencies). While the higher frequency of Cleft constructions in Italian in comparison to German (78 vs. 18 occ.) is expected on the basis of previous descriptions and data counts (cf. in particular Dufter 2009; Altmann 2009; De Cesare 2011), the higher frequency of Cleft constructions in Italian in comparison to English is surprising (78 vs. 51 occ.). The very close results shown in the use of Cleft constructions in Italian and French (78 and 75 occ., respectively) are also quite unexpected. And unexpected is also the difference between Italian and Spanish. Notwithstanding the partial count for Spanish (recall that we did not search the corpus systematically for the Cleft sentences with the relative que ‘that’), it seems that as far as the overall frequency of clefts is concerned Italian is closer to French than to Spanish, which is in turn closer to English (cf. 57 and 51 occ., respectively). The data provided in Table 11 thus seem to suggest that there is, at least partially, a division between genetic families. This finding contradicts the claim made in the literature that the distribution of clefts is an areal phenomenon, declining from west to east (cf. Miller 2006; Filppula 2009), rather than something accountable in terms of genetic families. As for the differences between Cleft constructions with medial and final cleft constituents respectively, the following general observations can be made. In the linguistic literature on Italian, it is sometimes claimed that clefts of Type-C (in particular with a fused pronoun, as in chi ha risposto è Eva, Lit. ‘who answered is Eva’) are less frequent than clefts of Type-B, which include the prototypical clefts (cf. Berretta 2002: 19). However, this is not the case in our data in Italian, as reported in Table 11. This claim seems to hold only when the Cleft construction type with a final cleft constituent is interpreted in a narrow sense, i.e., as a class which merely encompasses syntactic structures opened by a wh-form (in It. chi ‘who’, dove ‘where’) or a noun phrase headed by a pronoun (It. quello che ‘what, which’, ciò che ‘what, which’, etc.). If we compare the Cleft constructions of Type-B with the more traditional Cleft constructions of Type-C (which basically correspond to the class of pseudoscisse ‘pseudo-clefts’), we arrive at the following results for Italian: there are 42 occurrences of Cleft constructions of Type-B and 7 occurrences of Type-C (this second figure is based on the data provided in Table 15). This result thus clearly confirms the claims made in the bibliography about the different distribution of these two forms of clefts. As far as English is concerned, differently from both Biber et al. (1999: 961) and Collins (1991), who find that wh-clefts (based on structures opened by who, where, why etc., as well as the all-cleft type) are much less frequent than Cleft
80
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
sentences proper (in all the registers analyzed, i.e., in conversation, in fiction, in news and in academic prose; cf. Biber et al. 1999), our results show that the difference between clefts of Type-B and the traditional class of Pseudo-clefts is not significant:³⁰ we found 24 occurrences of the first type (in 100,000 words) as opposed to 17 occurrences of the second type (cf., again, Table 15). By contrast, in informative prose, Collins (1991: 179) found 51 occurrences of Cleft sentences proper and 19 occurrences of what he labels basic Pseudo-clefts (these figures are again normalized to 100,000 words).³¹
4.2 Clefts constructions with medial cleft constituent We will now discuss the data found for clefts with medial cleft constituent in more detail, by taking into account the frequency of the different subtypes of Type-B clefts identified in § 3.1. This data is provided in Tables 12 and 13. Table 12: Absolute frequency of Type-B clefts in the ICOCP corpus³² Absolute frequency B-I
B-II
TOT
I
208
8
216
F
255
1
256
S
31
47
78
E
100
0
100
G
38
8
46
30 The result of the statistical test cannot be considered significant (Pearson’s Chi-Square, χ2 = 3.885, df = 1, p = 0.0487). 31 In Collins’ view, basic Pseudo-clefts (which differ from Reverse pseudo-clefts) include the syntactic structures opened by a fused relative pronoun (what, who, where, when, why, how, including all) and by all the noun phrases corresponding to the fused relative pronouns (the thing / one / place / time / reason / way). See Collins (1991: 27). 32 The results of the statistical tests (Fisher’s exact test, p < 0.0001; Cramér’s V = 0.644) show both a very high level of significance (i.e., the different proportion of Type-B-I and Type-B-II clefts in the five languages is highly significant) and a strong association between rows and columns.
Form and frequency of Italian Cleft constructions in a corpus of electronic news
81
Table 13: Normalized frequency of Type-B clefts in the ICOCP corpus Frequency per 100,000 words B-I
B-II
TOT
I
40
2
42
F
64
0.25
64
S
9
13
22
E
24
0
24
G
11
2
13
If we look at the figures in bold provided in Table 13 it is clear that, as far as the clefts of Type B are concerned, the picture differs with respect to what we described in the previous paragraph on the basis of all the cleft types. In this case, Italian clefts are clearly outnumbered by French clefts. In fact, French clefts are used much more often than in the other languages taken into account. What is also clear is that the frequency of Clefts of the Type-B in Italian and in French is much higher than in English and German. Thus, there seems to be the same divide along genetic families, with Spanish patterning with the two Germanic languages rather with the two Romance ones. Moreover, these results confirm some basic assumptions in the literature, namely that Cleft sentences (of type B-I) are more frequent in Italian than in German, and that they are more frequent in French than in Italian. These results also confirm the claim made in Katz (2000) regarding the higher frequency of clefts in French than in English. The fact that we find more Type-B clefts in Italian than in English (42 vs. 24 occ. per 100,000 words, see Table 13) is surprising. The percentage of English Type-B clefts may seem low compared to Italian, but the data on clefts in journalistic texts offered in previous studies are very similar to our results: Biber et al. (1999: 961) find roughly 20 occurrences of Cleft sentences proper per 100,000 words in a corpus of news, and Gómez García (2007) finds 16 occurrences. As previously mentioned, the data provided by Collins (1991) is interesting because it shows a significant range of variation between different types of journalistic texts. Collins (1991: 187) finds 27 occurrences of Type-B clefts in press reportage (this figure is again aligned to 100,000 words), a number that is close to our findings, and as many as 70 in press editorials. In addition, our data confirm the idea put forward in the literature (cf. Gómez García 2007) that there are more clefts in British than in American English: in our corpus, there are 34 occurrences of Type B clefts in the British subcorpus and 13 occurrences of the same structures in the American one (the figures are normalized to 100,000 words).
82
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
Finally, if we focus on the two groups of Type-B clefts, we see that with the exception of Spanish all the languages taken into account behave similarly. A strong preference is shown for clefts with a relative clause opened by an opaque or a transparent relative pronoun (in English, also with a covert relative marking), i.e. for Type-B-I. Clearly, one form is selected and generalized; the other forms therefore become more marginal.
4.3 Cleft constructions with final cleft constituent If we look at the number of occurrences found in the ICOCP corpus for Type-C clefts and for each of the subtypes subsumed in this general and relatively broad category, we have a very different picture. Our findings, summarized in Tables 14 and 15 show that there is quite an important cross-linguistic variation both in terms of frequency and forms. Recall that Type-C clefts are constructions in which the final cleft constituent is preceded by the copula and by one of the two following forms: a free relative clause (Type C-I), which can be headed by a fused relative (A) or by another type of relative (B), in Italian by a ‘to’ followed by the infinite and in English by all; a noun phrase followed by a subordinate clause (Type C-II), which can be introduced by a pronoun (A) or a generic noun (B). Table 14: Absolute frequency of Type-C clefts in the ICOCP corpus³³ Absolute frequency C-I
C-II
TOT
A
B
A
B
I
2
123
33
28
186
F
0
0
36
8
44
S
10
0
76
36
122
E
64
10
0
40
114
G
13
0
1
2
16
33 Once again, the results of the statistical tests (Fisher’s exact test, p < 0.0001; Cramér’s V = 0.572) show both a very high level of significance (i.e., the different proportion of Type-C-I and Type-C-II clefts in the five languages is highly significant) and a strong association between rows and columns.
Form and frequency of Italian Cleft constructions in a corpus of electronic news
83
Table 15: Normalized frequency of Type-C clefts in the ICOCP corpus Frequency per 100,000 words C-I
C-II
TOT
A
B
A
B
I
0.4
24
6.4
5.4
36
F
0
0
9
2
11
S
3
0
22
10
35
E
15
2.4
0
9.4
27
G
4
0
0.3
0.6
5
Overall, Type-C clefts are used most commonly in Italian and Spanish. In this case, the Italian data is very close to the Spanish one. Notably, both Italian and Spanish use much more Type-C clefts than French (we have three times more clefts of this type in the Italian and Spanish subcorpora than in the French one). Additionally, we observe that English Type-C clefts outnumber the French ones (27 vs. 11 occ., respectively). The data found for German, on the other hand, was expected as it is perfectly in line with the claims made in the literature that clefts are generally rare in this language (cf. Altmann 2009 as well as Gast and Wiechmann 2012 and Gast and Levshina in this volume). As for the differences related to the form of Type-C clefts, we can observe that there is generally one subtype which predominates, but that this type varies cross-linguistically. In Italian, by far the most frequent form of Type-C cleft is the one inaugurated by the generic a ‘to’ followed by the infinitive (cf. a chiamare i soccorsi è stata una donna Lit. ‘to call for help was a woman’, ansa.it). This is a very significant result and shows that this structure – which was initially called pseudoscissa ‘Pseudo-cleft’, for instance by Berretta (1994), Gil (2004) and De Cesare (2005), and has more recently been relabeled scissa inversa ‘Reverse cleft’ by Roggia (2009) – should be taken into account more systematically in the research on clefts. With respect to the other types of clefts with final cleft constituent, much less has been said on this particular construction. This bias towards the clefts opened by chi ‘who’, quello che ‘what’ etc. could perhaps be explained by the fact that the literature on clefts has so far primarily focused on oral communication. In French and Spanish there is a clear preference for the construction opened by a noun phrase (generally headed by a pronoun) and followed by a relative clause: ce qui compte pour elle, c’est de témoigner ‘what is important for her is to testify’ (lefigaro.fr); lo que procedía era declararlos nulos ‘what was appropriate
84
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
was to nullify them’ (elpais.com); note that this form is almost three times more frequent in the Spanish than in the Italian data. In the two Germanic languages, by contrast, preference is shown for Type-C clefts inaugurated by a free relative pronoun. Upon closer scrutiny, we can observe an interesting difference between English and German Type-C clefts: while we only found one form of wh-marker in the German clefts (i.e. was, as in was bleibt, ist der Versuch der Großen, die Kleinen aus dem Markt zu drängen ‘What is left is the attempt of the big to drive the little ones away from the market’, faz.net), in the English subpart of the ICOCP corpus we found three distinct wh-forms (i.e. what, as in what I’m describing here is pain, guardian.co.uk, which is clearly predominant, where and who). This latter remark could be of interest for typological studies. Our data – which is clearly based on a restricted type of text – show that there is a strong division between the forms of Type-C clefts used in the Romance and in the Germanic languages and, thus again, that this division seems more genetic than areal in nature. The wh-forms that are part of Type-C clefts are much more typical of the Germanic than the Romance languages, which use an alternative strategy based on a cleft opened by a pronominal noun phrase followed by a relative clause. Note that this type of cleft (cf. quello che è radicalmente cambiato, è anche la cultura ‘what has radically changed is also the culture’, repubblica.it) is the second most frequent form found in the Italian corpus (cf. Agar Marco in this volume for more discussion on this issue).
4.4 Cleft construction in written contemporary Italian: a few sociolinguistic remarks A first question we would like to address, but which needs to be further investigated on the basis of a larger corpus, regards the potential differences in the frequency and forms of Italian Cleft constructions used in written texts produced in two different countries, namely in Italy and in Switzerland. In order to compare the clefts used in Italy and in Switzerland, we can take a first look at the data given in Tables 16 and 17, which provide two separate counts for two types of clefts: in the first row, we find the results based solely on the parts of the ICOCP corpus produced in Italy (ICOCP_Italian_Italy), and in the second row the results found in the texts produced in Switzerland (cf. ICOCP_Italian_CH³⁴).
34 The Swiss subcorpus amounts to 75,500 words and includes articles from three different sources: from the Corriere del Ticino – and specifically from Confederazione, Ticino e Regioni, Economia and Sport for a total of 52,500 words – from the Swiss online news portal Swissinfo (13,000 words) as well as from the news Agency ATS (10,000 words of news releases).
Form and frequency of Italian Cleft constructions in a corpus of electronic news
85
Before commenting on these results, it should be noted that Tables 16 and 17 only report the findings related to the two most frequent forms of Cleft constructions with medial and final cleft constituents respectively, namely the Cleft sentences with both an explicit and implicit cleft clause (cf. Type B-1 and Type B-4 in De Cesare in this volume), and the so-called Reverse cleft sentences (Type C-4 in De Cesare in this volume), which are only found in the implicit form. Here are two examples from the Swiss subpart of the ICOCP corpus: (70-I) Sono stati proprio gli ospiti ad aprire le marcature al 25’ Have been precisely the guests to open.INF the goals at the 25th minute grazie a Tschantré (cdt.ch) thanks to Tschantré ‘It was precisely the guest team that scored the first goal thanks to Tschantré in the 25th minute.’ (71-I) Steve Jobs, il fondatore di Apple, il guru dell’hi-tech, è morto ieri a 56 anni. […] A ucciderlo è stato il cancro (cdt.ch) ‘Steve Jobs, the founder of Apple, the high-tech guru, died yesterday at the age of 56. […] It was cancer that killed him.’ A uccider-lo è stato il cancro To kill.INF-him has been the cancer Table 16: Absolute frequency of two types of Cleft constructions in the Italian parts of the ICOCP corpus³⁵ Absolute Frequency B-I
C-I (Reverse clefts)
ICOCP_Italian_Italy
195
108
ICOCP_Italian_CH
13
15
35 In assessing these data, we should take into account that the different proportion of Type B and C clefts between these two parts of the corpus is not statistically significant (Fisher’s exact test, p = 0.067).
86
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
Table 17: Normalized frequency of two types of Cleft constructions in the Italian parts of the ICOCP corpus Frequency per 10,000 words B-I
C-I (Reverse clefts)
ICOCP_Italian_Italy
4.4
2.4
ICOCP_Italian_CH
1.7
2
The data provided in Tables 16 and 17 seems to show a different picture of the frequency of the two types of clefts in the two corpora (but see footnote 35). The data we have for the Swiss subcorpus is however too small to make any sorts of generalization as far as the form of the clefts is concerned. Nevertheless, it is interesting to note that more than half of the clefts are of the implicit type, both in the Swiss and in the Italian subcorpus (we have 7 / 13 occ., i.e. 54%, and 100 / 195 occ., i.e. 51% of implicit clefts, respectively). Looking now at Type C (or Reverse) clefts, we see that the results are very similar in both corpora of Italian. Surprisingly, in the Swiss part of our work corpus Type C (Reverse) clefts are as frequent as Cleft sentences proper, which are generally believed to be the prototypical form of clefting, and in the Italian subcorpus Cleft sentences proper are indeed clearly preferred. If we take into account the data from the two columns that are related only to implicit clefts, it becomes apparent that these clefts are by far the most frequent form of clefting used in the Swiss subcorpus (80%), while they are slightly less represented in the Italian subcorpus (72%). It is crucial to take these findings into consideration when discussing the main syntactic features of the so-called neo-standard Italian (or italiano dell’uso medio ‘medium variety of Italian’). In most of the literature, it is considered that the written texts of this variety of Italian are less formal than the (traditional) standard, literary Italian as a result of the fact that features from spoken discourse are used. However, as far as Cleft constructions are concerned, this claim should obviously be made without lumping together all the different types of clefts. As shown in detail in Roggia (2009) and De Cesare (2014), contemporary written and spoken varieties of Italian (found in Italy) do not use the same type of clefts, nor do they use the same forms with the same frequency. For instance, Roggia (2009) finds that implicit Cleft sentences are much more frequent in written than in spoken communication, making them a syntactic feature of more formal registers (cf. Roggia 2009: 166–167). His findings, which are based on both the LIP and
Form and frequency of Italian Cleft constructions in a corpus of electronic news
87
the C-ORAL-ROM corpora for oral and the LISUL for written communication,³⁶ are very clear: overall, there are roughly 10% of implicit Cleft sentences in the spoken data and 40% of implicit clefts in the written data. By contrast, the socalled “spurious” temporal clefts (It. scisse spurie), such as sono due ore che ti aspetto ‘I have been waiting for you for two hours’ [Lit. ‘it is two hours that I have been waiting for you’] (Roggia 2009: 166),³⁷ are more typical of spoken registers. Roggia (2009) finds as many as 20% of these “spurious” clefts in spoken data and as little as 2% in written data. Our findings allow us to fine-tune this analysis. Interestingly – and in some sense somewhat surprisingly in light of the fact that our reference corpus includes a wide range of reported speech, interviews etc. (cf. Section 2.1.2) – the ICOCP corpus does not contain or else contains very few occurrences of several other forms of clefts typically associated with oral communication. First, and this is certainly the most expected finding of the results reported here, we do not have any instance of the form given in (72), another variant of Type C (or Reverse) cleft considered by Berretta (1996) to be substandard. (72-I) che mi ha stupito è Giovanni that me surprised is Giovanni ‘it’s Giovanni that surprised me’ (Berretta 1996: 120) Second, in our reference corpus we only found two interrogative Cleft sentences and these two tokens are actually based on an implicit cleft: (73-I) “Non dovreste essere voi a dirmelo?” “Not should be you to tell me that?” ‘Shouldn’t it be you to tell me that?’
(corriere.it)
36 The LIP corpus can be found online (http://badip.uni-graz.at/), while the C-ORAL-ROM corpus is available on a DVD included in the volume Cresti and Moneglia (eds.) (2005). In turn, LISUL is a private corpus including different types of written texts assembled at the Universities of Lausanne and Basle. 37 Spurious clefts are special in that they lack the preposition in the cleft part of the sentence and can even show agreement between the copula and the cleft constituent (the canonical cleft given in the text being è da due ore che ti aspetto ‘I have been waiting for you for two hours’ [Lit. ‘it is two hours that I am waiting for you’]).
88
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
(74-I) e chi è stato, se non lo stesso Filippo Penati, a esporsi and who was if not the same Filippo Penati to expose così tanto? (repubblica.it) so much? ‘and who was it to expose himself so much, if not Filippo Penati?’. Thus, in the Italian section of the ICOCP corpus (including here both the texts produced in Switzerland and in Italy) there is not a single instance of the interrogative clefts chi è che, lit. ‘who is that’, com’è che, lit. ‘how is that’, dov’è che, lit. ‘where is that’, etc., which are reported as being typical of oral communication (for instance by Berretta 1995: 159). Based on a count of the C-ORAL-ROM and LIP, we can assess the frequency of these interrogative clefts as being around 6 occurrences per 100,000 words.³⁸ Consequently, this result seems to indicate that, when present, questions are expressed in a much more standard form in the texts we have analyzed. Another observation to make is that, as we have seen, our corpus contains a great deal of implicit Type C (or Reverse) clefts. Our impression, based on a very partial search of the LABLITA corpus (cf. Cresti 2000), is that these clefts are confined to formal and planned speech. In light of these observations, it is interesting to note that in our reference corpus implicit Type C clefts sometimes occur in quoted speech, a fact that seems more frequent in the Swiss subcorpus: (75-I) Sulla stessa linea il “senatore” ticinese Filippo Lombardi (PPD), secondo cui anche Berna deve ora muoversi: “a pagare le spese della situazione è il Ticino”, ha detto all’ats. (swissinfo.ch) ‘The Ticinese “senator” Filippo Lombardi (PPD) is on the same page. According to him, Berna has to hurry up as well: “it is Ticino that is suffering because of this situation”, he said to ATS’ A pagare le spese della situazione To pay.INF the expenses of the situation
è il Ticino is the Ticino
By contrast, in the Italian part of the ICOCP corpus, implicit Type C clefts are almost always found outside the sections of reported speech; one of their main functions is actually to provide the source of the information reported as a direct quote (this function accounts for roughly 40% of the occurrences). Here is one
38 The data reported here is based on a corpus of 185 clefts found both in the LIP and the CORAL-ROM. We would like to express our deepest gratitude to our colleague Carlo Enrico Roggia for handing us his entire corpus of Cleft sentences found in these two corpora.
Form and frequency of Italian Cleft constructions in a corpus of electronic news
89
example in point, which also shows that these clefts are generally based on a verb of saying (dire, annunciare, scrivere ‘tell, announce, write’ etc.; on this function, cf. again De Cesare 2012, and De Cesare 2014a/b): (76-I) “Arriva in ritardo. A lezione risponde al telefono, sbiancando. Dopo pochi minuti chiede scusa ed esce dall’aula col telefono in mano. Quindi a fine lezione dà l’annuncio”. A scriverlo è Silvio (lastampa.it) ‘“He came in late. He answered his phone during the lesson and he turned pale. After a few minutes, he begged pardon and went out of the classroom with the phone in his hands. Then, at the end of the class, he announced it.” It was Silvio who wrote it.’ A scriver-lo To write.INF-it
è Silvio is Silvio
This difference could hint at the fact that the Swiss texts tend to maintain a more formal style, even in reported speech. Following up on this observation, from our data it seems that the Swiss part of the Italian ICOCP corpus shows the most formal and standard features in the use of clefts, at least of the two cleft types we have analyzed here. As we have seen, in the Swiss subpart of our reference corpus we mostly find implicit clefts (of both Type B and C); moreover, and by contrast with the texts of the ICOCP corpus produced in Italy, there is no instance of the “spurious” temporal cleft. This claim should of course be verified on the basis of a larger corpus of texts.³⁹
39 It should also be noted that, due to its political and institutional autonomy, the varieties of Italian used in Switzerland are in part subject to a different norm with respect to the varieties of Italian used in Italy. As recently suggested by Pandolfi (2011) and discussed in Berruto (2011), Italian could be described as a (weakly) polycentric language. The validity of this claim is fairly evident when we consider the lexical specificities of the Italian written and spoken in Switzerland. Not much has been claimed yet with regard to syntax and little attention has been devoted to the written language (for some remarks on syntax, cf. De Cesare 2009). Further research thus needs to be conducted on this point (cf. De Cesare et al. In prep).
90
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
5 Concluding remarks 5.1 New insights for language typology The findings presented in this paper, which are based on a large-scale analysis of a corpus of journalistic prose amounting to approximately two million words in five main European languages, can provide new insights in particular for the field of language typology. These insights will be highlighted below, after having recalled some of the most common ideas circulating in the typological literature. According to Miller (2006), whose work constitutes one of the main reference points in the typological literature on clefts constructions, clefts enjoy a special status in English as opposed to other European languages. His main claims about the forms and frequency of clefts in English and in some of the languages on which we focused in this paper include the following: English is the most striking in having three Cleft constructions which are not only described in grammars of English, but are in frequent use and occur in the English map task dialogues (and, of course, in conversation and writing). Other languages in the west of Europe have one Cleft construction or two, but not three. French has it-cleft and a wh-cleft [… but no Reverse wh-cleft]. German has clefts, which occur far less frequently than in English, but it has frequently-occurring particles. (Miller 2006: 203–204)
The data we presented in this study allow disconfirming several claims made by Miller (2006), which were subsequently picked up by other studies (Filppula 2009; Valentini 2012; Stark in this volume). Before we proceed, it should be stressed that our data differs from the data analyzed by Miller (2006), which is primarily based on map-task dialogues, i.e., on a very specific text type.⁴⁰ First, as far as the repertoire of cleft forms in English and in other languages is concerned, it is already clear from previous research that Italian has in its repertoire a type of cleft that is available neither in English nor in the other languages taken into account in this study: the implicit type C (or Reverse) cleft (a parlare è stata la presidente, Lit. ‘to speak was the president’, i.e., ‘it was the president who
40 Map task “is a cooperative task involving two participants. The two speakers sit opposite one another and each has a map which the other cannot see. One speaker – designated the Instruction Giver – has a route marked on her map; the other speaker – the Instruction Follower – has no route. The speakers are told that their goal is to reproduce the Instruction Giver’s route on the Instruction Follower’s map. The maps are not identical and the speakers are told this explicitly at the beginning of their first session. It is, however, up to them to discover how the two maps differ.” (this description is taken from http://www1.uni-hamburg.de/exmaralda/files/z2-hamatac/ public/index.html)
Form and frequency of Italian Cleft constructions in a corpus of electronic news
91
spoke’). Italian, thus, seems to have as many as four Cleft construction types, i.e., using the traditional terminology, Cleft sentences proper, Reverse clefts, Pseudoclefts and Reverse pseudo-clefts.⁴¹ In addition to confirming that the paradigm of clefts available in the repertoire of the Italian language is wider than in other European languages, the data presented in this paper also allows us to claim that all these possible structural options are actually used, at least as far as journalistic prose is concerned. In this paper, we also showed that these different cleft types do not have the same distribution in journalistic prose: the most frequent type of cleft is the prototypical cleft (40 occurrences per 100,000 words), followed by the Reverse cleft (24 occurrences per 100,000 words); Pseudo-clefts sentences are much less frequent in electronic news. Another revision ought to be made for French. According to Miller (2006: 185), “French does not have a Reverse wh-cleft construction”. This claim, however, is perhaps too strong. In our reference corpus, we found one instance of what is considered to be (at least on the basis of some of the literature: cf. the discussion provided in De Cesare in this volume) a form of Reverse wh-cleft. It should be noted that this example is found in an interview and it is produced by the Swiss francophone actor Carlo Brandt (which amounts to say that, in this case, we should not be facing an occurrence of covert translation into French from another Swiss national language or from English): (77-F) C’est surtout la rencontre avec Alexandre Astier, quelqu’un de très simple, basique, ce qui fait son génie (Swissinfo) ‘It is especially the encounter with Alexandre Astier which makes him [i.e., a character played on a TV show] special.’ Alongside this form of cleft that in this study we call Type B-cleft on the basis of the fact that the cleft constituent is found after the copula, French also has in its repertoire the more traditional form of Type A-cleft (or Reverse wh-clefts), with the cleft constituent occurring before the copula (cf. again De Cesare in this volume). Here are two examples, which show that after the cleft constituent (ça ‘this’ / le sport ‘sport’) we can find a copula construction optionally preceded by the pronoun ce ‘it, this’ (in line with the claims made for “straight” Pseudo-clefts; on this issue, see for instance Roubaud 2000):
41 In this paper, we did not say anything on this last type of cleft, which is a rather difficult type of cleft to find in written texts (where prosody is lacking). A few remarks are provided in De Cesare (in this volume). Cf. also Berretta (2002), Garassino (2014) and Wehr (In press).
92
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
(78-F) Ça c’est ce que This it is what ‘That’s what I like’
j’aime I like
(79-F) Le sport est ce que je connais le mieux The sport is what I know the best ‘Sports is what I know best’ Based on these findings, it seems more accurate to claim – contra Miller (2006) – that Type-A clefts (i.e. Reverse pseudo-clefts) are part of the linguistic repertoire of French, but that this option is hardly ever used. The rarity of this form of cleft is evident in written texts, where prosody is missing and where using a structure that resembles Type-A cleft leads to ambiguities (even the context does not always clearly indicate if we are dealing with a real Type-A cleft or with a canonical predicative construction). Thus, in written texts, when there is the need to highlight a piece of information, French resorts instead to Type-B clefts (i.e. to Cleft sentences proper).⁴² Note that these claims could certainly be extended to Italian as well. In addition to revising claims about the cleft formats found in the repertoire of English, Italian and French, our research allows for the revision of claims about the frequency of clefts in these different languages. Miller (2006) suggested that English clefts are “in frequent use” and that these structures are more frequent in English than in other European languages; however, our research clearly denies this (at least for the text type we analyzed). As we have shown, all the types of clefts taken into account in our data (which basically correspond to the class of it- and wh-clefts considered by Miller 2006) are in fact more frequent in Italian than in English. Finally, our data leads to questioning another crucial point made in the typological literature, namely the fact that the distribution of clefts is an areal phenomenon, declining from west to east (cf. Miller 2006; Filppula 2009). In the data gathered from our corpus, the distribution of clefts is at least partly best accounted for in terms of genetic families, in the sense that French is closer to Italian than to English. Our data suggests that this assumption is true not only with respect to the family of Cleft constructions as a whole, but also with respect to the subtypes of Cleft constructions we have analyzed, i.e., with syntactic struc-
42 It should be noted that while the French structure c’est le sport que j’aime le mieux is also ambiguous and allows a reading as Cleft sentence (i.e., ‘it is sports that I like the most’) and as predicative structure (‘this is the sport that I like best’, with a restrictive relative clause), the structure c’est le sport ce que j’aime le mieux, ‘it is sports what I like best’, can only be interpreted as a cleft.
Form and frequency of Italian Cleft constructions in a corpus of electronic news
93
tures with the cleft constituent occurring in medial and final position. However, what is puzzling is that not all three Romance languages pattern together: while Italian patterns with French as far as Type-B clefts are concerned, Italian patterns with Spanish as far as Type-C clefts are concerned.
5.2 Some open questions As a way of conclusion, we would like to refer to two sets of important open questions that ought to be addressed in future research. The first set, which was mentioned by Miller (2006), regards the role English has played and still plays in spreading clefts in different European languages. Miller (2006: 205) suggests the following: It is quite possible that, just as in the past languages borrowed constructions from Latin, so languages will borrow constructions from English. The borrowings will be made by people who know English, particularly written English, and are likely to appear in writing. Typologists beware! (Miller 2006: 205)
The question of whether it is possible to borrow pieces of syntax that are not already part of one’s linguistic repertoire is subject to skepticism (cf. Benincà 1993; Prince 2001). While it is certainly true that it is difficult to borrow new pieces of syntax that will become productive, it is more plausible to accept that the frequency of clefts can rise as a result of language contact induced phenomena. This means that a language (i.e. English) could merely help spreading a construction that is already part of a language system (cf. Benincà 1993: 284–285).⁴³ The possible influence of English in the spread of clefts in contemporary Italian has been suggested by D’Achille, Proietti, and Viviani (2005) and has been shown to occur for instance in film dubbing, i.e., in translation from English into Italian, by Pavesi (2005). Other attempts to explain the rise of clefts in Italian
43 Several interesting cases of language contact phenomena involving clefts can be mentioned. There is the case of contemporary written Romanian, in which clefts are allegedly emerging as a result of language contact with English (Metzeltin, personal communication; on Romanian clefts, see Metzeltin 2010 and Wehr In press). See also the case of Finnish, which has developed clefts as a result of language contact first with Swedish (a language in which clefts are claimed to be frequent; cf. Johansson 2002) and today with English (a claim made by Ahlqvist 2002: 279). Another interesting case to take into account is Hiberno-English, the variety of English spoken and written in Ireland, which uses more clefts than other varieties of English as a result of language contact with Irish, a language in which clefts are also known to be frequent (cf. Ahlqvist 2002: 277, on the basis of Filppula 1999).
94
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
through possible language contact with English, however, have failed to show a connection. De Cesare (2012) questions the fact that English could be responsible for the (high) frequency of clefts used in journalistic prose, suggesting that clefts found in news releases written in Italian cannot be accounted for as English calques, i.e., as a result of covert translation. The influence of English on Italian clefts is a complex question, though, and should be addressed again in future research. The second series of open questions we would like to mention regard the explanation of the availability and frequency of Cleft constructions across languages. One of the most popular explanations put forward in the literature to explain cross-linguistic variation in the availability and frequency of clefts is based on the so-called compensation mechanism principle, proposed by von Wartburg (1936) for French and Italian and by Jespersen (1937) for a wider spectrum of language families (Romance, Germanic, and Slavic). According to this principle, Cleft sentences (proper) “may be considered one of the means by which the disadvantages of having a comparatively rigid grammatical word-order (SVO) can be obviated” (Jespersen 1937: 86). This means that languages with a relatively fixed word order, such as English and French, should use more clefts than languages with a relatively free word order, such as Italian, Spanish and German.⁴⁴ However, on the basis of our frequency counts of clefts in journalistic prose, this principle seems to work only for German. By contrast, it is problematic for Italian and Spanish, which do have and use Cleft constructions. In the specific case of Italian, this principle does not allow an explanation for why this language has both a rich repertoire of cleft forms and uses these forms to a much larger extent than in English. As will be discussed in more detail in our subsequent papers (see Garassino; Baranzini; as well as Agar Marco), several other factors must be considered in explaining cross-linguistic variation in the form and frequency of clefts. Along with syntax, in particular with the parameter related to the degree of freedom of clause-internal constituents, it is necessary to take other linguistic focusing strategies into account, such as prosody and the lexicon. In the literature, it has already been often claimed that the lack or the rarity of clefts in a language correlates with the degree of freedom of main sentence stress (see Lambrecht 2001), as well as with the use of special particles, such as focusing and modal particles (cf. Nowakowska 2002 on the lack of clefts in Polish and Russian and the use of focus particles, and Miller 2006 on the correlation between the partial development 44 As a V2 language, German is of course bound to realizing the verb in the second position of the clause. However, the other main sentence constituents (subject, object, etc.) can be fairly freely placed before or after the main finite verb.
Form and frequency of Italian Cleft constructions in a corpus of electronic news
95
of the cleft “system” and the rich paradigm of particles in German). There are of course other variables to take into consideration in explaining the form and frequency of clefts from a cross-linguistic perspective. One of these variables, which is much less acknowledged in the literature, has to do with the role played by different “rhetorical” traditions with regards to the codification of the same text type (e.g., with the fact that there are different “stylistic” traditions in journalistic writing; on this issue, see Garassino in this volume).
References Agar Marco, Rocío. This vol. Pseudo-cleft sentences. Italian-Spanish in contrast. Ahlqvist, Anders. 2002. Cleft sentences in Irish and other languages. In Markku Filppula, Juhani Klemola & Heli Pitkänen (eds.), The Celtic Roots of English, 271–281. Joensuu: Yliopistopaino. Altmann, Hans. 2009. Cleft- und pseudocleft-Sätze (Spalt- und Sperrsätze) im Deutschen. In Rita Brdar-Szabó, Elisabeth Knipf-Komlósi & Attila Péteri (eds.), An der Grenze zwischen Grammatik und Pragmatik, 13–34. Bern: Peter Lang. Atayan, Vahram & Ursula Wienen. This vol. Inferential cleft constructions in translation. French c’est que in political texts. Baranzini, Laura. This vol. Pseudo-cleft sentences. Italian-French in contrast. Benincà, Paola. 1993. Sintassi. In Alberto A. Sobrero (ed.), Introduzione all’italiano contemporaneo. Le strutture, 247–290. Bari: Laterza. Berretta, Monica. 1994. Ordini marcati dei costituenti di frase in italiano. Vox romanica 53. 79–105. Berretta, Monica. 1995. Ordini marcati dei costituenti maggiori di frase: una rassegna. Linguistica e Filologia 1. 125–170. Berretta, Monica. 1996. Come inseriamo elementi nuovi nel discorso/3. Che mi fa paura è la nebbia. Italiano e Oltre 2. 116–122. Berretta, Monica. 2002. Quello che voglio dire è che: le scisse da strutture topicalizzanti a connettivi testuali. In Gian Luigi Beccaria & Carla Marello (eds.), La parola al testo. Scritti per Bice Mortara Garavelli, 15–31. Alessandria: Edizioni dell’Orso. Berruto, Gaetano. 1987. Sociolinguistica dell’italiano contemporaneo. Roma: Carocci. Berruto, Gaetano. 2011. Italiano lingua pluricentrica? In Anja Overbeck, Wolfgang Schweickard & Harald Völker (eds.), Lexikon, Varietät, Philologie. Romanistische Studien. Günter Holtus zum 65. Geburtstag, 15–26. Berlin & New York: Mouton de Gruyter. Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad & Edward Finegan. 1999. Longman Grammar of Spoken and Written English. London: Longman. Birkner, Karin. 2008. Relativ(satz)konstruktionen im gesprochenen Deutsch. Syntaktische, prosodische, semantische und pragmatische Aspekte. Berlin & New York: Mouton de Gruyter. Bonomi, Ilaria. 2002. L’italiano giornalistico. Dall’inizio del ’900 ai quotidiani on line. Firenze: Franco Cesati.
96
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
Bonomi, Ilaria, Elena Catalfamo, Laura Nacci & Francesca Travisi. 2002. La lingua dei quotidiani on line. In Ilaria Bonomi, L’italiano giornalistico. Dall’inizio del ’900 ai quotidiani on line, 267–349. Firenze: Franco Cesati. Bonomi, Ilaria. 2003. La lingua dei quotidiani. In Ilaria Bonomi, Andrea Masini & Silvia Morgana (eds.), La lingua italiana e i mass media, 127–164. Roma: Carocci. Brianti, Giovanna. This vol. Cleft Sentences. A translation perspective on Italian and French. Collins, Peter C. 1991. Cleft and Pseudo-cleft Constructions in English. London & New York: Routledge. Cresti, Emanuela. 2000. Corpus di italiano parlato. Campionamento (vol. II). Firenze: Accademia della Crusca. Cresti, Emanuela & Massimo Moneglia (eds.). 2005. C-ORAL-ROM. Integrated Reference Corpora for Spoken Romance Languages. Amsterdam & Philadelphia: John Benjamins. D’Achille, Paolo, Domenico Proietti & Andrea Viviani. 2005. La frase scissa in italiano: aspetti e problemi. In Paolo D’Achille & Iørn Korzen (eds.), Tipologia linguistica e società. Due giornate italo-danesi di studi linguistici (Roma, 27–28 novembre 2003), 249–279. Firenze: Franco Cesati. Dardano, Maurizio & Gianluca Frenguelli (eds.). 2008. L’italiano di oggi. Fenomeni, problemi, prospettive. Roma: Aracne. De Cesare, Anna-Maria. 2005. La frase pseudoscissa in italiano contemporaneo. Aspetti semantici, pragmatici e testuali. Studi di grammatica italiana XXIV. 293–322. De Cesare, Anna-Maria. 2009. La lingua dei giornali ticinesi. I titoli. In Bruno Moretti, Elena Maria Pandolfi & Matteo Casoni (eds.), Linguisti in contatto. Ricerche di linguistica italiana in Svizzera. Atti del convegno dell’Osservatorio linguistico della Svizzera italiana (Bellinzona, 16–17 novembre 2007), 349–367. Bellinzona: Osservatorio linguistico della Svizzera italiana. De Cesare, Anna-Maria. 2011. L’ordine dei costituenti in italiano contemporaneo e in prospettiva contrastiva con il tedesco. Tra sintassi, pragmatica e tipologia linguistica. Habilitationschrift. University of Basle. De Cesare, Anna-Maria. 2012. Riflessioni sulla diffusione delle costruzioni scisse nell’italiano giornalistico odierno a partire dalla loro manifestazione nei lanci di agenzia in italiano e in inglese. Cuadernos de filología italiana 19. 11–39. De Cesare, Anna-Maria. 2014a. Sui lanci di agenzia online in italiano e in francese. Costruzioni scisse a confronto. In Enrico Garavelli & Elina Suomela-Härmä (eds.), Atti del XI Congresso SILFI. Dal manoscritto al web: canali e modalità di trasmissione dell’italiano. Tecniche, materiali e usi nella storia della lingua (Helsinki, 18–20 giugno 2012), 619–629. Firenze: Franco Cesati. De Cesare, Anna-Maria. 2014b. Les phrases clivées en italien contemporain. Quelle caractérisation dans les volumes de référence de l’italien et quelle description à partir de corpus?. In Georgia Veldre-Gerner & Sylvia Thiele (eds.), Sprachen und Normen im Wandel, [Romanische Sprachen und ihre Didaktik], 201–216. De Cesare, Anna-Maria. This vol. Cleft constructions in a contrastive perspective. Towards an operational taxonomy. De Cesare, Anna-Maria & Laura Baranzini. 2011. La variété syntaxique des dépêches d’agence publiées en ligne. Réflexions à partir d’un corpus de langue italienne. In Angela Ferrari & Letizia Lala (eds.), Variétés syntaxiques dans la variété des textes online en italien: aspects micro- et macrostructuraux. [Special issue]. Verbum XXXIII (1–2). 247–298.
Form and frequency of Italian Cleft constructions in a corpus of electronic news
97
De Cesare, Anna-Maria, Davide Garassino, Rocío Agar Marco, Ana Albom & Doriana Cimmino. Submitted. L’italiano come lingua pluricentrica? Riflessioni sull’uso delle frasi sintatticamente marcate nella scrittura giornalistica online. Studi di grammatica italiana. Dufter, Andreas. 2008. On explaining the rise of c’est-clefts in French. In Ulrich Detges & Richard Waltereit (eds.), The Paradox of Grammatical Change. Perspectives from Romance, 31–56. Amsterdam & Philadelphia: John Benjamins. Dufter, Andreas. 2009. Clefting and discourse organization: Comparing Germanic and Romance. In Andreas Dufter & Daniel Jacob (eds.), Focus and Background in Romance Languages, 83–121. Amsterdam & Philadelphia: John Benjamins. Dufter, Andreas. 2010. El que galicado: distribución y descripción gramatical. In Alfonso Zamorano Aguilar & Carsten Sinner (eds.), La ‘excepción’ en la gramática española. Perspectivas de análisis, 253–278. Madrid & Frankfurt am Main: Iberoamericana Vervuert. Filppula, Markku. 1999. The Grammar of Irish English. Language in Hibernian Style. London & New York: Routledge. Filppula, Markku. 2009. The rise of it-clefting in English: Areal-typological and contactlinguistic considerations. English Language and Linguistics 13 (2). 267–293. Garassino, Davide. This vol. Cleft sentences. Italian-English in contrast. Garassino, Davide. 2014. Reverse Pseudo-cleft sentences in Italian and English. A contrastive analysis. In Iørn Korzen, Angela Ferrari & Anna-Maria De Cesare (eds.), Tra romanistica e germanistica: lingua, testo, cognizione e cultura / Between Romance and Germanic: language, text, cognition and culture, 55–74. Bern: Peter Lang. Gast, Volker & Daniel Wiechmann. 2012. W(h)-Clefts im Deutschen und Englischen. Eine quantitative Untersuchung auf Grundlage des Europarl-Korpus. In Lutz Gunke & Gisela Zifonun (eds.), Jahrbuch des IDS 2011, 333–362. Berlin & New York: Mouton de Gruyter. Gast, Volker & Natalia Levshina. This vol. Motivating w(h)-cleft in English and German: A hypothesis-driven parallel corpus study. Gil, Alberto. 2004. Textstrukturelle Dimensionen der Satzspaltung im Italienischen. In Alberto Gil, Dietmar Osthus & Claudia Polzin-Hausmann (eds.), Romanische Sprachwissenschaft. Zeugnisse für Vielfalt und Profil eines Faches. Festschrift für Christian Schmitt zum 60. Geburtstag, vol. 2, 369–385. Bern: Peter Lang. Gómez García, Lidia. 2007. Clefting in journalistic texts. In María Losada Friend, Pilar Ron Vaz, Sonia Santano Hernández & Jore Casanova (eds.). Proceedings of the 30th International AEDEAN Conference. [Cd-Rom]. Huelva: Servicio de Publicaciones de la Universidad de Huelva. Gómez González, María de los Ángeles. 2007. It was you that told me that, wasn’t it? It-clefts revisited in discourse. In Mike Hannay & Gerard J. Steen (eds.), Structural-Functional Studies in English Grammar: In Honour of Lachlan Mackenzie, 103–139. Amsterdam & Philadelphia: John Benjamins. Gualdo, Riccardo. 2007. L’italiano dei giornali. Roma: Carocci. Guitart, Jorge M. 2013. Del uso de las oraciones hendidas en el español actual. Revista Internacional d’Humanitats 27. 89–104. Jespersen, Otto. 1937. Analytic syntax. London: Allen and Unwin. Johansson, Mats. 2002. Clefts in English and Swedish. A Contrastive Study of It-clefts and Wh-clefts in Original Texts and Translations. Lund: Lund University dissertation. Katz, Stacey. 2000. Categories of c’est-cleft constructions. Canadian Journal of Linguistics / Revue canadienne de linguistique 45(3–4). 253–273. Korzen, Iørn. This vol. Cleft sentences. Italian-Danish in contrast.
98
Anna-Maria De Cesare, Davide Garassino, Rocío Agar Marco and Laura Baranzini
Lambrecht, Knud. 2001. A framework for the analysis of cleft constructions. Linguistics 39(3). 463–516. Lepri, Sergio. 2011. News. Manuale di linguaggio e di stile per l’informazione scritta e parlata. Milano: Rizzoli Etas. Metzeltin, Michael. 2010. Erklärende Grammatik der romanischen Sprachen. Wien: Praesens Verlag. Miller, Jim. 2006. Focus in the languages of Europe. In Giuliano Bernini & Marcia L. Schwartz (eds.), Pragmatic Organization of Discourse in the Languages of Europe, 121–214. Berlin & New York: Mouton de Gruyter. Moreno Cabrera, Juan Carlos. 1999. Las funciones informativas. Las perífrasis de relativo y otras construcciones perifrásticas. In Ignacio Bosque & Violeta Demonte, Gramática descriptiva de la lengua española 3, 4245–4302. Madrid: Espasa Calpe. Nowakowska, Aleksandra. 2002. Problématique de la phrase clivée dans une approche plurilingue. Marges linguistiques. http://icar.univ-lyon2.fr/Equipe1/documents/ MargesNowakowska.pdf (accessed 13 May 2013) Pandolfi, Elena Maria. 2011. Contatto o mancanza di contatto nell’italiano della Svizzera italiana. Considerazioni quantitative. In Raffaella Bombi, Mari D’Agostino, Silvia Dal Negro & Rita Franceschini (eds.), Atti del X congresso della associazione italiana di linguistica applicata (Bolzano, 18–19 febbraio 2010), 235–258. Perugia: Guerra Edizioni. Panunzi, Alessandro. 2009. Strutture scisse e pseudoscisse: valori d’uso del verbo essere e articolazione dell’informazione nell’italiano parlato. In Angela Ferrari (ed.), Sintassi storica e sincronica dell’italiano. Subordinazione, coordinazione, giustapposizione. Atti del X Congresso SILFI 2008 (Basilea, 30 giugno-3 luglio 2008), vol. II, 1121–1137. Firenze: Franco Cesati. Pavesi, Maria. 2005. La traduzione filmica. Aspetti del parlato doppiato dall’inglese all’italiano. Roma: Carocci. Prince, Ellen F. 2001. Yiddish as a contact language. In Norval Smith & Tonjes Veenstra (eds.), Creolization and Contact, 263–290. Amsterdam & Philadelphia: John Benjamins. Roggia, Carlo Enrico. 2008. Frasi scisse in italiano e francese orale: evidenze dal C-ORAL-ROM. Cuadernos de filología italiana 15. 9–29. Roggia, Carlo Enrico. 2009. Le frasi scisse in italiano. Struttura informativa e funzioni discorsive. Genève: Slatkine. Roubaud, Marie-Noëlle. 2000. Les constructions pseudo-clivées en français contemporain. Paris: Honoré Champion. Sabatini, Francesco. 1980. Linee di tendenza dell’italiano contemporaneo e problemi di norma. In Lauri Lindgren (ed.), La lingua italiana in Finlandia. Atti del primo convegno degli insegnanti di italiano in Finlandia (Turku, 17–18 maggio 1979). 73–91. Turku: University of Turku. Sabatini, Francesco. 1985. “L’italiano dell’uso medio”: una realtà tra le varietà linguistiche italiane. In Günter Holtus & Edgar Radtke (eds.), Gesprochenes Italienisch in Geschichte und Gegenwart, 154–184. Tübingen: Narr. Scarano, Antonietta. 2003. Les constructions de syntaxe segmentée: syntaxe, macro-syntaxe et articulation de l’information. In Antonietta Scarano (ed.), Macro-syntaxe et pragmatique. L’analyse linguistique de l’oral, 183–201. Roma: Bulzoni. Sedano, Mercedes. 1990. Hendidas y otras construcciones con ser en el habla de Caracas. Caracas: Universidad Central de Venezuela.
Form and frequency of Italian Cleft constructions in a corpus of electronic news
99
Sedano, Mercedes. 1994. Seudohendidas y oraciones con verbo ser focalizador en dos corpus del español hablado de Caracas. Thesaurus 49. 491–518. Smits, Rik J. C. 1989. Eurogrammar: The Relative and Cleft Constructions of the Germanic and Romance Languages. Dordrecht: Foris. Sornicola, Rosanna. 1988. It-clefts and wh-clefts: Two awkward sentence types. Journal of Linguistics 24(2). 343–379. Stark, Elisabeth. This vol. Frequency, form and function of Cleft constructions in the Swiss SMS corpus. Valentini, Ada. 2012. Per una tipologia della struttura informativa: il caso delle frasi scisse in un dialetto italo-romanzo. Linguistica e filologia 32. 75–117. von Wartburg, Walther. 1936. La posizione della lingua italiana nel mondo neolatino. Leipzig: Keller.
Davide Garassino*
Cleft sentences. Italian-English in contrast 1 Introduction The aim of this contribution is twofold: first, to describe the link between information structure and the discourse functions of Italian and English Cleft sentences (which belong to Type B clefts according to the classification proposed by De Cesare in this volume and adopted in De Cesare et al. in this volume¹); second, to examine the specific functions displayed by Cleft sentences within journalistic texts in Italian and English. Both goals will be achieved through a corpus-based enquiry. One of the main issues in pursuing a contrastive pragmatic study of clefts is that it relies on an accurate model of information structure. My proposal is based on two distinct but interacting criteria: referential givenness and focus / background, which will be used to identify several types of clefts according to their different information properties. In addition to a qualitative analysis, I will provide quantitative data to show the distribution of different types of clefts in the two languages. Referring to quantitative data while discussing the functions of Cleft sentences, I will also reveal similarities and differences between Italian and English. The quantitative analysis is based on a subcorpus of online journalistic texts derived from the Italian Constituent Order in a Contrastive Perspective (ICOCP) corpus (see De Cesare et al. in this volume), which contains two hundred occurrences of clefts (one hundred from the English section and one hundred from the Italian one). In the construction of this subcorpus, I relied on the same basic criteria that inspired the composition of the main corpus, particularly thematic similarities between the Italian and the English subsections, i.e., the presence of equivalent or at least analogous newspaper sections. Moreover, besides “traditional” newspapers, related textual typologies, such as free newspapers and news releases, were considered to obtain a wider (and more representative) perspective on journalistic writing. Consequently, the Italian subsection is composed
* I wish to thank Anna-Maria De Cesare, Iørn Korzen, Enrico Roggia, Barbara Wehr, and an anonymous reviewer for their many insightful observations on a preliminary draft of this paper. 1 To be more precise, the label Type-B clefts proposed in De Cesare in this volume encompasses sub-types of clefts with medial constituents that I have not considered in this paper. For this reason, I will rely on the “traditional” terminology. Davide Garassino, University of Basle
102
Davide Garassino
of the following texts: La Stampa, Il Sole 24 Ore (section: “Italia-Usa”), La Repubblica (sections: “Scienze” and “Tecnologia”) (newspapers), Leggo (free newspaper), and finally Ansa (news agency). In the English subsection, the examined texts are The Guardian, The New York Times, USA Today (newspapers), AM New York (free newspaper), and AP and ATS (news agencies). The structure of the paper is as follows: in section 2, I will provide a brief and informal survey on the main syntactic and semantic properties of Italian and English Cleft sentences. Section 3 presents a detailed analysis of the information structure of clefts (with relevant quantitative information). In section 4, I will discuss the discourse functions associated with different types of Cleft sentences, and in section 5 I will deal with their specific functions, as observed in journalistic texts. In the conclusion, I will give a final assessment of the divergences and similarities between Italian and English Cleft sentences that emerged from both the qualitative and the quantitative analyses.
2 Italian and English Cleft sentences from a descriptive point of view Cleft sentences have attracted the attention of scholars for a considerable amount of time (Jespersen 1954 is probably the first linguist to offer both a fine-grained analysis of English clefts and a surprisingly “modern” definition).² The relation between clefts, shown in examples (1) and (2), and their monoclausal counterparts, (3) and (4), has also been well recognized in the literature (the English cleft example is from Jespersen 1954, 7: 148): (1)
È la moglie che decide.³
(2)
It is the wife that decides.
2 “A cleaving of a sentence by means of it is […] serves to single out one particular element of the sentence and very often, by directing attention to it and bringing it, as it were, into focus, to mark a contrast” (Jespersen 1954, 7: 147). 3 As far as the Italian examples are concerned, I chose the following modus operandi for their translation: in made-up examples, such as (1) and (2), the English sentence is already a direct translation of the Italian one, so that no further translation is required. In all other cases, Italian examples are always accompanied by a translation.
Cleft sentences. Italian-English in contrast
(3)
La moglie decide.
(4)
The wife decides.
103
Although examples (1) to (4) are truth-conditionally equivalent, they nonetheless show some interesting differences: (1) and (2) are specificational copular sentences (for English, see Declerck 1988; Collins 1991; Dikken 2009. For Italian, see Salvi 1991), i.e., sentences in which a value is provided for a variable. Informally, the semantic content of the subordinate in (2) can be represented as the predicate decide(x). The noun phrase the wife is introduced as the argument to which the predicate applies, decide(wife). Furthermore, Cleft sentences such as (1) and (2) convey both the existence presupposition that “there exists somebody who decides”, and an exhaustive / exclusive component,⁴ paraphrasable as “only the wife decides”. Both the existence presupposition⁵ and the exhaustive information are not available in sentences (3) and (4). From a superficial⁶ syntactic point of view, Cleft sentences are constructions consisting of a main clause containing the copula and a cleft clause, which is variously considered a pseudo-relative or a restrictive relative clause (see Reeve 2011 for a recent and extensive discussion on English clefts). By abstracting from 4 I am using here the generic term “component” instead of “presupposition” because the theoretical status of exhaustiveness in Cleft sentences is a hotly debated topic in the literature. For a detailed discussion see Dufter (2009: 95–98) and De Cesare and Garassino in press. 5 To be more precise, we have to distinguish here between local and global existence presuppositions (see Charnavel 2011: 134). The local one is activated at the noun phrase level and in sentences (1) to (4) is triggered by the definite determiner (i.e., there is a wife); the global one is, on the contrary, triggered at the sentence level (i.e., there is somebody that decides) and is conveyed only by examples (1) and (2). 6 Given the scope of this article (the connection between information structure and discoursepragmatic functions), I cannot delve into formal syntactic analyses of clefts. However, it is necessary to mention briefly the fundamental contribution of derivational works to our understanding of Cleft sentences. The syntactic derivation of cleft structures, and in particular it-clefts, has been a hotly debated topic since the early years of the generative enterprise (Akmajian 1970; Higgins 1973; Chomsky 1977). To simplify, it is possible to draw a line between two major approaches in the derivational literature: the extraposition-based analyses (see Akmajian 1970; Percus 1997) and the expletive ones (see Chomsky 1977; É. Kiss 1998). The former approach basically claims that (a) the cleft clause is a right-extraposed adjunct, (b) the (deep) syntactic structure of Cleft sentences and wh-clefts (or Pseudo-clefts) is the same and (c) the cleft pronoun it and the definite description are semantically equivalent (Percus 1997). The main tenets of the expletive analysis are, on the contrary, the expletive nature of the pronoun it (considered a dummy subject) and the syntactic movement of the clefted constituent to a higher functional projection (identified as a focus phrase, FocP, in É. Kiss 1998). More recent analyses, such as Hedberg (2000) and Reeve (2011), blend aspects of both approaches.
104
Davide Garassino
language-specific features, it is possible to claim from a linear syntactic point of view that English and Italian clefts basically overlap: (5)
È Gianni che ha partecipato alla festa. Is Gianni that has attended to the party⁷ ‘It is Gianni that attended the party.’
Structure: copula + clefted constituent (in Italian elemento scisso) + cleft clause (It. frase subordinata) (see Frison 1988; Sornicola 1988; Metzeltin 1989, 2010; Berretta 1995; Roggia 2009; Panunzi 2011). (6)
It is John that attended the party.
Structure: it + copula + clefted constituent + cleft clause (see at least Delin 1992; Biber et al. 1999; Huddleston and Pullum 2002; Gundel 2006, 2008; Huber 2006; Reeve 2011; Patten 2012). The most striking language-specific differences between Italian and English clefts include the following: Italian allows an implicit variant when the clefted constituent is a subject. In such cases, the cleft clause can be formally realized by the complementizer a / ad and an infinitive form. There is however no interpretive difference between (5) and (7) (Frison 1988: 201): (7)
È Gianni ad aver partecipato alla festa. Be.PRS.3S Gianni to have.INF attended to the party ‘It is Gianni that attended the party.’
Furthermore, in Italian when the clefted constituent is a subject noun phrase or pronoun, as in (5), (7), (8), and (9) or an object noun phrase, as in (10), the copula systematically agrees in person and number with it⁸ (Frison 1988: 211; D’Achille, Proietti, and Viviani 2005: 265):
7 Here and in the following examples I follow the Leipzig glossing rules. 8 The agreement in gender is also necessary when a past participle is involved (stata ‘been’ in the following example is the feminine form of the past participle of essere ‘to be’): è stata Maria che ha partecipato alla festa, ‘it was Mary that attended the party’.
Cleft sentences. Italian-English in contrast
105
(8)
Sono Gianni e Luca che hanno partecipato Be.PRS.3PL Gianni and Luca that have.PRS.3PL attended alla festa. to the party ‘It is Gianni and Luca that attended the party.’
(9)
Siete voi che avete partecipato alla Be.PRS.2PL you.PL that have.PRS.2PL attended to the ‘It is you that attended the party.’
(10)
Sono Gianni e Luca che ho invitato alla Be.PRS.3PL Gianni and Luca that have.PRS.1SG invited to the festa. party ‘It is Gianni and Luca that I invited to the party.’
festa. party
If the clefted constituent is an object pronoun, then two options are available: if the pronoun appears in the nominative form, the copula agrees with it, as in (11). If the pronoun is in the accusative form, the agreement is normally not realized, as in (12) (Metzeltin 1989: 161; Salvi and Vanelli 2004: 312): (11)
Sei tu che Be.PRS.2SG you.NOM that ‘It is you that they want.’
vogliono. want.PRS.3PL
(12)
È te che Be.PRS.3SG you.ACC that ‘It is you that they want.’
vogliono. want.PRS.3PL
In English, the pronoun it is always used as an invariant introducer of the construction (not by chance English Cleft sentences are known in the literature as it-clefts). Moreover, the choice of the complementizer is not restricted to that, as in example (13), but it also includes wh-forms, as in (14), and a zero option (or nocomplementizer), as in (15) (this last example is from the ICOCP corpus):⁹ (13)
It is John that attended the party.
9 The choice of one of these three options is not random, as shown by Dikken (2009).
106
Davide Garassino
(14)
It is John who attended the party.
(15)
It is then Ø I saw the clipboard. (guardian.co.uk)
A contrastive syntactic analysis of the two languages can of course be pursued more deeply, as shown by Sornicola (1988) and D’Achille, Proietti, and Viviani (2005) to which I refer for a more detailed discussion.¹⁰ For the purposes of this paper, I will now turn attention to some properties that concern the relationship among the syntax, semantics, and pragmatics of clefts. First, the biclausal structure of Cleft sentences allows the separation of presupposed and asserted content (Gómez-González 2007: 120–121): the first part, formed by the copula and the clefted constituent, introduces (i.e., asserts) the value that satisfies the variable x contained in the second part (the subordinate); the cleft clause conveys the presupposition that “there is an x to which the main predicate applies”. The biclausal syntax has also another important consequence because the isolation of the clefted constituent (whose referent is singled out as the value satisfying the variable) results in its pragmatic highlighting. Another motivation for using Cleft sentences is that they allow the interpretation without ambiguity of the semantic scope of logical operators (e.g., negation) within the sentence. According to Lambrecht (2001: 489–490) and Dufter (2009: 108), a monoclausal sentence, such as (16), is ambiguous between two readings (corresponding to sentences [17] and [18]); such ambiguity disappears by using a cleft as the narrow negation can have scope over only the clefted constituent, as in (17), or the cleft clause, as in (18): (16)
John did not attend the party.
(17)
It was not John that attended the party.
(18)
It was John that did not attend the party.
The Cleft sentence in (17) presupposes that someone attended the party, but that referent is not to be identified with John in contrast with someone else (“it was not John that attended the party, but Jen”). On the contrary, the sentence in (18) presupposes that someone did not attend the party and John is identified as the one that did not come. 10 I also refer to De Cesare (2012) for a discussion on the form, frequency, and functions of Italian and English Cleft constructions (including Cleft sentences) in a very specific textual typology, i.e., news releases.
Cleft sentences. Italian-English in contrast
107
In conclusion, the breaking of linear syntax into two parts is one of the most distinctive features of Cleft sentences. This cleaving also determines important consequences at the level of information structure as some of its relevant properties (e.g., the distribution of given and new information and focus / background) can be mapped onto linear cleft structure in different ways. In the next section, I will explore the information structure properties of Cleft sentences and offer a classification of Italian and English occurrences.
3 Information structure and Cleft sentences According to both prosodic and semantic-pragmatic criteria, Cleft sentences have traditionally been considered focalizing devices, as shown in examples (5) and (6): the highlighted constituent receives a pitch accent,¹¹ and it indicates the presence of relevant alternatives for semantic interpretation¹² (i.e., in sentences (5) and (6), Gianni and John are picked from a set of relevant contextual alternatives of the same semantic type, such as {Giorgio, Mary, Anna, Steve, …}). In cartographic approaches (É.Kiss 1998), the clefted constituent is claimed to be hosted in a special focus projection (FocP) in the syntax, where it receives an exclusive interpretation (i.e., the same semantic reading as exhaustive operators such as only. In example [6], for instance, given a set of alternatives, only John among them attended the party).¹³ However, not every cleft can be assigned a focalizing interpretation: in many occurrences of Cleft sentences no alternative exclusion and no overall contrast seem to be conveyed. Consider the following examples: (19)
Gianni ha bevuto troppo ieri sera. È per questo che oggi non si è presentato a lezione.
(20)
Gianni drank too much last night. It is for this reason that he didn’t show up to class today.
(21)
It was in 1886 that (…) Lewin published the first systematic study of the cactus. (Huber 2006: 551)
11 In the remainder of the paper, given my focus on written data, I will not delve into prosodic factors. 12 This intuitive definition of focus (see Krifka 2007: 18) is probably the most widespread in the current semantic literature. See also below in this section and note 19. 13 This is, however, a very controversial position (see Wedgwood et al. 2006).
108
Davide Garassino
(22)
È un anno che lavoro come traduttore.¹⁴ ‘I have been working as a translator for one year.’
(23)
È con grande piacere che scrivo la prefazione di questo volume.
(24)
It is with great pleasure that I am writing the preface to this volume.
In all these cases, the referent of the clefted constituent does not evoke relevant alternatives. In example (21), as a discourse starter, the prepositional phrase in 1886 does not normally contrast with other prepositional phrases, such as in 1885, in 1901 or in 2012, but instead creates a temporal setting. A similar observation applies to examples (23) and (24). Since Prince (1978), various proposals have countered the assumed monofunctionality of clefts. Prince first suggested a bipartition between canonical stressed-focus clefts (i.e., the focalizing clefts seen in examples [5] and [6]) and the so-called presupposition-informative clefts (a relative broad category under which we can subsume examples [19] to [24]). In comparison with monofunctional analyses, Prince’s proposal has the advantage of emphasizing the non-univocal mapping of information structure properties onto the linear syntax of clefts. As I will show below, information can in fact be distributed in different ways between the clefted constituent and the cleft clause. Moreover, recent research found a correlation between different types of clefts (e.g., stressed-focus clefts and informative-presupposition clefts) and different syntactic properties (see Dikken 2009).¹⁵ Prince’s suggestions have been further developed in the last decades. Declerck (1988) considered the informative-presupposition type described by Prince too undifferentiated and put forward a further subclassification that depended on the information status of the clefted constituent. He thus proposed two new labels: 1) unstressed-anaphoric-focus clefts with a given clefted constituent and a new cleft clause, as in examples (19) and (20); and 2) discontinuous clefts, in which
14 This kind of occurrence is known as scissa temporale or spuria ‘temporal or spurious cleft’ in the Italian literature. Grammatically, it is a quite peculiar construction as it allows the omission of the preposition da ‘since’: è un anno is thus equivalent to è da un anno, Lit. ‘it has been one year.’ Due to their idiosyncratic properties, temporal clefts are usually considered a special case of Cleft sentences (see Roggia 2009: 125–128 and Panunzi 2011). In the classification proposed in this paper, temporal clefts are included as well (see section 4.3). 15 For instance, in the English informative-presupposition clefts, the complementizer that cannot be omitted and that cannot be used if the subject is [+ human] (for a detailed discussion, see Dikken 2009). Different types of clefts thus seem to show different syntactic as well as prosodic features. These findings contribute to weakening the position of scholars who defend a unitary analysis of clefts, such as Lambrecht (2001).
Cleft sentences. Italian-English in contrast
109
both the clefted constituent and the cleft clause are new, as in (21) to (24) (these examples are usually found at the beginning of the discourse). Although it is not based on givenness, but on the distribution of Topic and Comment within the sentence, Gundel’s and Hedberg’s classifications (see Gundel 2006, 2008; Hedberg and Fadden 2007) of three types overlap Declerck’s classification. According to Gundel and Hedberg, Cleft sentences can be divided into the following: Topicclause clefts (corresponding to stressed-focus clefts), Comment-clause clefts (corresponding to Declerck’s unstressed-anaphoric-focus clefts) and, finally, all-Comment clefts (corresponding to Declerck’s discontinuous clefts). Recently, Gómez-González (2007), following Collins (1991), proposed a classification of English clefts based on referential givenness, specifically recoverability; in this perspective, “given information is that which the speaker presents as recoverable from prior linguistic context or the situation, and new information is that which is presented as non recoverable” (Collins 1991: 91). Her classification displays three basic subtypes of clefts: new-given, given-new, and new-new.¹⁶ Whereas the aforementioned studies are based almost exclusively on the analysis of English data, Dufter (2009) offers a cross-linguistic enquiry on German and Romance Cleft sentences. In developing his taxonomy, he relies on the notions of focus and background and their distribution in the sentence. By so doing, he distinguishes three main classes: a focus-background (represented by examples [5] and [6] above), a background-focus (examples [19] and [20]), and finally an all-focus type (examples [23] and [24]). Finally, Roggia (2009) puts forward a fine-grained taxonomy of Italian clefts primarily based on information status (given, new, and inferable) and, as a secondary criterion, the cognitive activation of referents in the discourse. The interplay of these two parameters results in a very detailed description of the data (Roggia 2009: 139–156). In summary, it is interesting to note that these proposals, although often founded on different theoretical grounds, are substantially similar in their results. This fact strongly suggests that Cleft sentences are sensitive to distinct levels of information structure and that a truly comprehensive view of their information properties can be reached only if we consider these separate levels to interact.¹⁷ 16 Within every type she also distinguishes between “inferable” and “contrastive” information (see Gómez-González 2007: 127). Quite interestingly, both Collins (1991: 110) and GómezGonzález (2007: 124), by introducing the “contrastive” label (regarded as a category of newness), implicitly refer to another level of information structure (focus / background). 17 In other words, I want to avoid the “theoretical trap”, which is well explained in Dufter (2009: 101). According to Dufter, many analyses fail “to distinguish between two kinds of givenness, called referential and relational givenness […] Put simply, a referring expression is referentially given if and only if the discourse referent that it denotes has been previously introduced into the discourse representation. Relational givenness, by contrast, must be evaluated relative to
110
Davide Garassino
My aim in this paper is to propose an “integrated” classification, based on two independent criteria: referential givenness (Declerck 1988; Collins 1991; GómezGonzález 2007) and the distribution of focus and background in the sentence (Dufter 2009). In the following pages, I will show that this choice is motivated by both empirical and theoretical factors. Regarding the basic terms of my analysis, I will use the following definitions. The given information label includes information that has been previously introduced in the discourse or that is otherwise pragmatically retrievable (the so-called inferable information, see Prince 1981).¹⁸ On the contrary, the new information label subsumes information that has not been previously mentioned and is not immediately retrievable (see Gómez-González 2007: 123). Focus and background, on the other hand, operate at the sentence level and signal an asymmetry between the part of the sentence (usually a constituent) “that relates an utterance to a set of relevant alternatives” (Zimmermann and Onea 2011: 1652)¹⁹ and the portion of the sentence that does not evoke alternatives (background). Considering these parameters at the same time, it is possible to derive a basic taxonomy: Table 1: A basic information-structural typology of Cleft sentences
Type 1
Type 2
Type 3
Clefted Constituent
Cleft Clause
New
Given
Focus
Background
Given
New
Background
Focus
All-New All-Focus
the sentence wherein an expression occurs”. Relational givenness refers thus to other levels of information structure, i.e., to the distribution of focus and background (Dufter 2009: 101) and Topic and Comment (Gundel 2006, 2008) within the Cleft sentence. 18 Given and inferable information are usually grouped together (see Ward and Birner 2004: 156 and the bibliographical references they provide). 19 In the sentence JOHN attended the party, the focalized subject points out that alternatives of the form x attended the party are relevant for semantic interpretation (i.e., in a more communicative perspective, the uttered sentence is evaluated by the hearer with respect to the potential knowledge that someone attended the party, see Zimmermann and Onea 2011: 1654). Please note that alternatives do not need to be introduced in the discourse (see also section 4.1) as focus does not necessarily express an explicit opposition between the selected value and the discarded alternatives. For a more detailed and technical discussion see Zimmermann and Onea (2011: 1652–1655).
Cleft sentences. Italian-English in contrast
111
Type 1 represents the canonical Cleft sentence. This kind of cleft can be the answer to wh-questions such as: “who is working today?”. Consider the following examples in Italian and English: (25)
È Gianni che lavora.
(26)
It is Gianni who is working.
The content of the cleft clause is given, while the clefted constituent conveys new information and is associated with focus. Within a set of contextually relevant alternatives, John is the selected one (and is consequently pragmatically highlighted). However, as we will see below, focus and new information do not always coincide. Furthermore, this subtype of clefts apparently violates a basic information principle according to which old information usually precedes new information (i.e., the given before new principle, see Gundel 1988: 229). Type 2 shows the reverse image of the information distribution just examined (thus providing a more “natural” distribution from old to new information). The clefted constituent refers to given information (represented by the anaphora “questo / this” in sentences [27] and [28]), whereas the cleft clause provides (totally or partially) new information. In addition, at the level of focus / background, we observe the mirror image of Type 1: the clefted constituent becomes the background, while the cleft clause (or part of it) is associated with focus. In a communicative situation, in which a famous athlete had to quit competing because of a severe injury, we could find the following cleft: (27)
È per questo che i fan verseranno molte lacrime.
(28)
It is for this reason that fans will shed many tears.
Type 2 clefts can also allow the resumption of a spatio-temporal setting (a “frame Topic”), previously introduced in the discourse (Dufter 2009: 102). Consider the following examples in which we imagine somebody talking about the goals he achieved in a certain city. At some point in the conversation, the speaker adds that he also met his future wife in the same place: (29)
È sempre qui che ho incontrato la mia futura moglie.
(30)
It is also here that I met my future wife.
112
Davide Garassino
Finally, Type 3 is an all-new or all-focus²⁰ cleft and is typically located at the beginning of a discourse / text: (31)
È con piacere che vi presento i nostri ospiti.
(32)
It is with pleasure that I am going to introduce to you our guests.
(33)
Sono tre anni che vive in Scozia.²¹ ‘He has lived in Scotland for three years.’
Before discussing in detail the discourse functions of the three types (section 4), it is important to observe from a methodological point of view some problematic occurrences found in the ICOCP data and how I chose to deal with them. Consider for instance the following two examples: (34)
“Assolta”, “Una donna libera”, “Amanda liberata dalla corte”. Le testate statunitensi si inseguono, la Knox campeggia su tutte le prime pagine. Cnn e Fox News. New York Times, Wall Street Journal, Boston Globe, Miami Herald e Los Angeles Times: non c’è un sito che non dia risalto alla notizia. Senza contare i giornali locali di Seattle, la città dove è nata la Knox: la concittadina è al centro dell’informazione. Anche in Gran Bretagna al caso viene dato grande risalto, con occhio molto più critico, però, sulla fama di Amanda, che offusca e quasi sembra far dimenticare la vera vittima di questa storia, la cittadina britannica Meredith Kercher. La decisioni della Corte d’appello di Perugia risuona anche in Francia, Spagna e Germania. Su tutti i quotidiani è Amanda a conquistare la scena²²: Sollecito è citato raramente, e sempre in secondo piano. (repubblica.it)²³
20 These clefts are known as all-focus, because “the whole cleft clause is the domain of a single sentence-focus” (Dufter 2009: 99). 21 As already observed in note 14, such sentences are known in the Italian literature as scisse temporali or spurie ‘temporal or spurious clefts.’ English does not have, however, a truly equivalent construction (see D’Achille, Proietti, and Viviani 2005: 258). In the ICOCP corpus I found only one occurrence similar to the Italian examples: it’s only been 40 years that we have the women’s vote (swissinfo.ch). 22 From now on, the Cleft sentences will be indicated in boldface within the examples. Please also note that the examples taken from newspapers are left unaltered: possible mistakes are included. 23 This example is drawn from the ICOCP corpus but not from the specific subcorpus analyzed in this paper (see the Introduction). I chose nonetheless to discuss this occurrence because it perfectly represents the kind of problematic examples that I am going to examine.
Cleft sentences. Italian-English in contrast
113
‘“Acquitted”, “A free woman”, “Amanda set free by the court”. American newspapers run similar headlines. Amanda Knox stands out on the front pages of the CNN and Fox News, The New York Times, The Wall Street Journal, The Boston Globe, The Miami Herald, and The Los Angeles Times: every website gives prominence to this news. In local newspapers in Seattle, the native city of Amanda Knox, the fellow citizen is the center of attention. In Great Britain, the news also has been given much prominence, but with a critical eye on Amanda’s reputation, who obfuscates and relegates to the background the true victim of this story, the British citizen Meredith Kercher. The decision of the court of appeal of Perugia echoes in France, Spain, and Germany. In all newspapers, it is Amanda that is in the limelight; Sollecito is seldom mentioned and [is] always in the background.’ (35)
These reports don’t stop to explain that more than 40% of people found ‘fit for work’ appeal against the decision, and of these, around 40% have the decision overturned. Instead they tend to assume that those found ‘fit for work’ have essentially been trying to cheat the system, whereas in fact it is often the system that has treated them badly. So if you are at one of the party conferences, take a minute to pick up the Daily Stigma newspaper and read the full story behind the headlines. It exposes the truth behind benefit fraud – 99.5% of sickness benefit claims are genuine. (guardian.co.uk)
In such cases, “è Amanda a conquistare la scena”, ‘it is Amanda that is in the limelight’, and “it is often the system that has treated them badly”, the clefted constituent represents the focus and pragmatically contrasts with alternatives more or less explicitly presented in the text. For instance, Amanda contrasts with Sollecito (the other young indictee of the process) and the system in sentence (35) contrasts with the explicitly mentioned people ‘fit for work’. Amanda and the system are both foci, but they are not at all discourse-new information. When the Cleft sentences in (34) and (35) are used, Amanda and the system are already discourse-given. In principle, if only the level of referential givenness were considered, it would be possible to make up another category for such examples, e.g., an “all-given” type (Johansson 2002 and Hasselgård 2004 actually introduced a special type / class for all-given clefts), but in the end this choice is problematic from a theoretical point of view. First, an all-given cleft would go against what has repeatedly been observed by many authors (e.g., Delin 1989; Delin and Oberlander 2005; Berretta 1996: 116; Dufter 2009: 101), i.e., that putting together old
114
Davide Garassino
elements always results in a new connection between them.²⁴ Second, creating a separate class for all-given occurrences would deny the intuitive but striking resemblance between clefts such as (34) and (35) and prototypical Type 1 examples, such as (25) and (26) (a similarity caused by the highlighted and focal nature of the clefted constituent). These considerations suggest that referential givenness alone is not sufficient to capture the informational and functional properties of Cleft sentences. Integrating referential givenness with insights from other levels of information structure (in particular focus / background) thus seems a necessary move. More specifically, the introduction of a cleft type that includes occurrences whose clefted constituent is both given and focalized is required. Another aspect is worth mentioning: in many examples (especially when the clefted constituent is a subject noun phrase, such as [34] and [35]) the clefted constituent not only identifies with the focus, but also presents Topic features.²⁵ Like Dufter (2009), I am keen to analyze such examples as contrastive Topics, “which arguably exhibit topic and focus properties” (Dufter 2009: 101; see also Krifka 1999; Hedberg 2013). A sentence containing a contrastive Topic should also present a secondary focus;²⁶ consequently, the cleft clause should also be partially or totally associated with focus as seems the case in (34) and (35).²⁷ Analogue cases (i.e., where we can observe more than one focus in the Cleft sentence) are provided in our corpus by the occurrence of the so-called multiple foci clefts (see Huber 2006: 569–570; Dufter 2009: 100; Roggia 2009: 113–115).²⁸ I found only four examples in the Italian section and one in the English section:
24 As Delin and Oberlander claim: “even a conjunction of ‘old’ elements presents a novel, second-order connection between the two” (Delin and Oberlander 2005 based on Delin 1989: 213). 25 The role of the clefted constituent as Topic can be suggested by its information status. However, topichood cannot always be equated with givenness (a Topic constituent can also be made up of new material, see Molnár 1998). 26 “Contrastive topics always occur in expressions that have another focus outside of the contrastive topic” (Krifka 2007: 47). Consider the following example: [My sister]Focus 1 / Topic studies [History]Focus 2. [My brother] Focus 1 / Topic studies [Civil engineering]Focus 2. 27 In example (34) Amanda is said “to be in the limelight”, while Raffaele Sollecito is “seldom mentioned” in newspapers. In (35) the contrast holds between people fit for work that are assumed “to cheat the system” and the system that “treats them badly”. 28 Although in my corpus all the multiple foci occurrences are represented by given or inferable clefted constituents (this is probably due to the fact that in multiple foci construction Focus and Topic can overlap, see Huber 2006: 570), in principle “this givenness might not be a necessary requirement” (Huber 2006: 570). Following this observation, it might be more accurate to introduce a further subtype of Type 1 for only multiple foci clefts. However, I think this move is unnecessary here.
Cleft sentences. Italian-English in contrast
(36)
115
Anche questo settembre, circa quattromila ragazzi italiani hanno iniziato l’anno accademico in un’università americana. In Italia la chiamano “fuga dei cervelli”, ma in inglese si dice “brain drain”, che letteralmente significa “perdita di cervelli” […] sono loro a scappare, non è l’Italia che perde capitale umano qualificato a causa di sistemi accademici e industriali poco competitivi […] (ilsole24ore.com) ‘Once again this September, about four thousand Italian young persons started the academic year in an American university. In Italy, this is called “brain escape”, but in English one says “brain drain”, which literally means “leaking of brains” […] it is they who leave, it is not Italy that loses qualified human resources because of a lack of competitive academic and industrial systems […]’
(37)
So what were the signs, and when did they occur? Was it she who revealed them, gradually; or was it I, as my senses and my body and my mind gradually came back to life, who noticed them, one by one, reassembling the parts into a whole? We were in a cafe one morning, and she visibly tensed up. (guardian.co.uk)
I will consider examples (34) to (37) as a subtype of Type 1 cleft (Type 1b), where the clefted constituent is both a given element and a focus. The information status of the cleft clause does not appear to be relevant but the subordinate must be associated wholly or partly with focus as well. Specifically, in my classification Type 1b includes occurrences with a given and focalized clefted constituent, instances of contrastive Topics, and multiple foci constructions. In the light of these observations, a more refined version of Table 1 can now be proposed (see also Korzen in this volume for a similar classification). The final result perfectly overlaps Dufter’s taxonomy (Dufter 2009: 100). The only difference is the inclusion of the dimension of referential givenness in my proposal:
116
Davide Garassino
Table 2: A more refined information-structural typology of Cleft sentences
Type 1a
Type 1b
Type 2
Type 3
Clefted Constituent
Cleft Clause
New
Given
Focus
Background
Given
Given or New
Focus1
Focus2
Given
New
Background
Focus
All-New All-Focus
In summary, referential givenness and focus / background seem to combine well. On the one hand, as already shown, referential givenness could not be used alone to describe in detail the different informational classes (i.e., it would fail to recognize the functional similarity between prototypical Type 1a clefts and the occurrences hosting a clefted constituent both given and associated with focus). On the other hand, considering information status in addition to focus / background permits a more accurate assessment of the functional differences concerning the role of the clefted constituent in Types 1a and 1b (e.g., in the latter the coexistence of both givenness and focal features often hints at the presence of a contrastive Topic). Moreover, referential givenness in addition to focus / background also suggests the expected contexts of use of certain cleft types (e.g., a Type 3 occurrence entirely made up of new information is typically encountered as a discourse starter). Based on this classification, I will provide the result of a quantitative analysis conducted on 200 occurrences (I classified 97 of 100 clefts in both Italian and English for a total of 194 examples):²⁹
29 The study of clefts in discourse is not an exact science. A number of examples required very careful analysis and in almost every case a second (or third) rethinking. I could not come to a decision concerning three examples in Italian and another three occurrences in English.
Cleft sentences. Italian-English in contrast
117
Table 3: The distribution of the different types of Italian clefts in the corpus Number of occurrences
%
Type 1a
26
27%
Type 1b
49
51%
Type 2
15
15%
Type 3
7
7%
Total
97
100%
Table 4: The distribution of the different types of English clefts in the corpus Number of occurrences
%
Type 1a
51
53%
Type 1b
29
30%
Type 2
11
11%
Type 3
6
6%
Total
97
100%
The data revealed several interesting asymmetries and similarities between the two languages. Starting with the differences, Type 1a clefts or prototypical clefts (focalizing devices with a new clefted constituent) are the most represented occurrences in English, whereas the use of Type 1b is less frequent although still relevant. In this respect, Italian data shows the reverse of the English situation: prototypical clefts are not as widespread as the “special” Type 1b, whose occurrences are almost twice as frequent as the Type 1a examples. In section 5, I will suggest some possible explanations concerning this quantitative difference. The distribution of Type 2 and Type 3 clefts, on the contrary, is quite similar in the two languages, although Type 2 clefts are slightly more common in Italian. The distributions shown in Tables 3 and 4 are not very different from the picture offered by other quantitative studies (see Roggia 2009: 142 for Italian; Collins 1991: 111; Gómez-González 2007: 124 for English). A direct, one-to-one comparison among these works is not an easy task because these classifications, although similar, do not overlap. However, by abstracting from the most idiosyncratic features of every analysis, it is possible to state that the larger diffusion
118
Davide Garassino
of Types 1b and 2 in Italian is supported by the findings in Roggia (2009: 142).³⁰ Gómez-González (2007), on the other hand, points out the large predominance of Type 1 clefts in English (i.e., Type 1a in my taxonomy), whereas in her study the distribution of Type 2 and 3 is quite different from the one observed in my corpus.³¹
4 Discourse functions of Italian and English Cleft sentences As already mentioned, the monofunctional view of clefts as focalizing devices was abandoned in the light of Prince’s works. In the previous section, I examined the information profiles of clefts; I will show how the differences in the information structure of clefts play a role in determining their discourse-pragmatic functions.
4.1 Type 1a and Type 1b According to my data, the main discourse functions carried out by Type 1a and 1b clefts are the highlighting of the clefted constituent and the expression of pragmatic contrast (Huber 2006: 559; Dufter 2009: 99). These functions are strictly dependent on the semantic properties of clefts as specificational copular structures and the role of the clefted constituent as a focus (see section 2). Hence, as I showed above (section 3), a value for the variable contained in the cleft clause is provided, and this value is picked from a set of relevant alternatives.
30 His three categories focus dato attivo ‘active given focus’, focus dato semiattivo / inattivo ‘semiactive / inactive given focus’, and inferibile ‘inferable’ are equivalent to Types 1b and 2 in my classification. According to Roggia (2009: 142), whose analysis is based on a corpus containing 530 occurrences, these clefts are the most widespread in Italian, representing cumulatively 66.7% of his data. 31 This fact largely depends on the different parameters used in Gómez-González (2007), which includes, for instance, multiple foci clefts within Type 3. However, in Gómez-González (2007) Type 1 clefts represent 51.65% of the data (the total number of her corpus occurrences amounts to 422 examples). In Collins (1991: 111), whose corpus contains 752 occurrences, Type 1 clefts are only slightly predominant (36%). In Gómez-González (2007), Type 2 represents 18.72% and Type 3 is equal to 29.62% of the data. In Collins (1991), Type 2 accounts for 34.6% and Type 3 for 29.4% of the data.
Cleft sentences. Italian-English in contrast
119
The highlighting of the clefted constituent is the result of focalization and refers to its pragmatic importance; the clefted constituent is, so to speak, put at the forefront of the message. However, highlighting does not require an explicit opposition within the text between the clefted constituent and other discourse referents (in other words, focus does not necessarily express an explicit opposition between the selected value and relevant alternatives). In example (38), for instance, there is no obvious contrast involving la forte pioggia ‘heavy rain’; the same observation applies to the struggle of indigenous Inuits and shrinking ice in example (39): (38)
Un muro di contenimento è crollato travolgendo due bambini, rispettivamente di 4 e 6 anni, che fortunatamente sono rimasti illesi. Le due piccole vittime sono state portate in ospedale, al momento non si conoscono le loro condizioni. È accaduto in serata a Somma Vesuviana, nel napoletano. Potrebbe essere stata la forte pioggia che si è abbattuta oggi pomeriggio sul napoletano ad avere provocato il cedimento di un terrapieno. (leggo.it) ‘A retaining wall collapsed, crushing two children, four and six years old, who were fortunately unharmed. The two little victims were brought to the hospital, but at the moment their conditions are unknown. It happened last evening at Somma Vesuviana, in the outskirts of Naples. It could have been the heavy rain, which fell this afternoon in the Naples area, which caused the embankment to collapse.’
(39)
Last year, Verheggen, a cultural ambassador for UNESCO, erected a huge sculpture on an iceberg off the coast of Greenland, an area he has visited annually for many years. It was the struggle of the indigenous Inuits to cope with extreme temperatures and shrinking ice that prompted thoughts of building an ice-making piece of art in the desert. “Let’s accept the climate is changing,” he says. “We have to see that as a challenge, to find new ways to deal with the changes in climate circumstances.” (nytimes.com/pages/aponline)
The label pragmatic contrast³² (see Huber 2006: 563) is used in this paper only for contexts in which the focalized referent is explicitly opposed to alternatives, such as in example (40) (a reduced form of [34]) and example (41):
32 Pragmatic contrast obviously involves highlighting of the clefted constituent as well. However, what I want to point out here is the functional difference between cleft occurrences, which
120
(40)
Davide Garassino
“Assolta”, “Una donna libera”, “Amanda liberata dalla corte”. Le testate statunitensi si inseguono, la Knox campeggia su tutte le prime pagine […] Su tutti i quotidiani, è Amanda a conquistare la scena: Sollecito è citato raramente, e sempre in secondo piano. (repubblica.it) ‘“Acquitted”, “A free woman”, “Amanda set free by the court”. American newspapers run similar headlines […] In all newspapers, it is Amanda that is in the limelight; Sollecito is seldom mentioned and [is] always in the background.’
(41)
I have felt for some time that there has been an element in all this of “don’t frighten the horses” and, personally, I think it underestimates – nay, insults – the intelligence of women. Screening is not like vaccination. We are not going to infect anybody else if we don’t go for breast screening. If a cancer is missed, it is an individual who suffers, not the population as a whole. (guardian.co.uk)
In these cases, Amanda is contrasted with Sollecito and, in the English example, an individual is opposed to the population as a whole. The following tables show the distribution of these functions in the data: Table 5: The discourse-functions of Type 1a and b clefts in Italian Mere highlighting only
Pragmatic contrast
Total
Type 1a
17 (65%)
9 (35%)
26 (100%)
Type 1b
23 (47%)
26 (53%)
49 (100%)
Table 6: The discourse-functions of Type 1a and b clefts in English Mere highlighting only
Pragmatic contrast
Total
Type 1a
24 (47%)
27 (53%)
51 (100%)
Type 1b
11 (38%)
18 (62%)
29 (100%)
present an explicit opposition among alternatives in the discourse (pragmatic contrast), and other cleft occurrences, which do not (mere highlighting only).
Cleft sentences. Italian-English in contrast
121
Whereas in Italian Type 1a is specialized for the mere highlighting of the clefted constituent and Type 1b preferably expresses pragmatic contrast³³, in English the contrastive function is more widespread in both types (but especially in Type 1b). At the level of Topic / Comment organization in the discourse (Ferrari et al. 2008: 58), Type 1a clefts introduce new referents (in the clefted constituent), which can be used as sentence Topics³⁴ in the subsequent text; e.g., Lewis Hamilton, who is then referred to as il pilota britannico ‘the British driver’, in (42) and Tom Watson in (43): (42)
È stato l’incostante ma sempre determinato Lewis Hamilton, al volante di una McLaren tornata in gran forma, a creare la sorpresa del giorno […] Il pilota britannico[Topic] si è mantenuto concentrato anche oggi e nell’ultimo giro ha piazzato un tempone […] (leggo.it) ‘It was a fickle but always determined Lewis Hamilton who, behind the wheel of his McLaren now back in shape, produced the surprise of the day […] The British driver[Topic] maintained his concentration today and set the fastest time on the final lap […]’
(43)
In many ways, and over two hours and 37 minutes, Murdoch acquitted himself coolly, even at moments when he could easily have slipped up. It was Tom Watson MP who produced the surprise of the day with new evidence from Neville Thurlbeck, the former NoW chief reporter. Watson[Topic] said that Thurlbeck told him he had been told by the former company lawyer Tom Crone that James Murdoch had in fact seen the critical “for Neville” email in 2008 – the evidence that showed phone hacking went beyond the actions of a single reporter. In a gripping moment, Watson[Topic] read out Thurlbeck’s words on the matter and invited the mogul elect to respond. (guardian.co.uk)
This function is known in the literature as Topic launching (Hasselgård 2004). Topic launching and contrast can also overlap (as they belong to different levels). It may well be the case in example (43), where Tom Watson is selected in contrast to Murdoch, who “acquitted himself coolly” but did not “produce any surprise of the day”. Consider now the following example: 33 In Italian Type 1b clefts, there is however no dramatic difference between the frequencies of the two functions. 34 I define a sentence Topic in terms of aboutness, following Reinhart (1981) and Lambrecht (1994): “A referent is interpreted as the topic of a proposition if in a given situation the proposition is construed as being about this referent” (Lambrecht 1994: 131).
122
(44)
Davide Garassino
Twins Mary-Kate[Topic 1] and Ashley Olsen[Topic 2] are still getting confused for each another, and now it’s having an effect on their dating lives. Reports circulated Wednesday that Ashley[Topic 2] got cozy with “Saturday Night Live” funnyman Jason Sudeikis at the recent premiere after-party for the flick “Tower Heist,” but a source confirmed to amNewYork that it was actually Mary Kate who chatted up with the comedian. Mary-Kate[Topic 1], 25, and Sudeikis, 36, were “super-flirty” at the bash, held at Stone Rose Lounge at the Time Warner Center, an eyewitness told us. (amny.com)
This case is similar to (42) and (43), but it is not exactly the same. The main difference is obviously that the Cleft sentence belongs to Type 1b, i.e., the clefted constituent is given. Therefore, the Cleft sentence in (44) launches as a Topic a referent previously introduced in the discourse (Mary-Kate Olsen) and that has already served as a sentence Topic, but was temporarily set aside in favor of another Topic (Ashley Olsen). I will thus refer to this function as Topic relaunching. Tables 7 and 8 show that both functions are not at all common in the two languages: Table 7: Total number of Cleft sentences carrying out Topic Launching and Relaunching functions in Italian Topic Launching
Topic Relaunching
Type 1a
2
0
Type 1b
0
3
Table 8: Total number of Cleft sentences carrying out Topic Launching and Relaunching functions in English Topic Launching
Topic Relaunching
Type 1a
4
0
Type 1b
0
2
Cleft sentences. Italian-English in contrast
123
4.2 Type 2 clefts The mapping of information structure onto the biclausal structure of Type 2 (i.e., given / new) suggests that their “natural” usage is to connect different parts of the discourse:³⁵ (45)
Schmidt è […] un ingegnere col pallino degli affari, che da Sun Microsystems è passato a Novell, dove ha potuto assaggiare la sconfitta al termine di un lungo braccio di ferro col gigante e concorrente Microsoft. E forse è anche per questo che non solo è stato chiamato a guidare Google, ma a far parte del consiglio di amministrazione di Apple nell’agosto del 2006. (repubblica.it) ‘Schmidt is […] an engineer who has a bent for business, moving from Sun Microsystems to Novell, where he could sense defeat at the end of a long arm wrestling with the rival giant Microsoft. It is perhaps also for this reason that he has been appointed not only to lead Google but also to be part of the Apple board of governors in August 2006.’
(46)
Angela Merkel said to David Cameron: either you allow us to go ahead with treaty change at 27 or others [Nicolas Sarkozy] will want a separate treaty with separate institutions for the 17. She said she did not want that but others did. Relations between Cameron and Merkel have improved since then after the prime minister indicated that Britain accepts the need for treaty change and will table relatively modest demands. The repatriation of social and employment laws will be for a later treaty negotiation. It is amid this background that José Manuel Barroso, the president of the European Commission, entered the fray on Wednesday with a major speech in Berlin. One line caused some irritation in London when Barroso had a pop at Britain: The speed of the European Union, and a fortiori of the euro area, cannot be the speed of its slowest member or its most reluctant member. […] The Anglophile former Portuguese prime minister, whose favourite English publication is the Spectator, said “a split union will not work”. Barroso believes that Britain should hug him and Merkel close because a view is developing in Brussels that France is attempting
35 As Hedberg (2013) rightly observes, although the information contained in the cleft clause is new, the conveyed presupposition is marked as a “known fact” (as noted by Prince 1978) and “it reminds more than inform” (Delin 1992). Therefore the new piece of information can be easily accommodated by the hearer.
124
Davide Garassino
to unravel two key British achievements over the last 20 years. (guardian. co.uk) In both examples, the clefted constituent ensures (because of its givenness) an anaphoric connection with the previous discourse, while the cleft clause introduces (because of its newness) new content and establishes a link with the following sentences. The generic cohesive function of Type 2 clefts, which is recognized in the literature (Berretta 1995, 2002; Roggia 2009; Dufter 2009: 103), acquires specific subfunctions depending on their position within the discourse. Consider again examples (45) and (46): in the former, the cleft links different text portions within the same paragraph, whereas in the latter the Cleft sentence provides a thematic connection between two distinct paragraphs. Understandably, at the end of a text / paragraph, the cohesive nature of clefts results in a “summative” use (Hasselgård 2004), i.e., a kind of conclusion and / or resume: (47)
Nel laboratorio intitolato a Enrico Fermi, che contribuì allo sviluppo dell’energia atomica all’università di Chicago, il futuro sembra in ogni caso denso di impegni. Nei prossimi mesi, infatti, i ricercatori cercheranno di verificare i risultati dell’esperimento Cngs 4 (Cern Neutrino to Gran Sasso), secondo il quale i neutrini sarebbero stati più veloci della luce di circa 60 nanosecondi. Il Fermilab è uno dei due laboratori al mondo ad avere le carte in regola per farlo. L’altro, in Giappone, ha subito rallentamenti a causa del terremoto e dello tsunami dello scorso marzo, per cui è da qui che si attendono i primi risultati. (repubblica.it) ‘In the lab named after Enrico Fermi, who contributed to the development of atomic energy at the University of Chicago, the future looks busy. In the following months, researchers will try to confirm the results of experiment Cngs 4 (Cern Neutrino to Gran Sasso), which found that neutrins are about 60 nanoseconds faster than light. Fermilab is one of two labs in the world that are fully equipped to do it. The other lab, in Japan, was hindered by the earthquake and tsunami of last March; therefore it is from here [i.e., Fermilab] that the first results are awaited.’
(48)
“It is with great sadness and a heavy heart that I have decided to end my six-year marriage to Ashton. As a woman, a mother and a wife, there are certain values and vows that I hold sacred, and it is in this spirit that I have chosen to move forward with my life.” (amny.com)
Cleft sentences. Italian-English in contrast
125
The following quantitative inquiry reveals that most Type 2 occurrences in the corpus are located within a paragraph: Table 9: Text position of Type 2 clefts in the Italian data (number of occurrences)
Type 2
Within a paragraph
Between two paragraphs
End of a paragraph / text
Total
9
3
3
15
Table 10: Text position of Type 2 clefts in the English data (number of occurrences)
Type 2
Within a paragraph
Between two paragraphs
End of a paragraph / text
Total
8
1
2
11
Furthermore, Type 2 can introduce in the cleft clause a referring expression that may be subsequently resumed as a sentence Topic. This discourse strategy (named Topic shifting in Hasselgård 2004) is used to lead to a Topic change.³⁶ The following is the only (clear) example found in the corpus (a reduced form of [46]): (49)
It is amid this background that José Manuel Barroso […] entered the fray on Wednesday with a major speech in Berlin. One line caused some irritation in London when Barroso [Topic] had a pop at Britain: The speed of the European Union, and a fortiori of the euro area, cannot be the speed of its slowest member or its most reluctant member […] The Anglophile former Portuguese prime minister[Topic], whose favourite English publication is the Spectator, said “a split union will not work”. Barroso[Topic] believes that Britain should hug him and Merkel close because a view is developing in Brussels that France is attempting to unravel two key British achievements over the last 20 years. (guardian.co.uk)
Interestingly, in principle, a similar function can also be fulfilled by Type 1b clefts. In (50), the cleft allows the journalist a successful transition from a previous Topic to a new one, i.e., from il Capo dello Stato Giorgio Napolitano, sua moglie Clio e il Presidente del Consiglio Silvio Berlusconi ‘the President Giorgio Napolitano, his wife Clio, and the Prime Minister Silvio Berlusconi’ to il Pontefice ‘the Pope’:
36 See Agar Marco (in this volume) for some considerations on Pseudo-cleft sentences and Topic shifting.
126
(50)
Davide Garassino
Napolitano e signora con Berlusconi per saluto italiano al Papa. Saranno il Capo dello Stato Giorgio Napolitano, sua moglie Clio e il Presidente del Consiglio Silvio Berlusconi ha [sic] rendere omaggio, con un breve saluto, a Benedetto XVI al termine della cerimonia, a nome della delegazione italiana. Dopo il rito di beatificazione e la recita del Regina Coeli il Pontefice[Topic] riceverà infatti i principali rappresentanti delle delegazioni ufficiali […] (repubblica.it) ‘Napolitano and his wife, with Berlusconi, will pay homage to the Pope on behalf of Italy. It will be the President Giorgio Napolitano, his wife Clio, and the Prime Minister Silvio Berlusconi who will briefly pay homage to Benedict XVI at the end of the ceremony in the name of the Italian delegation. After the beatification and the saying of the Regina Coeli, the Pope[Topic] will receive the spokesmen of the official delegations […]’
This functional contiguity should come as no surprise as Type 1b and Type 2 clefts are very similar in their information status (see Table 2). On the other hand, they also differ considerably in the repartition of focus / background within the sentence (the clefted constituent in Type 2 clefts is entirely backgrounded) and this fact determines the overall different distribution between the two groups.
4.3 Type 3 clefts Ideally, all-new (or all-focus) clefts introduce a discourse and are thus located at the beginning of a speech. They can be formulaic discourse starters, with clefted prepositional phrases, such as con piacere ‘with pleasure’, with honor, etc., in both Italian and English: (51)
In tribuna, la famiglia Monti al completo, la moglie Elsa, i figli Giovanni e Federica. In scuro e senza le consuete cravatte azzurro-celesti, Monti parte così: “È con grande emozione che mi rivolgo a voi”. (lastampa.it) ‘In the parliament gallery is the whole Monti family: his wife, Elsa, and their son and daughter, Giovanni and Federica. [Dressed] in a dark suit without his customary pale-blue tie, Monti begins: “It is with profound emotion that I address you.”’
(52)
“It is with great sadness and a heavy heart that I have decided to end my six-year marriage to Ashton. As a woman, a mother and a wife, there
Cleft sentences. Italian-English in contrast
127
are certain values and vows that I hold sacred, and it is in this spirit that I have chosen to move forward with my life.” (amny.com) Example (52) ([48]) is particularly noteworthy. In addition to the cleft at the end of the quoted speech (the Type 2 one that I have already discussed in section 4.2), this example contains a Type 3 occurrence right at the beginning. According to Hasselgård (2004), such clefts are used to give particular prominence to what the speaker experiences in a particular communicative situation. However, such description does not consider that these constructions, as well as Type 2 clefts such as (45) and (46), are probably undergoing a grammaticalization process that is turning them into textual connectors (see Berretta 2002)³⁷ or rhetorical devices used as discourse starters. Their evolving grammaticalized status is in fact accompanied by the loss of an alternative evoking function (see Roggia 2009: 137) and, formally, by the presence of formulaic prepositional phrases and / or adverbials in the clefted constituent. Type 3 includes other kinds of occurrences, such as Cleft sentences that provide a spatio-temporal setting. As Huber (2006: 566) points out, these clefts tend to appear in a discourse-initial position, such as formulaic ones. The clefted constituent usually hosts a temporal (or spatial) adverbial that serves as a “frame topic”, i.e., “it anchors the proposition within a temporal or local situation” (Huber 2006: 566). The Italian data also showed temporal clefts (section 3) such as in example (53), which are usually avoided in English³⁸ (see D’Achille, Proietti, and Viviani 2005: 258). Example (54) is the only occurrence found in the corpus: (53)
Oggi lo scienziato ha ritirato il premio Ig Nobel per la Biologia con il giusto umorismo: “Era molto che me l’aspettavo, perché ci avete messo così tanto?” (repubblica.it) ‘Today the scientist collected the Ig Nobel prize for Biology with an appropriate sense of humor: “I have been waiting for this for a long time [Lit. It has been a while that I have been waiting for this]. Why did it take you so long?”’
37 Type 3 clefts seem more constrained than Type 2 ones because of their information status (allnew), which in most cases favors their presence at the beginning of a text. 38 Instead of the Cleft sentence in (54), one would normally expect a monoclausal alternative (see D’Achille, Proietti, and Viviani 2005: 258) such as “we have had the women’s vote for only 40 years.”
128
(54)
Davide Garassino
“We will have to analyse it and see if we can do something. There are many things that have to do with society, where can we influence it,” said Verrey. “It’s only been 40 years that we have the women’s vote, it’s unbelievable but it is the way it is. In Switzerland it takes a very long time to do anything. It takes one step after another. And a big leap at the end.” (swissinfo.ch)
Finally, I chose to include within this information type the following examples, (55) and (56), that are by no means prototypical Type 3 clefts, but are entirely composed of new information. Because of their appearance (they report a quotation in the cleft clause and the author of the quoted material in the clefted constituent), they can be referred to as “quotation clefts”: (55)
Il pendolo oscilla tra individuo libero e individuo egoista: i nuovi mercati di Steve non sono né liberi né paritari. Sono giardini recintati, dove un solo doganiere decide cosa si vende, cosa non si vende, inventando una censura “locale” perfino sulle parolacce, e un diritto separato, che è la pretesa di tutti i Facebook e colossi di questo mondo. Era stato Orwell a scrivere “la libertà è schiavitù” e il pendolo di Steve, che era partito proprio con l’immagine di un colpo di martello sull’icona del totalitarismo orwelliano di Ibm (nel famoso commercial di Ridley Scott del 1984) è andato regolarmente al polo opposto alla libertà, economica in questo caso. (repubblica.it) ‘The pendulum sways between a free individual and an egotistic one; the new markets [created] by Steve are neither free nor equal. They are like fenced gardens where a sole custom officer decides what can be sold and what cannot be sold, thus creating “local” censorship even on swear words and distinct laws, which is the implicit ambition of Facebook and other [financial] giants. It was Orwell that wrote “freedom is slavery”; Steve’s pendulum, which in the beginning was identified with a hammer blow on IBM, the symbol of Orwellian totalitarianism (as conveyed in the famous 1984 ad by Ridley Scott), has moved constantly towards the opposite of liberty, which in this case is economic liberty.’
(56)
“From then on, the body was the site of a battle between germs and disease, and I think that’s something that’s still very vivid in our cultural memory – even though the idea has become more nuanced and these days we’re aware of things like ‘good bacteria’, and even though some scientists believe that we are cleaning our environs too harshly and that this is
Cleft sentences. Italian-English in contrast
129
leading to a rise in things like asthma, you still have all these ads on TV that talk of ‘waging war on dirt and germs’.” “But it’s a complex issue,” she adds. “It was the anthropologist Mary Douglas who said: ‘There is no such thing as absolute dirt. It exists in the eye of the beholder.’ ” (guardian.co.uk) For the sake of simplicity, Type 3 tokens can be split between two classes: formulaic and non-formulaic (i.e., frame Topics and quotation clefts) occurrences. Tables 11 and 12 below show the result of a quantitative inquiry: Table 11: Formulaic and non-formulaic Type 3 cleft sentences in Italian (number of occurrences)
Type 3
Formulaic
Non-formulaic
Total
1
6
7
Table 12: Formulaic and non-formulaic Type 3 cleft sentences in English (number of occurrences)
Type 3
Formulaic
Non-formulaic
Total
3
3
6
The few available examples notwithstanding, we can observe that (at least in the corpus) non-formulaic occurrences seem preferable to formulaic ones in Italian, while English shows a perfect symmetry between the two groups. Type 3 clefts are usually claimed to be typical of English (Dufter 2009: 106; Hasselgård 2004), but my data might suggest that in fact only formulaic Type 3 clefts seem more widespread in English than in Italian.
5 Specific functions within journalistic texts I have examined so far the most typical discourse-pragmatic properties of Cleft sentences in both Italian and English as well as the connection between these functions and the different information structure properties exhibited by clefts. Because our corpus is made of journalistic texts, it would be interesting to determine whether more specific functions can be fulfilled within this text typology. The corpus analysis suggests the existence of specific uses of Cleft sentences, and it also hints that they are not new functions at all, but instead local, text-
130
Davide Garassino
dependent instantiations of the more general functions already examined (e.g., discourse-cohesive uses). As the present inquiry makes no pretence of being exhaustive, the remainder of this section will be largely descriptive.³⁹ The most striking function is perhaps the “connection” or “link” that a Cleft sentence can build between different textual units, i.e., headlines, subheads, and the body of the article. Consider the following examples: (57)
Steve Jobs non ce l’ha fatta addio al fondatore della Apple Uno scarno comunicato dell’azienda rivela al mondo la scomparsa di uno dei simboli dell’èra digitale. Poche settimane fa aveva dovuto rinunciare ad ogni incarico nel gruppo che aveva creato […] SAN FRANCISCO – È uno scarno comunicato della sua azienda, la Apple di Cupertino, che dà la notizia attraverso l’Associated Press […] (repubblica.it) ‘Steve Jobs did not make it. Farewell to the founder of Apple. A brief public announcement by his company discloses the death of one of the symbols of the digital era. A few weeks ago he had to renounce his appointment in the company he created […] San Francisco – It was a brief public announcement by his company, Apple [headquarters] in Cupertino, that disseminated the news through the Associated Press.’
(58)
Red Cross opens new logistics centre A centre to house stocks for the humanitarian missions of the International Committee of the Red Cross (ICRC) was opened in Satigny near Geneva on Wednesday. “It’s here that a large part of the medicines, orthopedic equipment, water treatment equipment and the other material to be sent to help victims of conflicts will be received, stored and then dispatched.” (swissinfo.ch)
In examples (57) and (58), the content of the clefted constituent and the cleft clause is at first glance given (it has already been introduced in the headline or,
39 For other analyses of the functions exhibited by clefts in journalistic texts, see Banks (1999) and Gómez-González (2007) for English and Bonomi (2002) for Italian.
Cleft sentences. Italian-English in contrast
131
in the Italian example, the summary. Both the headline and the summary are written in small caps in the examples). Furthermore, in this case we refer to different textual units, i.e., units that have a different status in the discourse and are in principle independent from each other. However, this is true only in principle, as the anaphoric use of the adverbial in (58) points out that the reader cannot possibly reconstruct the reference for here without knowing the headline and the summary content. Moreover, a closer look reveals that these two Cleft sentences are not entirely given: new information is in fact presented. In (57), the content of the cleft is more detailed than the information presented in the summary; in (58), a Type 2 cleft, the content of the cleft clause is almost completely new. These clefts fulfill a metatextual function to the extent that (a) they ensure thematic continuity between different levels of the text that in principle can be read separately and (b) they seem to elaborate on the main discourse Topic(s) already presented in higher textual units. This function is very rare in the English corpus (the only example is [58] from the Swiss agency ATS), but it is quite common in Italian, as examples (59) and (60) show: (59)
STIPENDI DEI PARLAMENTARI, DECIDERANNO LE CAMERE. FINI: “MODIFICHE A GENNAIO” […] ROMA – Saranno le Camere a provvedere al taglio delle indennità di deputati e senatori e non un decreto del governo come prevedeva la manovra. (leggo.it) ‘Parliamentary Salaries, the chambers will decide. Fini: “Changes in January” […] Rome- It will be the Chambers that will cut the salaries of representatives and senators and not a decree, as previously planned in the financial measures.’
(60)
Russia, partito oligarca arruola popstar Primo “colpo” per Causa Giusta del magnate in vista elezioni […] MOSCA, 6 SET – È l’oligarca Mikhail Prokhorov a fare il primo colpaccio della campagna elettorale per le legislative del 4 dicembre. (ansa.it)
132
Davide Garassino
‘Russia, oligarch party enlists a pop star. First “hit” of the tycoon for Right Cause in view of the elections. […] Moscow, September 6 – It is the oligarch Mikhail Prokhorov who first hit the jackpot during the campaign for the legislative elections on December 4.’ The clefts in (59) and (60) are probably best analyzed as Type 1b occurrences: the clefted constituents, le Camere, ‘the Parliament’, and the oligarch Mikhail Prokhorov, respectively, are given referents (already introduced in the title or the summary) but also appear to be associated with focus. The different distributions of Italian and English in this respect require further observation; strictly speaking, this divergence does not seem to be caused by language-internal factors, but by the different construction and organization of journalistic articles in the two languages. Italian newspapers in fact present a higher degree of discourse cohesion between different subparts of the text, whereas in English these units seem in principle more independent from one another. In this light, the strategy of Italian journalists is more oriented towards a repetition / elaboration of the main discourse Topic(s), even at the cost of a certain redundancy. Another function observed in the corpus is the ability of clefts to link quoted speech and the journalist’s own writing, as the following examples show: (61)
“Non ho paura di niente” e “non mi pento di niente”. È con queste parole che Cesare Battisti apre un’intervista a uno dei più diffusi settimanali brasiliani, Istoè (repubblica.it) ‘“I am not afraid of anything” and “I do not regret anything.” It is with these words that Cesare Battisti starts an interview with one of the most widespread Brazilian weekly publications, the Istoè.’
(62)
Marie, 23, doesn’t offer any comfortable excuses for how she ended up in such a dark place. Growing up in a three-bedroom house in Leicester with two older brothers, she was a loved child. “My dad left when I was one, but my mum did the best she could. She always put us first.” It was at school, rather than at home, that Marie had problems. The bullying started in primary school, but got worse when she started secondary school. “It went on every day, in and out of school, physical and verbal.” (guardian.co.uk)
Cleft sentences. Italian-English in contrast
133
In such occurrences, the Type 2 occurrence in (61) refers back to the words used by Cesare Battisti in the quoted speech (“‘non ho paura di niente’ e ‘non mi pento di niente’”, ‘“I am not afraid of anything” and “I do not regret anything”’) and introduces new content in the cleft clause (“che Cesare Battisti apre un’intervista…”, ‘that Cesare Battisti starts an interview…’). In example (62), there are two different discourse levels: the story told by the interviewee in the quoted speech and the commentaries made by the journalist in between the different pieces of the interview. The cleft helps to construct the transition from one sequence of the story to another and allows the journalist to make some introductory statements about what is going to come next. Similar examples are (63) and (64): in (63) we observe a mix of the journalist’s own writing and the quoted speech within the cleft, while in (64) the use of the Cleft sentence is a way for the journalist to comment on the claims of the media and democracy analyst in the quoted speech above:⁴⁰ (63)
Speroni (parlando a Radio 24) se la prende proprio con il sindaco di Verona che, ammette “ha avuto un grandissimo consenso popolare”, ma “alcune sue posizioni sono al di fuor della linea del partito come ad esempio quando appoggia certe posizioni centraliste”. Quanto al congresso provinciale di Varese per Speroni è stata “la persona a non essere gradita, non la linea politica della Lega […]” (ilsole24ore.com) ‘Speroni (speaking on Radio 24) gets angry with the mayor of Verona himself who, he concedes, “has gained tremendous public approval”, but “some of his positions lie outside the party policy as when, for instance, he endorses certain centralistic stances”. Regarding the provincial convention in Varese, according to Speroni, “it is the person that is not appreciated, not the policy of the Lega [Nord party] […]”’
(64)
Other analysts expressed fears the result would dent the credibility of Sirleaf’s government. “The government will need to strengthen the belief and participation in democracy to win back the people,” said Abdullai Kamara, a media and democracy analyst based in Monrovia. “There was no competitor so it didn’t give people any impetus to go out and vote.”
40 This example is also interesting because it allows some considerations about the rhetorical effects achieved by clefts in discourse. The occurrence in (64) introduces a relevant alternative (fear besides the one-party race) in order to satisfy the open proposition expressed by the cleft clause; the use of a negative cleft and a truncated (i.e., without the cleft clause, see Hedberg 2000) cleft containing also suggests that the expert’s analysis is incomplete and indirectly implies the “expert” status of the journalist in that particular field.
134
Davide Garassino
It was not only the “one party” race that failed to entice people to the polls, it was also fear. The day before the vote, the CDC organised a protest march intended to be peaceful at its headquarters, but it ended in bloodshed. (guardian.co.uk)
6 Conclusion This survey has revealed the distribution and the functions of different types of clefts in both Italian and English. These types were motivated by their different information structure properties: in particular, the different information statuses displayed by the clefted constituent and the cleft clause as well as the repartition of focus / background in the sentence. As a result, several classes were considered: Type 1a (whose clefted constituent is new and focalized while the cleft clause is given and backgrounded); Type 1b (characterized by a given, but focalized clefted constituent. This type also includes instances of contrastive Topics and multiple foci constructions); Type 2 (which is the reverse image of Type 1a, with a fully backgrounded clefted constituent); and finally a Type 3 (reserved for all-new and all-focus cleft occurrences). The following are some final observations about specific divergences and similarities between Italian and English. First, I noted an overall functional overlap between the two languages: the information types and their discourse-pragmatic properties are basically the same. However, the distribution of these features is sometimes quite different. For instance, in Italian the most common cleft type is Type 1b, whereas in English Type 1a is the most widespread. The distribution of Type 2 and Type 3 in both languages is (more or less) symmetrical, but it is also more marginal in terms of frequency. Type 2 is, however, slightly more common in Italian. Type 3 needs further explanation because the typology of the clefts I chose to include under this label is quite heterogeneous. Formulaic clefts seem typical of English, which is in line with the existing literature (Hasselgård 2004; Dufter 2009), whereas clefts presenting a “frame Topic” occur more frequently in Italian. As far as the discourse-pragmatic functions are concerned, Italian uses Type 1a clefts to highlight the clefted constituent more than to stress an explicit contrast. On the contrary, Type 1b is preferably employed to signal an explicit opposition. In English, the explicit contrastive function is more frequent in both Type 1a and Type 1b. Type 2 clefts perform cohesive and metatextual functions in both Italian and English. Type 3 examples are specialized as rhetorical devices for starting a discourse in English (formulaic clefts), while in Italian they commonly provide a temporal setting, mostly at the beginning of a text.
Cleft sentences. Italian-English in contrast
135
In summarizing the divergences and similarities between Italian and English, it is also necessary to reiterate that the goal of this paper is not to pursue an “absolute” comparison, but a contrastive analysis based on a very specific textual corpus. Put differently, my analysis also had to deal with language-external factors. In particular, the source of the data and the specific uses of clefts in journalistic texts suggest that how an article is structured in different journalistic traditions does matter and does affect the distribution of linguistic structures. More specifically, Italian articles, in which the main discourse Topics are usually expressed and elaborated in different textual units, seem to promote a more cohesive use of clefts in the discourse. Consequently, Italian newspapers show a tendency to repeat given information, thus favoring a more consistent presence of Type 1b and Type 2 clefts, whose information structure is fully compatible with such strategies of text construction.
References Agar Marco, Rocío. This vol. Pseudo-cleft sentences. Italian-Spanish in contrast. Akmajian, Adrian. 1970. On deriving cleft sentences from pseudo-cleft sentences. Linguistic Inquiry 1. 147–168. Banks, David. 1999. Decoding the information structure of journalistic clefts. Interface, Journal of Applied Linguistics 14 (1). 3–24. Berretta, Monica. 1995. Ordini marcati dei costituenti maggiori di frase: una rassegna. Linguistica e Filologia 1. 125–170. Berretta, Monica 1996. Come inseriamo elementi nuovi nel discorso/3. Che mi fa paura è la nebbia. Italiano e Oltre 2. 116–122. Berretta, Monica. 2002. Quello che voglio dire è che: le scisse da strutture topicalizzanti a connettivi testuali. In Gianluigi Beccaria & Carla Marello (eds.), Dalla parola al testo. Scritti per Bice Mortara Garavelli, 15–31. Alessandria: Edizioni dell’Orso. Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad & Edward Finegan. 1999. Longman Grammar of Spoken and Written English. Harlow, Essex: Pearson Education Ltd. Bonomi, Ilaria. 2002. L’italiano giornalistico dall’inizio del ’900 ai quotidiani on line. Firenze: Franco Cesati. Charnavel, Isabelle. 2011. On French un même and antispecificity. In Ingo Reich, Eva Orch & Dennis Pauly (eds.), Sinn und Bedeutung 15. Proceedings of the 2010 Annual Conference of the Gesellschaft für Semantik (Universität des Saarlandes, 9–11 September 2010), 133–147. Saarbrücken: Saarland University Press. Chomsky, Noam. 1977. On wh-movement. In Peter W. Culicover, Thomas Wasow & Adrian Akmajian (eds.), Formal Syntax, 77–132. New York: Academic Press. Collins, Peter C. 1991. Cleft and Pseudo-cleft constructions in English. London: Routledge. D’Achille, Paolo, Domenico Proietti & Andrea Viviani. 2005. La frase scissa in italiano: aspetti e problemi. In Paolo D’Achille & Iørn Korzen (eds.), Tipologia linguistica e società. Due
136
Davide Garassino
giornate italo-danesi di studi linguistici (Roma, 27–28 novembre 2003), 249–279. Firenze: Cesati. De Cesare, Anna-Maria. 2012. Riflessioni sulla diffusione delle costruzioni scisse nell’italiano giornalistico odierno a partire dalla loro manifestazione nei lanci di agenzia in italiano e in inglese. Cuadernos de Filología Italiana 19. 11–39. De Cesare, Anna-Maria. This vol. Cleft constructions in a cross-linguistic perspective. Towards an operational taxonomy. De Cesare, Anna-Maria, Davide Garassino, Rocío Agar Marco & Laura Baranzini. This vol. Form and frequency of Italian cleft constructions in a corpus of electronic news. A comparative perspective with French, Spanish, German and English. De Cesare, Anna-Maria & Davide Garassino. In press. On the status of exhaustiveness in Cleft sentences: A data-driven and cross-linguistic study of English also- / only-clefts and Italian anche- / solo-clefts. Folia Linguistica. Declerck, Renaat. 1988. Studies on Copular Sentences, Clefts and Pseudo-clefts. Berlin & New York: Mouton de Gruyter. Delin, Judy. 1989. Cleft Constructions in Discourse. Edinburgh: University of Edinburgh dissertation. Delin, Judy. 1992. Properties of it-cleft presupposition. Journal of Semantics 9. 179–196. Delin, Judy & Jon R. Oberlander. 2005. Clefts constructions in context. Some suggestions for research methodology. Ms., University of Stirling. http://www.fb10.uni-bremen.de/ anglistik/langpro/projects/GeM/delin-publications.html (accessed 22 June 2013) Den Dikken, Marcel. 2009. Predication and specification in the syntax of cleft sentences, Ms. http://www.gc.cuny.edu/CUNY_GC/media/CUNY-Graduate-Center/PDF/Programs/ Linguistics/Dikken/predication_specification_Clefts_short_paper.pdf (accessed 22 January 2013) Dufter, Andreas. 2009. Clefting and discourse organization: Comparing Germanic and Romance. In Andreas Dufter & Daniel Jacob (eds.), Focus and Background in Romance Languages, 83–121. Amsterdam & Philadelphia: John Benjamins. É. Kiss, Katalin. 1998. Identificational focus versus informational focus. Language 74 (2). 245–273. Ferrari, Angela, Luca Cignetti, Anna-Maria De Cesare, Letizia Lala, Magda Mandelli, Claudia Ricci & Carlo Enrico Roggia. 2008. L’interfaccia lingua-testo. Natura e funzioni dell’articolazione informativa dell’enunciato. Alessandria: Edizioni dell’Orso. Frison, Lorenza 1988. L’ordine degli elementi della frase e le costruzioni marcate. In Lorenzo Renzi & Giampaolo Salvi (eds.), Grande grammatica italiana di consultazione, vol. 1, 194–225. Bologna: il Mulino. Gómez González, Maria de los Ángeles. 2007. ‘It was you that told me that, wasn’t it?’ It-clefts revisited in discourse. In Mike Hannay & Gerard J. Steen (eds.), Structural-functional Studies in English Grammar, 103–140. Amsterdam & Philadelphia: John Benjamins. Gundel, Jeanette K. 1988. Universals of topic-comment structure. In Michael Hammond, Edith A. Moravcsik & Jessica Wirth (eds.), Studies in Syntactic Typology, 209–239. Amsterdam & Philadelphia: John Benjamins. Gundel, Jeanette K. 2006. Clefts in English and Norwegian: Some implications for the grammarpragmatics interface. In Valéria Molnár & Susanne Winkler (eds.), The Architecture of Focus, 517–548. Berlin & New York: Mouton de Gruyter. Gundel, Jeanette K. 2008. Contrastive perspectives on cleft sentences. In Maria de los Ángeles Gómez González, John Lachlan Mackenzie & Elsa M. González Álvarez (eds.), Languages
Cleft sentences. Italian-English in contrast
137
and Cultures in Contrast and Comparison, 69–87. Amsterdam & Philadelphia: John Benjamins. Hasselgård, Hilde. 2004. Adverbials in it-cleft constructions. In Karin Aijmer & Bengt Altenberg (eds.), Advances in Corpus Linguistics, 195–212. Amsterdam & New York: Rodopi. Hedberg, Nancy. 2000. The referential status of clefts. Language 76 (4). 891–920. Hedberg, Nancy. 2013. Multiple focus and cleft sentences. In Andreas Haida, Katharina Hartmann & Tonies Veenstra (eds.), The Structure of Clefts, 227–250. Amsterdam & Philadelphia: John Benjamins. Hedberg, Nancy & Lorna Fadden. 2007. The information structure of it-clefts, wh-clefts and reverse wh-clefts in English. In Nancy Hedberg & Robert Zacharski (eds.), The Grammarpragmatics Interface: Essays in honor of Jeanette K. Gundel, 49–76. Amsterdam & Philadelphia: John Benjamins. Higgins, Francis R. 1973. The Pseudo-cleft construction in English, Cambridge, Ma: MIT dissertation. Huber, Stefan. 2006. The complex function of it-clefts. In Valéria Molnár & Susanne Winkler (eds.), The Architecture of Focus, 549–578. Berlin & New York: Mouton de Gruyter. Huddleston, Rodney D. & Geoffrey K. Pullum (eds.). 2002. The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press. Jespersen, Otto. 1954. A Modern English Grammar on Historical Principles. London: Allen & Unwin. Johansson, Mats. 2002. Clefts in English and Swedish: A Contrastive Study of it-clefts and wh-clefts in Original Texts and Translations. Lund: University of Lund dissertation. Korzen, Iørn. This vol. Cleft Sentences. Italian-Danish in contrast. Krifka, Manfred. 1999. Additive particles under stress. In Devon Strolovitch & Aaron Lawson (eds.), Proceedings of the 8th Semantics and Linguistic Theory Conference (MIT, May 8–10, 1998), 111–128. Cornell: CLC Publications. Krifka, Manfred. 2007. Basic notions of information structure. In Caroline Féry, Gilbert Fanselow & Manfred Krifka (eds.), Interdisciplinary Studies on Information Structure 6, 13–55. Potsdam: Universitätsverlag. Lambrecht, Knud. 1994. Information Structure and Sentence Form. Topic, Focus, and the Mental Representation of Discourse Referents. Cambridge: Cambridge University Press. Lambrecht, Knud. 2001. A framework for the analysis of cleft constructions. Linguistics 39 (3). 463–516. Metzeltin, Michael. 1989. La scissione relativa in italiano e nelle altre lingue romanze. In Fabio Foresti, Elena Rizzi & Paola Benedini (eds.), L’italiano fra le lingue romanze. Atti del XX Congresso SLI (Bologna, 25–27 settembre 1986), 151–169. Roma: Bulzoni. Metzeltin, Michael. 2010. Erklärende Grammatik der romanischen Sprachen: Satzkonstruktion und Satzinterpretation. Wien: Praesens Verlag. Molnár, Valéria. 1998. Topic in focus. On the syntax, phonology, semantics and pragmatics of the so-called ‘contrastive topic’ in Hungarian and German. Acta Linguistica Hungarica 45. 89–166. Panunzi, Alessandro. 2011. Frasi scisse. In Raffaele Simone, Gaetano Berruto & Paolo D’Achille (eds.), Enciclopedia dell’italiano, 1284–1287. Roma: Istituto dell’Enciclopedia Italiana Giovanni Treccani. Patten, Amanda L. 2012. The English it-cleft: A Constructional Approach and a Diachronic Investigation. Berlin & New York: Mouton de Gruyter.
138
Davide Garassino
Percus, Orin. 1997. Prying open the cleft. In Kiyomi Kusumoto (ed.), Proceedings of the NorthEastern Linguistic Society 27 (McGill University, October 18–20, 1996), 337–351. Amherst, Mass.: GLSA Publications. Prince, Ellen F. 1978. A comparison of wh-clefts and it-clefts in discourse. In Language 54. 883–906. Prince, Ellen F. 1981. Toward a taxonomy of given-new information. In Peter Cole (ed.), Radical Pragmatics, 223–256. New York: Academic Press. Reeve, Matthew. 2011. The syntactic structure of English clefts. Lingua 121 (2). 142–171. Reinhart, Tanya. 1981. Pragmatics and linguistics: An analysis of sentence topics. Philosophica 27 (1). 53–94. Roggia, Carlo Enrico. 2009. Le frasi scisse in italiano. Struttura informativa e funzioni discorsive. Genève: Slatkine. Salvi, Giampaolo. 1991. Le frasi copulative. In Lorenzo Renzi & Giampaolo Salvi (eds.), Grande grammatica italiana di consultazione, vol. 2, 163–189. Bologna: il Mulino. Salvi, Giampaolo & Laura Vanelli. 2004. Nuova grammatica italiana. Bologna: il Mulino. Sornicola, Rosanna. 1988. It-clefts and wh-clefts: Two awkward sentence types. Journal of Linguistics 24 (2). 343–379. Ward, Gregory & Betty J. Birner. 2004. Information structure and non-canonical syntax. In Laurence R. Horn & Gregory Ward (eds.), Handbook of Pragmatics, 153–174. Oxford: Blackwell. Wedgwood, Dan, Gergely Pethő & Ronnie Cann. 2006. Hungarian ‘focus position’ and English it-clefts: The semantic underspecification of ‘focus’ readings, Ms. http://www.lel.ed.ac.uk/~dan/web_HungFP_wpc.pdf (accessed 22 January 2013). Zimmermann, Malte & Edgar Onea. 2011. Focus marking and focus interpretation. Lingua 121. 1651–1670.
Laura Baranzini*
Pseudo-cleft sentences. Italian-French in contrast 1 Introduction This paper presents the results of a corpus-based analysis of Italian and French Pseudo-cleft sentences (henceforth, PC sentences) in a contrastive perspective. The empirical basis for our discussion was provided by the ICOCP (Italian Constituent Order in a Contrastive Perspective) reference corpus, built on the framework of the Swiss National Science Foundation project of the same name. ICOCP is a multilingual comparable corpus of online newspaper texts of different types and sizes, ranging from press releases to larger, in-depth articles. For the present work we have taken into account the Italian and French subsections which comprise 520,000 words and 420,000 words respectively. We collected all the texts constituting the corpus within the same short period (the last trimester 2011) from sources in different languages, using equivalent criteria (same category of daily newspaper and press agencies, comparable types of sections and topics). The study has been conducted at different levels of analysis. After a brief theoretical discussion on the definition of the object of study (Section 2), we will present the formal types of PC sentences in the corpus (Section 3), their information structure (Section 4), and, finally, their textual functions, as well as some of the ways in which they can be exploited on an argumentative level (Section 5). For each level of analysis comparable data will be provided, and, where relevant, the differences between French and Italian will be described and analyzed.¹
* I would like to thank Claudia Ricci for her help in translating this paper. 1 A first version of this research appears in Baranzini (2014). Changes in data calculation and analysis reflect the evolution and refinement of our work on the subject. Laura Baranzini, University of Basle, University of Neuchâtel
140
Laura Baranzini
2 The Pseudo-cleft sentence: The problem with its definition Within the macro-category of cleft constructions,² the following can be identified: Cleft sentences in the strict sense (see [1] below), Reverse clefts (2), PC sentences (3) and Reverse PC sentences.³ Among these, only PC sentences will be examined here. Although Cleft sentences have been the object of in-depth study in various languages for many decades, less research has been conducted on PC sentences (among the older works we should, however, mention Akmajian 1970), partly because they occur relatively seldom compared with cleft constructions. The term Pseudo-cleft indicates a derivation from Cleft sentences (but see Akmajian 1970), to which PC sentences have been related on the basis of similarities between the two constructions on a syntactic-semantic level. In other cases, the interest in Pseudo-clefts can be ascribed to their similarities with other constructions, such as copulative sentences; in Salvi (1991), for instance, PC sentences are analyzed in the chapter dealing with copulative sentences. PC constructions are identified by Salvi as a subset of copulative sentences on account of their semantic features, the only additional formal feature being the presence of a relative clause in the first part of the structure. As briefly mentioned above, the PC sentence was initially described as being linearly similar to a Cleft (1), or better still, to a Reverse cleft (2),⁴ based on the presence of the verb essere ‘to be’ as well as on a common information function, namely a generically “focalizing” function: (1)
È Mario che ci viene a trovare spesso. IsCOP Mario that-REL us comes to visit often ‘It is Mario that comes to visit us often.’⁵
2 For a description of the relations between different cleft constructions, and for a new taxonomy proposal, see De Cesare (in this volume); according to this new taxonomy, the sentences under consideration here should be classified within type C (constructions with final cleft constituent), though they would not be the only possible realization of this type. In view of the perspective taken for this analysis, whose focus is on a subtype of type C constructions, we have chosen for now to keep using the traditional terminology. 3 È Mario chi ci viene a trovare spesso ‘It is Mario who comes to visit us often’. 4 In the typology proposed by De Cesare (in this volume), Pseudocleft and Reverse cleft sentences together should form type C precisely because of their similarity in the position of the cleft constituent. 5 The translations of the examples are provided to help understanding the phenomena under study and are therefore intentionally literal.
Pseudo-cleft sentences. Italian-French in contrast
(2)
?
Che ci viene a trovare spesso often That-REL us comes to visit ‘That comes to visit us often is Mario.’
è Mario.⁶ is-COP Mario
(3)
Chi ci viene a trovare spesso Who-REL us comes to visit often ‘Who comes to visit us often is Mario.’
è Mario. is-COP Mario
141
In the literature, the prototypical example of a Pseudo-cleft sentence is formally equivalent to (3) above; it is introduced by a fused relative pronoun and has a post-copular element constituted by a noun phrase (usually a simple noun): x (pre-copular constituent) è (copula) y (post-copular constituent or cleft constituent)
Furthermore, this type of structure (in its variant forms) is associated with a particular semantic movement, referred to as specificational⁷ (or, in other studies, identificational), which will be discussed in more detail in Section 4. For the purpose of identifying the object of study, we retain for now the double nature of the definition of the PC, according to which this construction is to be viewed as a subtype of copulative sentence on the basis of formal as well as semantic criteria (the presence of a relative clause and a specificational movement respectively). The “prototypical” PC⁸ is introduced by a fused relative pronoun. This necessarily involves the presence of a relative subordinate clause in the pre-copular constituent; furthermore, considering the substantial semantic equivalence between fused relative pronouns and some sequences of the type ‘demonstrative pronoun + relative pronoun’ (for instance, chi ‘who’ → quello che ‘that who’), the PC traditionally also includes all structures characterized by the presence in the pre-copular element of a relative subordinate clause headed by a demonstrative pronoun such as ciò, quello, ‘it, this, that’ etc. (ce, celui, etc. in French). Such linguistic realizations – with purely pronominal introducers – are the only possible ones as far as a strict definition of PC sentences is concerned (see Roggia 2009: 36–37). It has been suggested, notably in Collins (1991), that English should also include generic (full) noun phrases which can head the relative clause while preserving the overall meaning of the construction. Collins proposes a list of possible
6 Compared with the other examples, (2) is marked as being more typical of spoken discourse. 7 Roubaud (2000) describes PC sentences in terms of “spécification progressive” (‘progressive specification’). 8 A formal description can be found in Panunzi (2009); for an analysis in terms of dispositif (‘device’) see Roubaud (2000).
142
Laura Baranzini
generic nominal heads obtained by assigning pronouns a corresponding noun (a similar kind of correspondence in Italian would be chi → la persona che ‘who → the person who’, ciò che → la cosa che ‘what → the thing that’, dove → il luogo in cui ‘where → the place in which’, etc.); such inclusion of nominal heads can, on a theoretical level, easily be applied to increasingly less generic nouns and, ultimately, to any type of noun modified by a relative subordinate clause, as illustrated in the examples proposed by Salvi (1991).⁹ As we have seen, the grammatical category (and some semantic features) of the elements introducing a PC sentence is not always agreed upon, and its treatment in the literature is often uneven. The presence of the pre-copular relative subordinate clause, considered a defining feature of the construction, may also be problematic on a theoretical level. Indeed, the question arises whether it is appropriate to draw a definitive boundary between types of modification of the pre-copular head, thus ruling out, for instance, other types of subordinate clauses or phrase expansions through adjectives or participles. The following examples show how the different degrees of complexity and types of expansion of a phrase can code very similar semantic content within a highly comparable morpho-syntactic structure: (4)
La soluzione che potrebbe risolvere i nostri problemi The solution that-REL could solve the our problems è il licenziamento. is the dismissal ‘The solution that could solve our problems is dismissal.’
(5)
La soluzione per risolvere i nostri problemi è il The solution to solve the our problems is the licenziamento. dismissal ‘The solution to solve our problems is dismissal.’
(6)
La soluzione ai nostri problemi è il licenziamento. The solution to our problems is the dismissal ‘The solution to our problems is dismissal.’
We shall not further discuss this definition issue on a theoretical level. As we have seen, the formal aspect of the definition of a PC sentence is ultimately based
9 For a more detailed analysis of this aspect of the definition see Agar Marco (in this volume).
Pseudo-cleft sentences. Italian-French in contrast
143
upon a theoretical a priori choice, which is justified on a case-by-case basis but is not exclusive. Such a choice sets an arbitrary boundary on the continuum of the (more or less complex) morpho-syntactic realization of the pre-copular element of a copulative specificational clause.¹⁰ The possible realizations range from nongeneric full noun phrases to fused relative pronouns, including generic full noun phrases and relative pronouns with an antecedent. Our choice is mainly based on the semantic nature of this structure (see Section 4). On a formal level, it is among the least restrictive descriptions, as it includes pronominal heads of various kinds, and also generic as well as nongeneric nominal heads. Conversely, the restriction on the type of phrase expansion has been maintained in our definition: we only take into account expansions in the form of a restrictive relative clause. Furthermore, for reasons of convenience, we have chosen to take into consideration only PC sentences with the verb essere ‘to be’ as a copula, thus excluding other eventual formulations (for example, sembrare / parere / dovere – ‘seem / appear / must’ etc. + essere). This should have no impact on either qualitative or quantitative observations, as copulative verbs other than essere are very rare in the texts taken into account here and can be assumed to function in an equivalent way.
3 Formal types of Pseudo-clefts in our corpus: Quantitative data On the basis of the aforementioned criteria,¹¹ the number of PC sentences found in our corpus is 77 for the Italian section and 46 for the French section, an average of 14.8 occurrences per 100,000 words in Italian and 11.5 per 100,000 words in French. The first relevant quantitative data therefore relate to the frequency of
10 On the basis of similar reflections, less restrictive definitions have been proposed; in Apothéloz (2008), for instance, a PC sentence can also be introduced by noun phrases without subordinate clauses, exactly as it is introduced by a phrase modified by a relative clause; among older descriptions see also that of Valli (1981), in which nouns without subordinate clauses or nominalized adjectives are seen as possible pre-copular elements in a PC construction. 11 The corpus was searched for the following keywords: chi ‘who’, dove ‘where’, perché ‘why / because’, come ‘how’, quando ‘when’, quanto ‘how much / how long / how / what’, che ‘who / that / which’, cui ‘whose / of which / to which / to whom’, quale, quali ‘what / which / which one(s)’, a ‘to’ for Italian and où ‘where’, pourquoi ‘why’, comment ‘how,’ quand ‘when’, combien ‘how much’, que, qui, qu’ ‘that / which / who’, dont ‘whose’, lequel, laquelle, lesquels, lesquelles ‘that / which / who’, auquel, auxquels, auxquelles ‘to which / to whom’, duquel, desquels, desquelles ‘whose’, and à ‘to’ for French.
144
Laura Baranzini
distribution of PC sentences within the corpus: PC constructions are indeed more frequent in the Italian section of the corpus than in the French section. It should also be noted that the PC sentence is to be found within texts reporting others’ speech in 63% of occurrences in French and only in 47% of occurrences in Italian. We have already briefly illustrated the standard representation of a PC sentence in the literature; since it proceeds from a theoretical description, such a representation inevitably leads to a forced homogenization of data, as all types of PC sentences are shown as substantially equivalent. In this section, we intend to present real corpus data concerning the forms we have identified, so as to restore a dynamic, hierarchical representation of PC sentences.
3.1 Formal types of Pseudo-clefts in Italian 3.1.1 Types of introducers Let us examine the formal types of PC constructions found within our corpus. First, we identified the types of introducer and classified the PC sentences according to the head of the subordinate clause within the pre-copular element. We then drew a distinction between fused relative pronouns¹² such as chi, quando, perché (‘who’, ‘when’, ‘why’) etc., which included the pronoun introducing the relative clause, and other types of pronouns, for example demonstratives (followed by a relative pronoun) such as quello (‘that’) (in all its inflected forms) or ciò (‘this’, ‘that’), as well as full noun phrases. Within the latter, based on a semantic, not formal, criterion, a further distinction was drawn between generic and non-generic nominal heads (as mentioned in Section 2, generic nominal heads are considered in some typologies as being the analytic equivalent of a fused pronoun).
12 For a definition of fused relative pronoun, see the Oxford Dictionary of English Grammar under Free relative clause: “Free relative clause: A type of relative clause which, like many relative clauses, begins with a wh-word (variously referred to as a free relative pronoun or fused relative pronoun) but, unlike canonical relative clauses, has no overt antecedent” (Aarts/Chalker/ Weiner, eds., Oxford Dictionary of English Grammar, Oxford, Oxford University Press, 1994, 165).
Pseudo-cleft sentences. Italian-French in contrast
145
Table 1: Introducers of Pseudo-cleft sentences in Italian¹³ Total Fused relative pronoun
2
Pronoun – Demonstrative pronoun – Other
50
Full noun phrase¹³ – Generic – Non-generic
25
% 2.6 65 30 20 32.4 8 17
The first category we have taken into account is that of fused relative pronouns. As already mentioned, these pronouns appear as the typical introducer in most examples of PC sentences in the literature. Yet, as shown in Table 1, they are by far the least represented category in our corpus: here, the PC sentence is introduced by a fused relative pronoun in only two cases (chi ‘who’ and dove ‘where’). (7)
[...] dove Monti ha rivelato sapienza politica oltre ogni previsione, è stato nell’affrontare i dossier esplosivi Ici, patrimoniale, regole sui licenziamenti, pensioni [...]¹⁴ (lastampa.it) ‘[…] Where Monti revealed political wisdom beyond all predictions was in handling critical files: town council tax, wealth tax, regulation on dismissal, pensions […].’
Within the category of the other pronouns (those followed by a relative clause), we have focused on demonstrative pronouns, which are the largest group from a quantitative point of view: (8)
[...] quello che non abbiamo fatto – ha concluso – è cambiare Washington [...] (ilsole24ore.com) ‘[…] what we have not done – he concluded – is change Washington [...].’
If we kept to a strict definition of a PC sentence and thus exclusively considered fused relative pronouns and demonstrative pronouns as heads of the subordinate clause, the total occurrences in our corpus would fall to less than half of those ini-
13 Among full noun phrases, we have also counted those with a pronominal head modified by a noun (for instance, uno dei temi ‘one of the topics’). 14 All authentic examples are transcribed without any revision or correction.
146
Laura Baranzini
tially considered. The category which we refer to as “other” includes, in addition to the indefinite expressions uno, ‘one’ or un altro, ‘another’ (which are very rare), il primo, ‘the first’, l’ultimo, ‘the last’, l’unico, ‘the only one’, and superlatives. Together, these represent a significant percentage of the introductory pronouns: (9)
Il primo a prendere la parola è stato il sostituto procuratore generale di Perugia Giancarlo Costagliola. (repubblica.it) ‘The first to speak was the Assistant General Public Prosecutor of Perugia Giancarlo Costagliola.’
As regards full noun phrases (be they modified or not), they appear in most cases in the form of non-generic noun phrases. In 10 occurrences they are generic noun phrases; in this latter case, the noun occurring is almost systematically cosa ‘thing’,¹⁵ often modified by an adjective (altra ‘other’, prima ‘first’, etc.): (10)
La lezione che abbiamo appreso – dice la Merkel – è quella di allontanarci da un’unione indebitata per muoverci verso un’unione stabile e sostenibile. (repubblica.it) ‘The lesson that we learned – says Merkel – is to leave an indebted union and move towards a stable and sustainable union.’
(11)
“L’altra cosa che vorrei dire – conclude Marrazzo – è che attualmente lavoro in Rai come assistente del direttore di Rai3 Antonio Di Bella e mi occupo di documentari.” (leggo.it) ‘“The other thing that I would like to say – concludes Marrazzo – is that I am currently working at Rai as an assistant to Rai3 director Antonio di Bella and I am in charge of documentaries.”’
Finally, we provide data concerning the syntactic function, within the relative clause, of the relative pronoun referring to the head representing the introducer; in a corresponding monoclausal variant, this would be equivalent to the syntactic function of the cleft constituent. Our data (see Table 2 below) show that the subject function (Ciò che conta prima di tutto […] ‘What counts most’) strongly prevails, and that subject and direct object functions (La prima cosa che ha fatto la signora […] ‘The first thing the lady did’) represent almost all the occurrences found in our corpus.
15 For the “grammatical” role of the term chose (‘thing’) in such structures, see Blanche-Benveniste (1986).
Pseudo-cleft sentences. Italian-French in contrast
147
Table 2: Syntactic function of the relative pronoun in Italian¹⁶ Total
%
Subject
48
62.3
Direct object
22
28.6
Other (indirect complements, adjuncts)¹⁶
7
9.1
3.1.2 Subordinate clauses As mentioned when defining the PC sentence, we have taken into account for the present study only the structures whose pre-copular element has an expansion taking the form of a restrictive relative subordinate clause, be it explicit (that is, introduced by a relative pronoun and containing an inflected verbal form, for instance Quello che va ancora per la maggiore […] ‘What is still most popular’) or implicit (introduced by the preposition a followed by an infinitive verb form, such as Il primo a prendere la parola […] ‘The first to speak’). The PC sentences taken into account here are therefore to be considered as a subtype of standard copulative specificational sentences, on the basis of the presence and type of the subordinate expanding the pre-copular pronoun / noun (or containing it, in the case of fused relative pronouns). Table 3: Form of the relative subordinate clause Total
%
Explicit relative subordinate
65
84.4
Implicit relative subordinate
12
15.6
As Table 3 above shows, considering the formal restrictions (type of antecedent) on its possible occurrences, the implicit variant of the PC sentence – which presupposes a subject function for the antecedent within the subordinate clause – represents a significant alternative to the explicit form.
16 “L’unica speranza alla quale si possono aggrappare […]”, ‘The only hope to which they can cling’.
148
Laura Baranzini
3.1.3 Formal types of the cleft constituent We provide hereafter the data concerning the morphological realization of the second element of the PC structure, namely the post-copular constituent. Contrary to some exemplifications in the literature (for other accurate descriptions, see for example Roubaud 1998, 2000), our data show a certain variety in its manifestations. The highest rate of occurrence, however, corresponds to nominal realization, as shown in Table 4 below: Table 4: Form of the cleft constituent in Italian Total Noun phrase – With noun – With pronoun – With pronoun + sentence
60
Prepositional phrase
2
Clause – explicit: che / se ‘that / if’ – implicit: (di ‘of’) + infinitive
15
% 77.9 53 4 3 2.6 19.5 7 8
Three main categories have been taken into consideration: noun phrases, prepositional phrases and clausal phrases. Within the first one, we have drawn a distinction between standard realizations with a nominal head (for instance il Real Madrid ‘Real Madrid’), pronominal realizations (in particular quello/a/i/e di ‘that, those of’ + noun, such as quello di casa mia ‘that of my own home’), and more complex pronominal realizations including an implicit subordinate (quello di ‘that of, that to’ + infinitive, such as quello di allontanarci da un’unione indebitata ‘(that) to leave an indebted union’). Within the category of clausal phrases, we have distinguished implicit variants (introduced or not introduced by a preposition, for instance smentire la sua morte ‘to deny his death’) from explicit variants, containing an inflected verb form (che vada sulle reti Rai ‘that it is aired on Rai channels’). These two types of realization have shown to be, roughly, numerically equivalent. The prototypical manifestation is thus the one which has a full lexical head within the cleft constituent, as shown in example (12) below: (12)
Ma l’ultimo ad utilizzarla era stato l’ex sottosegretario alla presidenza del Consiglio Paolo Bonaiuti. (corriere.it) ‘But the last one to use it had been the former Under-Secretary at the Presidency of the Council of Minister Paolo Bonaiuti.’
Pseudo-cleft sentences. Italian-French in contrast
149
3.2 Formal types of Pseudo-cleft in French 3.2.1 Types of introducers and subordinates The data observed for French suggest a partially different situation from that of Italian as regards the formal features of PC constructions. Let us take a closer look at the elements composing PC sentences in our corpus. The distribution of introducers in French (see Table 5 below) is of extreme interest, as it shows an almost total degree of specialization in comparison with Italian: most PC constructions which occur in our corpus are introduced by a demonstrative pronoun, especially the neutral pronoun ce ‘it’ (in some cases modified by tout ‘all’), or by the personal masculine form (celui / ceux ‘that / those’). Valli (1981) presented the sequence ce qui / que ‘that which’ as being characteristic of PC sentences in spoken French and essentially different from copulative predicative sentences whose pre-copular phrase is a syntactic subject. The head of the relative clause is a full noun phrase (la première chose ‘the first thing’) only in seven cases; in most of them it is a non-generic one. Table 5: Introducers of Pseudo-cleft sentences in French Total
%
Fused relative pronoun
0
0
Pronoun – (tout) ce ‘(all) that’ – celui / ceux ‘that, the one / those’ – other (le moins, le premier, le seul ‘the least, the first, the only’)
39
Noun phrase – Generic – Non-generic
7
84.8 32 3 4 15.2 2 5
The situation for French is thus closer to a description in which a strict definition of a PC sentence would apply, as the inclusion of noun heads, particularly nongeneric ones (for example la détresse ‘the distress’), has received very little, if any, confirmation in the corpus data. It would be interesting to find similar quantitative data referring to a larger corpus, possibly of a different text type. This would allow us to understand to what extent the distribution we have observed reflects a real characteristic of the construction. Other studies which provide quantitative data on forms of PC sentences, e.g. Roubaud (2000), take as a starting-point for observation only the forms introduced by a pronoun and cannot therefore provide an answer at this level.
150
Laura Baranzini
Let us look at the syntactic functions within the relative subordinate clause, listed in Table 6 below: Table 6: Syntactic function of the relative pronoun in French Total
%
Subject
29
63
Direct object
15
32.6
Other
2
4.4
The data referring to syntactic functions are different from those of Italian only in that Italian shows a greater presence of functions which correspond neither to the subject function nor to the object function (examples in French are le seul moment où vous pouvez un peu souffler ‘the only time in which you can relax a little’ and l’autre opposition sur laquelle le film devait travailler ‘the other opposition on which the movie had to work’). As regards the distribution of the two main syntactic roles, however, the two languages show similar data: relative pronouns with a subject function (Ce qui pose problème, c’est à nouveau la proportionnalité ‘What is problematic, is once again proportionality’) are twice as frequent as direct objects (Ce qu’ils vont découvrir, c’est leur univers intime respectif ‘What they are going to discover, is each one’s intimate universe’). As regards subordinates (see Table 7 below), the data for French show an extremely low rate (6.5%) of implicit subordinates (Le premier à s’exprimer est Benoît Lutgen, le président du CDH ‘The first to express himself is Benoît Lutgen, CDH president’), whereas explicit subordinates with inflected verbs (Ce qui compte pour elle, c’est de témoigner ‘What counts for her, is to bring her testimony’) represent more than 90%: Table 7: Form of the relative subordinate Total
%
Explicit relative subordinate
43
93.5
Implicit relative subordinate
3
6.5
Pseudo-cleft sentences. Italian-French in contrast
151
3.2.2 The form of the copula For French too, as for Italian, we have taken into account only copulas constituted by the verb to be, ruling out other copulative verbs (+ to be); a statistically relevant alternation of two forms is nevertheless to be observed in French: we are talking about the alternation between the verb être alone and the same verb preceded by the pronoun ce ‘that / it’ (on occasions, in its elided form c’): (13)
Il me semble que ce qu’ils dénoncent est aussi le concept d’Etat-nation qui semble archaïque aujourd’hui. (lemonde.fr) ‘It seems to me that what they criticize is also the concept of nation-state which seems archaic today.’
(14)
Tout ce que nous voulons montrer c’est que dans la crise actuelle il faut faire des efforts. (lefigaro.fr) ‘All we want to show is that in the current state of crisis it is necessary to make efforts.’
We will not discuss here the variation between the two forms, a debate which has continued since Vaugelas ([1647] 2009).¹⁷ For now, we limit ourselves to highlighting quantitative data, which indicate the distribution shown in Table 8 below: Table 8: Form of the copula in French Total
%
Verb être (‘to be’)
13
28.3
Ce / c’ + verb être
33
71.7
The two possible forms of manifestation of the copula in French have, thus, a clear-cut distribution: in more than 70% of cases, the verb être is preceded by the pronoun ce, whereas it occurs alone in less than a third of the Pseudo-clefts considered here.¹⁸
17 For an overview of the research on the variation est / c’est in the history of grammar from the sixteenth to the twentieth century, see Roubaud (2000: 34); for a recent analysis of the status of pronoun ce in specificational copulas, see Amary-Coudreau (2010). 18 In Blanche-Benveniste (2010), for instance, the prototypical scheme provided for PC sentences is A c’est B, and the expression c’est is treated as a verbal unit.
152
Laura Baranzini
3.2.3 The cleft constituent: Formal types Table 9: Form of the cleft constituent in French Total Noun phrase – With noun – With pronoun
22
Prepositional phrase
1
Clause – explicit: que ‘that’ / quand ‘when’ – implicit: (de ‘of’) + infinitive
23
% 47.8 19 3 2.2 50 19 4
Our data (see Table 9 above) show that the cleft constituent is mainly constituted by a whole clause, usually explicit, with an inflected verb, as in (15) below: (15)
Ce qui s’est passé cette année, c’est que la crise ne s’est pas faite sentir même s’il y a avait clairement un peu moins de gens d’Europe que les années précédentes ‘What happened this year, is that the crisis did not make itself feel although there were clearly fewer people from Europe as compared with the previous years’
The presence of noun phrases is nevertheless significant: in more of 47% of cases, the cleft constituent takes the form of a noun phrase. Such a case is shown in (16) below: (16)
Le seul à répondre clairement est Martin Hirsch ‘The only one to answer clearly is Martin Hirsch’
Conversely, the realizations in the form of a prepositional phrase are negligible.
3.3 Contrastive observations On the basis of the analysis of the quantitative data shown above, some aspects emerge which characterize a PC sentence in the two languages: as regards introducers, aside from a shared tendency to privilege pronominal ones, in French PC sentences it is not uncommon for the antecedent of the relative clause to be a
Pseudo-cleft sentences. Italian-French in contrast
153
noun; this option is only moderately exploited in Italian. Furthermore, it is significant that Italian PC sentences within our corpus seem to specialize in presenting as a cleft constituent a noun phrase whose “weight” or complexity can vary. This confirms the traditional view on the construction as a “focalizer” of referential elements. Data from French do not correspond to this situation. In French, an inverse proportion can be observed: clausal realizations prevail, albeit negligibly, over nominal ones. It may further be noted that the implicit variants of the clausal phrase are less frequent than the explicit ones.
4 Semantics and information structure of Pseudocleft sentences 4.1 Semantic pattern of Pseudo-cleft sentences In Section 2, we mentioned the semantic movement associated with the PC structure, claiming, in particular, that it is a movement of specification. We will now take a closer look at this semantic configuration. One of the main features of all cleft constructions is that they are divided into two parts: the predication is conveyed by two separate units, i.e. two successive moments from a linear point of view. In a PC construction such bipartition, constructed around the copula which acts as a “hinge point”, is concretized according to a dynamic semantic pattern of this type: Expression of an open proposition → Expression of the value allowing saturation of the variable within the proposition.
In other words, in a PC construction the pre-copular phrase is not construed as a referential expression; on the contrary, it allows indication of an unspecified abstract entity (through the description given by the phrase and the subordinate); the predication expressed by the copula then assigns the abstract entity a specific referent, which is expressed by the post-copular phrase: (17)
Quello che riesce sempre a stupirmi è la sua impudenza. ‘What always manage to surprise / surprises me is his impudence.’
In (17), the pre-copular phrase introduces a variable v and says that it satisfies the property riuscire sempre a stupire x ‘always manage to surprise or surprises x’ (where x = speaker); the predication of the PC consists in assigning that variable
154
Laura Baranzini
its value, giving it referential content (here la sua impudenza ‘his impudence’). The existence of referential content which can be assigned to the variable is, therefore, presented as presupposed, and the assertion is constituted by the identification of a referent as the specific concretization of that variable. These two aspects (presupposition of existence and identification of the presupposed referential content) determine the difference between the PC sentence and a linguistic structure which is a potential competitor and is comparable in terms of truth conditions, namely the monoclausal structure; an example is given in (18): (18)
La sua impudenza riesce sempre a stupirmi. ‘His impudence always surprises me.’
In effect, in this case the predication has as a starting-point an identified referent, of which something is said. On the other hand, the two-stage predication of the PC construction is characterized by the introduction of a path: we are led to the identification of a referent starting from the presupposition of its existence. As we shall see, this peculiar structure plays a fundamental role in how information is dealt with in a PC sentence, as well as in the ability of this construction to create links within the text.
4.2 Information structure of Pseudo-cleft sentences in our corpus Traditionally, the PC sentence is mentioned among the linguistic focalizing strategies which can be implemented by means of specific syntactic configurations. For example, in spite of their structural differences, a basic analogy between Pseudo-clefts and Clefts is recognized by Salvi (1991) precisely on the basis of what he refers to as their communicative function, namely the focalization of the post-copular constituent. Thus, it is assumed that the PC sentence has a fundamental informational function that consists in the linguistic signaling of utterance focus¹⁹ through a specific morpho-syntactic structure. Such a function is
19 The concept of focus is almost systematically associated with these constructions, but descriptions often lack problematization or a more precise definition of such notions. Here we choose a notion of focus related to the idea of salient information saturating the open variable of the structure, typically characterized as new, although this is not to be taken as a defining criterion.
Pseudo-cleft sentences. Italian-French in contrast
155
generally recognized in the literature on PC sentences,²⁰ although proximity to Cleft sentences is not always mentioned, and it is even explicitly minimized in Molinier (1996: 92), in favor of the idea of a type of non-contrastive focus: “Cependant, on ne perçoit pas (ou du moins aussi clairement) dans les pseudo-clivées l’effet contrastif propre aux clivées.” [However, the contrastive effect typical of the Clefts is not perceived in Pseudo-clefts (or at least not as clearly)]. As regards the cognitive status of the two members of a PC sentence, the possible configurations are diverse: in Italian, the two members can either carry information which is given (or inferred) within the text and kept more or less active until the reactivation operated by the PC sentence, or they can carry new information. Although relatively free, this informational articulation is nevertheless subject to a preferential pattern, which is based on the configuration most expected in the case of a structure of the type presupposition / assertion: it is the pattern corresponding to the order given / new. In particular, if new information were to be expressed within a PC sentence, the first member p would tend to be interpreted as already active – or at least easily activated from the context – by means of a process of accommodation, so as to make more relevant the fact that the information is presented as being presupposed.²¹ A situation in which all the material is given or inferable from the context is, conversely, unproblematic. In fact, what remains unchanged on an informational level, and thus belongs to the structure, is the feature of “newness” associated with the identification of the referent, that is, with the identification of the referent as the value to be assigned to the variable v; in other words, what brings new information is the assertion itself, the predication of the PC sentence, regardless of the cognitive status of the information within the two members of the construction. The following example clearly shows that new information is provided by the predication itself, whereas the information within the two members of the construction is already given in the immediately preceding co-text (the givenness
20 Müller (2007) proposes a different view of the way in which the information is dealt with and of the focalizing function; it is assumed that the latter is often realized within a larger textual span, including part of the right co-text. 21 Roggia (2009: 155) on Cleft sentences where new information appears within the presuppositional element: “Dal punto di vista dei destinatari della comunicazione il presupposto potrà in questi casi essere già noto indipendentemente dal testo, perché parte di conoscenze enciclopediche condivise con il locutore, oppure diventare noto solo grazie al testo, via accomodamento pragmatico.” [From the addressee’s point of view, in those cases the presupposed information can either already be known regardless of the text, because it is part of encyclopedic knowledge shared with the speaker, or become known only through the text, by means of pragmatic accommodation.].
156
Laura Baranzini
of the post-copular phrase is even lexically signaled through the expression la stessa, ‘herself’): (19)
È caccia sul web ad una serie di foto ‘in desabille’di Scarlett Johansson²². La prima ad inseguirle è la stessa attrice americana che si sarebbe scattata le immagini con un iPhone e che le sono state trafugate da alcuni hacker. (repubblica.it)²³ ‘Everybody on the Internet is after a series of “deshabillé” pictures of Scarlett Johansson. The first one to chase them is the American actress herself who apparently took the photos of herself with an iPhone and which have been stolen from her by some hackers.’
There is a significant difference between Italian and French in how the information is dealt with on a textual level (thematic progression, linking to the co-text, etc.). This is owed to the specialized role which the construction acquires on a meta-textual, argumentative level, which will be described in more detail in Section 5.
5 Textual functions of Pseudo-cleft sentences in our corpus Traditionally, cleft constructions are associated with a strictly focalizing function by which the cleft constituent is highlighted. We have seen in Section 4 that this feature also tends to be associated with the kind of PC sentences observed in our corpus. In addition to this tendency to highlight the cleft constituent, PC sentences in our corpus carry out peculiar, specific functions on a textual level (considering also that focalization alone can be obtained through other types of linguistic strategies). As pointed out earlier, a particular feature of the PC construction (and of cleft constructions in general) consists in dividing the proposition into two parts. This takes place in two successive stages. The first stage corresponds to the expression of the pre-copular complex phrase, which activates a presupposition of existence.
22 In our examples, italics are used to highlight elements or text portions that are considered relevant for our analysis. 23 As already mentioned, the examples have not been modified. In this particular case, the incoherence of the syntactic structure has not been corrected.
Pseudo-cleft sentences. Italian-French in contrast
157
This results in the “opening” of a variable to saturate, thus creating the need for the assignment of a referent. The second stage corresponds to the expression of the post-copular phrase. This element hinges on the identificational predication expressed by the copula to present a second set of information about the referent that is to be associated with the pre-copular phrase. This bipartite structure is based on a preferred cognitive articulation according to which, as already mentioned, the given – or already active – information is the information most often conveyed within the first phrase. Thanks to these two features (information bipartition and given / new sequential order) the PC sentence is particularly well suited to functioning as a “hinge” in specific articulation points in the text where it is required to link backwards to the left co-text – through recovery or reactivation of expressed or inferred information – and then to introduce potentially complex information, which can be new or enriched (or newly characterized by specificational predication), to which the upcoming text is in turn to be linked. The function of the PC as a textual “pivot”, already identified by De Cesare (2005) and related to the notion of “textual connector” introduced by Berretta (2002) with regard to certain types of Cleft sentences, is confirmed by the observation of the data of our corpus, as will be shown in this section. The features allowing the construction to function as described above are therefore of both a formal and an informational-semantic nature. Let us take for instance constructions similar to a PC sentence on an informational level which have the same basic semantic value but display information in the opposite sequential order ([20] and [21] below, obtained by modifying the authentic example [19], are instances of prosodic focalization and Reverse PC sentence respectively). In cases like these, the processing and further expansion of a “heavy” constituent in initial position would be made more difficult. Moreover, the focal element is here in a “confined” position (it is necessarily followed by syntactic material and is placed at the boundary of an information unit but lacks the necessary punctuation allowing for local recovery and development). This position prevents the focal element from developing forwards within the subsequent context. The focalization obtained through prosodic stress or through insertion in a Cleft structure probably facilitates a contrastive interpretation of the highlighted element. Contrary to what happens in a PC sentence, however, such an element does not benefit from the additional focalizing effect obtained through final position within the utterance. Exploiting the focus as the next discourse topic could thus be more difficult; for instance, in the example we have taken into account, the theme following the main sentence recovers the information about the pictures. In these types of focalization, such information is no
158
Laura Baranzini
longer in final position, and is therefore less amenable to being further developed on a textual level: (20)
È caccia sul web ad una serie di foto ‘in desabille’di Scarlett Johansson. La stessa attrice americana, che si sarebbe scattata le immagini con un iPhone e che le sono state trafugate da alcuni hacker, è la prima ad inseguirle. Al momento sulla stampa spagnola sono state pubblicate solo due ma ne esisterebbero altre in compagnia della collega Vanessa Hudgens. ‘Everybody on the Internet is after a series of “deshabillé” pictures of Scarlett Johansson. The American actress herself, who apparently took the photos of herself with an iPhone and which have been stolen from her by some hackers, is the first to chase them. At the moment in the Spanish press only two of them have been published but apparently there are others (showing her) with her colleague Vanessa Hudgens.’
(21)
È caccia sul web ad una serie di foto ‘in desabille’di Scarlett Johansson. È la stessa attrice americana, che si sarebbe scattata le immagini con un iPhone e che le sono state trafugate da alcuni hacker, la prima ad inseguirle. Al momento sulla stampa spagnola sono state pubblicate solo due ma ne esisterebbero altre in compagnia della collega Vanessa Hudgens. ‘Everybody on the Internet is after a series of “deshabillé” pictures of Scarlett Johansson. It is the American actress herself, who apparently took the photos of herself with an iPhone and which have been stolen from her by some hackers, (who is) the first to chase them. At the moment in the Spanish press only two of them have been published but apparently there are others (showing her) with her colleague Vanessa Hudgens.’
Placing the constituents in the same sequence as in a Pseudo-cleft produces effects closer to this construction. Such is the case for reverse cleft (22) or postponed subject clauses (23). In our view, however, the effect of the splitting of information is somewhat less sharp in these cases, especially as regards the selfactivation of the existence of a unidentified referent needing to be saturated; in these cases, the first constituent of the structure seems more dependent, and, as such, more “in need” of being completed on a syntactic level (see in particular
Pseudo-cleft sentences. Italian-French in contrast
159
the second case, which shows a verb-subject structure which is syntactically very cohesive). The activation of a presupposition of existence is weaker in this case. (22)
È caccia sul web ad una serie di foto ‘in desabille’di Scarlett Johansson. Ad inseguirle per prima è la stessa attrice americana che si sarebbe scattata le immagini con un iPhone. Al momento sulla stampa spagnola sono state pubblicate solo due ma ne esisterebbero altre in compagnia della collega Vanessa Hudgens. ‘Everybody on the Internet is after a series of “deshabillé” pictures of Scarlett Johansson. To chase them in the first place is the American actress herself, who apparently took the photos of herself with an iPhone. At the moment in the Spanish press only two of them have been published but apparently there are others (showing her) with her colleague Vanessa Hudgens.’
(23)
È caccia sul web ad una serie di foto ‘in desabille’di Scarlett Johansson. Le insegue per prima la stessa attrice americana che si sarebbe scattata le immagini con un iPhone. Al momento sulla stampa spagnola sono state pubblicate solo due ma ne esisterebbero altre in compagnia della collega Vanessa Hudgens. ‘Everybody on the Internet is after a series of “deshabillé” pictures of Scarlett Johansson. Lit.: is chasing them in the first place the American actress herself, who apparently took the photos of herself with an iPhone. At the moment in the Spanish press only two of them have been published but apparently there are others (showing her) with her colleague Vanessa Hudgens.’
In addition to this type of textual linking, our data show a second configuration, more frequent in the French corpus (a rough preliminary count seems to show that this second configuration appears twice as often in French as in Italian). In this case, the PC sentence has a partially different textual function, with interesting implications on an argumentative level: (24)
Si la gauche gagne, la face de la République n’en sera pas changée... et en sera toute changée en même temps. Le gouvernement pourra continuer de légiférer en donnant, comme la Constitution le lui permet, le dernier mot à l’Assemblée nationale. Il fera juste une croix sur la “règle d'or” budgétaire pour laquelle, de toutes façons, il n’avait déjà pas de majorité. Mais quel symbole qu’une défaite de la droite dans un Sénat que l’on n’avait jamais
160
Laura Baranzini
vu à gauche! le moins qu’on puisse dire est que ça ne va pas favoriser la mise sur orbite de Nicolas Sarkozy pour une réélection en 2012. Au fait, qui avait dit que le Sénat était “une anomalie parmi les démocraties”? Lionel Jospin, le 21 avril 1998, dans une interview au Monde. Preuve que la date du 21 avril ne portait vraiment pas bonheur à Lionel Jospin. (lefigaro.fr) ‘If the left wins, the face of the Republic will not be changed as a result… and at the same time it will be completely changed. The government will be able to continue to legislate, leaving, as allowed by the Constitution, the last word to the national Assembly. It will just have to forget about the budget “golden rule”, for which, anyway, it already didn’t have a majority. But how symbolic this defeat of the right is in a Senate that had never been seen leaning to the left! The least one can say is that this won’t help the launch of Nicolas Sarkozy’s campaign for re-election in 2012. In fact, who said that the Senate was “an anomaly among democracies”? [it was] Lionel Jospin, on April 21, 1998, in an interview to Le Monde. A proof that the date of April 21th didn’t bring luck to Lionel Jospin.’ It may be noted that the above PC sentences have similar pre-copular elements,²⁴ which are meta-discursive in nature. They introduce the speaker’s opinion, feeling or thinking about the discourse topic. Here, the textual link is not established through the more or less literal recovery of content that is active or inferable within the discourse, but through the expression of a comment about the discourse. It follows that, whereas in Pseudo-clefts of the first type the lexical material filling the first constituent is of various kinds, recurring lexical items are instead to be found in Pseudo-clefts of the second type (for example, in French ce qui m’intéresse / ce qui m’agace / ce qui est bien / ce que l’on veut / ce qui est impressionnant etc., i.e. ‘what interests me’ / ‘what annoys me’ / ‘what is good’ / ‘what we want’ / ‘what is impressing’; and in Italian uno dei problemi che abbiamo / l’altra cosa che vorrei dire / quello che mi ha dato fastidio / ciò che è importante
24 The data proposed in Roubaud (2000: 213) for French showed relevant lexical uniformity among the verb items used in the relative clause, and particular numerical prominence of the category of “verbs évaluatifs” (opinion, or evaluative verbs). In Roubaud (2005: 97) this aspect of the corpus is described thus: “Il est intéressant de constater que la construction pseudo-clivée met en avant de nombreux verbes ‘psychologiques’ ou comportant une évaluation: intéresser, plaire, compter […] et que certains autres verbes entreront dans ce dispositif mais en prenant un sens évaluatif.” [It is interesting to observe that the pseudo-cleft construction promotes many verbs that are ‘psychological’ in nature or imply evaluation: to interest, to like, to count […] and that other kinds of verbs will be accepted in the construction but will take an evaluative sense.].
Pseudo-cleft sentences. Italian-French in contrast
161
etc., i.e., ‘one of the problems we have’ / ‘the other thing I would like to say’ / ‘what annoyed me’ / ‘what is important’); such items refer to more specific lexical fields. This latter type of PC does not function in terms of content linking but on a more abstract level, namely a meta-textual level. It allows the speaker to introduce an evaluation or a textual organization comment and to signal it as referring to the current discourse and to previously expressed content.²⁵
5.1 The position of PC constructions in texts Before dealing with the relations between PC sentences and the surrounding text, we looked at their text position within our corpus. We drew a distinction between initial position, middle position and final position so as to verify the presence of utterances in the left and right co-text. As no occurrences of PC constructions have been found in the titles of the articles we took into account, the initial position should be considered as referring to the body of text. Initial and final positions also comprise the sub-distinction concerning reference text units delimited as (partially) autonomous (the whole text or a single paragraph). Furthermore, the PC constructions appearing at the end of a reply in an interview were considered as being in final position. Data relative to initial and final position also include PC sentences that open – or close – direct speech when the latter is introduced or followed only by quoting clauses, for instance.
25 A similar case for the analysis of the textual function of the PC sentences was made in Furukawa (1994) for French. This proposal is based on the observation, in a corpus of movie screenplays, of many marks which can be referred to the speaker within the first segment of the PC (see also Agar Marco in this volume). Conversely, in Păunescu (2002: 123) the PC is assigned a function which is “essentiellement thématisante, au sens où [elles] permettent au locuteur d’introduire un nouveau référent dans l’univers du discours” [fundamentally thematizing, in that [they] allow the speaker to introduce a new referent in the discourse universe]. This description thus underlines how the construction looks forward to the right co-text and the thematic prosecution of the discourse rather than the link with the preceding co-text.
162
Laura Baranzini
Table 10: Position of Pseudo-cleft sentences within the text in Italian Total Initial position – beginning of a paragraph
12
Position within the text
45
Final position – end of a paragraph
20
% 15.6 12 58.4 26 9
Table 11: Position of Pseudo-cleft sentences within the text in French Total Initial position – beginning of a paragraph
13
Position within the text
22
Final position – end of a paragraph
11
% 28.2 12 47.8 24 6
The data shown in Tables 10 and 11 above indicate that the number of PC sentences in middle position is the highest both in Italian and in French; furthermore, whereas a PC sentence rarely occurs at the beginning of the text, it can occur within the text but at the beginning of a paragraph. This particular position is said to prove that a PC sentence can be a preferential device for recovering information from the left co-text. Specifically, the PC globally reactivates a portion of text, even a large one, as well as the inferences drawn from that portion, and marks the passage to a new topic, rising to a hierarchically higher discursive and informational level within the text. The example below shows a prototypical case of distribution of the Pseudo-cleft sentence within a portion of text: (25)
Certo, dietro quel viso rotondo e in apparenza bonario non deve essere un bel carattere, don Juan Pablo Esquivel, se prima di attizzare gli strali dei fedeli di Pairola ha dovuto lasciare, nel 2007, la parrocchia di Atri, cacciato dal vescovo Michele Seccia per ragioni non facilmente decifrabili. Quel che rimane inequivocabile è la sua furibonda reazione, vergata a futura memoria in una lettera in cui bollava come «ipocrita, cinico e grottesco» il suo superiore, concludendo con una minaccia di querela e un perentorio «si consideri preavvisato». È evidente che al quarantottenne
Pseudo-cleft sentences. Italian-French in contrast
163
sacerdote argentino, arrivato in Italia da Santa Fe nel 2002, non piacciono le mezze misure nell’esercizio del suo ministero. [...] (corriere.it) ‘Of course, behind that round and apparently kind face, he must not be a good natured person, don Juan Pablo Esquivel, if before provoking the rage of the parishioners of Pairola he had to leave, in 2007, the parish of Atri, thrown out by the Bishop Michele Seccia for reasons not easy to decipher. What remains unmistakable is his furious reaction, written for future memory in a letter in which he accused his superior of being “hypocrite, cynical and grotesque,” ending with a threat of lawsuit and a peremptory “consider yourself forewarned.” It’s obvious that the forty-eight-year-old Argentinean priest, who came to Italy from Santa Fe in 2002, does not tolerate half-measures in exercising his ministry. […].’ A PC sentence can also occur in final position; in such cases, the post-copular phrase provides all the necessary information about the referent without further recovery or development. It is to be expected that this will mostly happen in cases where the post-copular phrase conveys given information. In such cases, the PC brings new information only through the predication itself. It is indeed less common to see cases in which completely new material is introduced and the text is concluded without any further resumption and development of the new phrase. This can nevertheless happen if one of the two elements of the construction is rich on an informational level and well-developed syntactically: (26)
Il giorno del giudizio dunque sta per arrivare e chi ritiene di aver ancora qualcosa da fare in questo passeggero mondo terreno è bene che si sbrighi. La modalità dovrebbe essere un terremoto, di un’entità tale che quello recente registratosi in Giappone sembrerà quasi uno scherzo. L’unica speranza alla quale si possono aggrappare tutti coloro che non fanno parte degli eletti destinati alla salvezza è il fatto che Camping si sia già candidato all’edizione 2012 degli Ig-Nobel. (corriere.it) ‘The judgment day is near and those who think they still have something to do in this temporary earthly world better hurry up. The modality should be an earthquake of such size that the recent one registered in Japan will almost seem like a joke. The only hope to which all those who are not among the elected destined to salvation can cling is the fact that Camping has already nominated himself for the 2012 Ig-Nobel prize.’
164
(27)
Laura Baranzini
Après, ce que l’on veut, c’est redonner la parole aux citoyens et aux experts pour que se tienne un vrai débat public”. (lefigaro.fr) ‘Then, what we want is give back the right of expression to citizen and experts for a real public debate to take place.’
It may further be noted that there is one only case in French of a PC construction occurring at the very beginning of a text. This seems to confirm that the conveyed information in this construction cannot be entirely new. Even when this appears to be the case, part of such information (preferably, the part conveyed by the first element) is in any case presented as having the cognitive status of known information or information that can be activated, as follows: (28)
Ils ont dit… Kevin Escoffier (Barreur / Régleur): “Ce qui est impressionnant, c’est qu’on est tout le temps à 30 nœuds et ça passe tout le temps. Les questions qu’on pourrait être amené à se poser, c’est combien de temps ça peut tenir? C’est pour ça qu’il faut vraiment préserver le bateau. Aller le plus vite possible quand les conditions de mer le permettent, et dès que la mer est un peu plus hachée et moins ordonnée, il faut lever le pied et faire des moyennes un peu moins élevées. Sinon tout se passe bien. On a tout de même rajouté des couches parce que l’eau est à 4 degrés. Jusqu’ici, il n’y a pas eu de problèmes avec les icebergs. Le prochain danger se trouvera à l’Est des îles Kerguelen.” (lefigaro.fr) ‘They said… Kevin Escoffier (helmsman / trimmer): “What is impressive is that we are all the time at 30 knots and it works all the time. The question one could ask, is how long can it work? It’s for this reason that it is really necessary to manage the boat. Going as fast as we can when the conditions at sea allow it, and as soon as the sea is a little choppier and less calm, one has to ease off the pace, with slightly lower average speeds. Apart from that, everything is working fine. We did add layers, because the water temperature is four degrees. Up to now, no problems with icebergs. The next dangerous area will be east of the Kerguelen Islands.”’
It must be said, however, that the above paragraph was in fact added to a full article and used to quote the words of a person who was involved in the reported event. The discourse topic is therefore strongly active. It is on the basis of this topic that the PC construction introduces an evaluation such as we have described for the second type of textual function identified.
Pseudo-cleft sentences. Italian-French in contrast
165
As we anticipated, the PC sentence can link to the preceding co-text in different ways; more precisely, the variation criteria are the distance between the PC sentence and the element it recovers, its extension and the difference between the original element and the expression used to recover it: 1. the recovery can be local, in contact, by exploiting information activated in the adjacent utterance (as in [29] and [30]), or long-distance, by retrieving information introduced (or inferred) in the preceding text and then reactivating it (as in [31]). It can be observed that, typically, the recovery does not occur at a very long distance within the text,²⁶ although it is difficult to provide precise quantitative criteria. In some cases the recovery is not easily classified as being of the first or of the second type, since the introduced information is kept active by being proposed again one or several times in the text in the form of a discourse topic; the information is therefore both available at a distance and recoverable from the immediately preceding co-text ([32]). (29)
“Ma di quali salotti parli? L’unico che frequento è quello di casa mia”. (corriere.it) ‘What drawing rooms are you talking about? The only one I know is that of my own home.’
(30)
Peu après, le porte-parole de la Maison Blanche, Jay Carney, a tempéré les propos de M. Obama: “Je pense que ce qu’a voulu dire le président c’est qu’il est dans nos moyens de faire ce qu’il faut pour aider le peuple américain.” (lemonde.fr) ‘Shortly after that, the spokesperson of the White House, Jay Carney, tempered the remarks of M. Obama: “I think that what the president meant to say is that it is within our means to do what is needed to help the American people.”’
(31)
Ci sono almeno tre pentiti, dietro l’inchiesta sulle tangenti EnavFinmeccanica. E se mai si decidesse a parlare anche Guido Pugliesi, l’amministratore delegato Enav, sarebbe il terremoto più completo. Sta collaborando infatti Lorenzo Borgogni, il manager che teneva i contatti con la politica per conto di Finmeccanica. Racconta di piccoli clientelismi quale l’assunzione della figlia dell’ex deputato di Forza Italia, Ilario Floresta, attuale membro cda di Enav. Ma sa molto di più. “Vi è una conversazione telefonica inter-
26 Apothéloz (2008) observes that the pre-copular part is generally small. That is justified through arguments related to memory limitations in managing information.
166
Laura Baranzini
cettata dalla quale si evince con solare evidenza che il ruolo di Lorenzo Borgogni dentro Finmeccanica fosse anche quello di occuparsi di contribuzioni illecite al sistema dei partiti”, scrivono i pm Paolo Ielo e Alberto Caperna. E nonostante le sparate di chi si sentiva nel mirino (“Ielo pensa di fare il milanese, ma a Roma le cose si fanno alla romana. O si calma o lo calmano”, dice in un’intercettazione Giampaolo Pinna, responsabile della funzione audit di Enav a Pugliesi), questa inchiesta promette di andare molto lontano. Un altro che sta svelando i meccanismi più segreti è Lorenzo Cola, l’ex consulente personale del presidente di Finmeccanica, Pier Lorenzo Guarguaglini. (lastampa.it) ‘There are at least three informants behind the inquiry on the EnavFinmeccanica bribery scandal. And if ever Guido Pugliesi, the Enav’s CEO decided to talk too, there would be turmoil. Lorenzo Borgogni, the manager in charge of political relations on behalf of Finmeccanica, is collaborating. He is talking about patronage on a small scale such as the hiring of the daughter of the former Forza Italia member of parliament, Ilario Floresta, current member of the board of Enav. But he knows much more. “There is an intercepted telephone conversation which very clearly leads to the assumption that the role of Lorenzo Borgogni inside Finmeccanica was also that of being in charge of illegal funding to the system of political parties,” write prosecutors Paolo Ielo and Alberto Caperna. And in spite of the slams of those who felt under scrutiny (“Ielo thinks he can behave as a Milanese. But in Rome you do things the Roman way. Either he calms down or they will calm him down,” says in an intercepted call Giampaolo Pinna, head of the Enav Audit department, to Pugliesi), this inquiry promises to go very far. Another one who is revealing the most secret mechanisms is Lorenzo Cola, former personal consultant for Finmeccanica’s chairman, Pier Lorenzo Guarguaglini.’ (32)
Les statistiques montrent que le vainqueur de la première étape est le vainqueur final de la Volvo… Iker Martinez : (rires) Exactement! Non, cela a marché sur les précédentes éditions, parce que la différence entre le premier et les autres avait été énorme ici, au Cap. Mais cette fois-ci, je n’y crois pas. Là, le vainqueur aurait pu être Puma s’il n’avait pas démâté, Camper est allé très vite hier et pour Groupama4 qu’on a vu très rapide au départ, c’est une option qui
Pseudo-cleft sentences. Italian-French in contrast
167
n’a pas payé. Non, celui qui gagnera sera celui qui sera le plus rapide à développer son bateau à chaque étape. (lefigaro.fr) ‘Statistics show that the winner of the first leg is the final winner of the Volvo cup… Iker Martinez: (laughs) Exactly! No, that worked this way in the past editions, because the difference between the first and the others had been enormous here, at the Cape. But this time, I don’t think so. In this case, the winner could have been Puma if their mast hadn’t broken, Camper was very fast yesterday and for Groupama4, which we saw was very fast at the start, this option has not proven to be successful. No, the one who will win will be the one who will be the fastest to develop his boat’s skills at each leg.’ 2.
(33)
The recovery may vary also depending on how large a part of the text is reused in the PC sentence, ranging from a single phrase ([33]) to a larger section of the text; the single phrase might also express the core concept developed in a section of the text. This results in the impossibility of distinguishing between the two situations, as in (34). Il presidente fa quindi un breve bilancio dei risultati raggiunti dalla sua presidenza: “In questi due anni e mezzo alla Casa Bianca abbiamo fatto molto, su fronte delle riforme, da quella sanitaria a quella scolastica, abbiamo fatto passi avanti sull’uguaglianza salariale. Ma quello che non abbiamo fatto – ha concluso – è cambiare Washington, la gente così ha perso fiducia. (ilsole24ore.it) ‘The president then proceeds to a brief assessment of the results achieved by his presidency: “In these two and a half years at the White House we have done a lot, as regards reforms, from the health system reform to the school system reform, we have made progress on equal pay. But what we haven’t done – he concluded – is change Washington, and so the people has lost faith.’
(34)
Déléguée générale de Syntec Ingénierie, Karine Leverger dresse un constat similaire. «De plus en plus de bureaux d'ingénierie sont contraints de délocaliser une part croissante de leur activité à l’étranger sous la pression de donneurs d’ordre. On a franchi, ces derniers mois, une étape supplémentaire. En 2009, il y avait eu peu de suppressions d’emplois dans l’ingénierie industrielle, sauf dans l’automobile. Cette fois, je crains des pertes sèches», observe-t-elle. Safran, notamment, aurait ainsi exigé de
168
Laura Baranzini
ses prestataires en R&D qu’ils diminuent leurs coûts de 10%. «C’est quasiment mission impossible sans délocaliser», constate Karine Leverger. Pour autant, ce que les entreprises vont chercher à l’étranger, ce n’est pas avant tout le moindre coût. (lefigaro.fr) ‘General manager of Syntec Ingénierie, Karine Leverger draws similar conclusions. “More and more engineering firms are forced to relocate an increasingly large part of their activity under the pressure of ordering parties. We have reached, these last few months, a further milestone. In 2009, there had been few job losses in industrial engineering, except for the automobile sector. This time, I’m afraid there will be deadweight loss,” she observes. Safran, in particular, has apparently asked his R&D providers to reduce costs by 10%. “It’s almost an impossible task without relocating,” observes Karine Leverger. And yet, what firms look for abroad is not primarily the lowest cost.’ 3.
(35)
Furthermore, as the above examples have shown, a more local recovery can be expressed by resuming essentially the same words as in the first mention or by using equivalent words. In such cases, too, new information is brought by the resumption expression (even only as an effect of variation). At the same time, the resumption can also take place indirectly. In this case, reference is made not to the same entity occurring in the preceding speech but rather to an aspect which can be derived from it, or to an inference, or something else of the kind. This is, indeed, what happens in most cases: “È il motore che uso anch’io, e apprezzo che ci sia una continua innovazione, una notevole vitalità. Non è più solo motore ma anche tante altre cose. Quello che non mi piace è l’aspetto business. La pubblicità su internet comporta che si facciano dossier sulle persone, e questo non mi va. (repubblica.it) ‘That’s the search engine I also use, and I appreciate that there is ongoing innovation, remarkable vitality. It is no longer just an engine, it is many other things. What I do not like is the business aspect. Advertising on the Internet involves the creation of files on people, and that I don’t like.’
4. An extreme manifestation of such indirect recovery could be implicit reference, within the description provided for the second type of textual function of the PC sentence (“metadiscursive” function). No contents are recovered in this case, but reference is made (often implicitly) to the relevance of a topic within the discourse. The Pseudo-cleft therefore allows for an evaluation or
Pseudo-cleft sentences. Italian-French in contrast
169
a reflection on that topic. (36) is an example in which a linguistic link is still present;²⁷ this occurrence is thus close to the type of indirect recovery illustrated in (35). Conversely, in (37), the domain of relevance of the observation introduced by the PC is to be reconstructed solely on a pragmatic level. (36)
J’étudie à l’Asian Institute of Technology, à Rangsit. Notre campus est sous trois mètres d’eau depuis une semaine maintenant. Tout le monde a été évacué, mais beaucoup ont perdu leurs affaires et nous avons été livrés à nous-mêmes du jour au lendemain sans avoir d’explications. Ce qui est frappant ici, c’est la difficulté à avoir une information fiable. Nous suivons surtout le fil d’infos de notre école et le groupe Facebook « Flood situation at AIT ». (lemonde.fr) ‘I am studying at the Asian Institute of Technology, in Rangsit. Our campus has now been under three meters of water for one week. Everybody was evacuated, but many have lost their possessions and we were left completely alone from one day to the next without any explanations. What is impressive here is the difficulty of getting reliable information. We are mostly following the information thread of our school and the “Flood situation at AIT” Facebook group.’
(37)
L’opposition, qui n’a toujours pas digéré sa défaite aux dernières élections, a tout intérêt à jouer sur la dramatisation. Elle tente, pour une partie de ses membres, de prouver l’incompétence de l’actuel gouvernement. Concrètement, ce que je constate pour le moment c’est que le district de Saun Luang-Pra Khanong, où j’habite, n’a pas une goutte d’eau dans ses rues. (lemonde.fr) ‘The opposition wing, who still haven’t digested their defeat in the last elections, have all interest in playing on dramatization. A certain part of its members is attempting to prove the incompetence of the current government. Concretely, what I observe for now is that the district of Saun LuangPra Khanong, where I live, does not have a single water drop in its street.’
27 The linguistic link in this specific instance is particularly suitable for illustrating the case we are dealing with. In effect, within the news, reference is made to certain places. The resumption via a location adverb might indeed refer to such places, but is more likely to be interpreted as a metaphorical reference to the whole situation. Suppressing the adverb would solve the ambiguity and lead to this latter interpretation.
170
Laura Baranzini
The data in Tables 10 and 11 indicate that the final position, although more frequent than the initial position, concerns only a minority of cases compared with the middle position within the text. This means that, in most occurrences, the PC construction is followed by utterances with which it establishes different kinds of informational-textual relations. In the light of the above observations, we claim that the PC construction establishes a bidirectional link on a textual level. Let us look at some examples of textual progression within our corpus: (38)
“La Fiat continuerà a produrre automobili e quello che è da stabilire è se l’Italia vuole fare automobili”. È questa la domanda posta al governo dal presidente di Fiat, John Elkann, in un passaggio del suo intervento a Rimini. Elkann è arrivato alla terza giornata di lavori del convegno organizzato da Comunione e Liberazione in compagnia dell’amministratore delegato del Lingotto Sergio Marchionne. (corriere.it) ‘“Fiat will continue to produce cars and what is to be established is if Italy wants to make cars”. This is the question addressed to the government by Fiat chairman John Elkann, during his speech in Rimini. Elkann came to the third day of works of the congress organized by Comunione e Liberazione with Fiat CEO Sergio Marchionne.’
In (38), the cleft constituent is categorized as a question in the immediately subsequent utterance; it functions as a support for the introduction of new elements (strictly related, on a conceptual level, to the notion of “question”, that is, the source of the question, the situation in which it is asked, etc.: Elkann, intervento a Rimini ‘Elkann, speech in Rimini’). These elements will in turn be developed in the following utterance. Here, the first member of the PC sentence is also recovered, through a sort of reformulation summarizing the whole construction (è questa la domanda ‘this is the question’); this creates a particularly cohesive textual sequence and allows for a progressive introduction of the information. French examples like (39) (m’intéresse → je serai très heureux; voir la Tunisie développer une démocratie [...] → verrai cela ‘interests me → I will be very happy → to see Tunisia developing a democracy […]→ I see that’) also show the same kind of reformulation which summarizes the whole PC sentence as a cohesive means for developing or introducing new information in the text: (39)
Je me trouve beaucoup mieux dans l’univers de la pensée que dans celui de la politique. Ce qui m’intéresse, c’est de voir la Tunisie développer une démocratie qui marie l’islam et la modernité. Je serai très heureux quand je verrai cela. Je suis vice-président de l’organisation mondiale
Pseudo-cleft sentences. Italian-French in contrast
171
des savants musulmans, et s’il ne me reste plus rien à faire en Tunisie, le monde musulman est vaste… (lemonde.fr) ‘I feel much more comfortable in the universe of thought than in that of politics. What I am interested in is to see Tunisia develop a democracy which could combine Islam and modernity. I will be very happy when I see that. I am vice-president of the International Union of Muslim Scholars, and if there is nothing left for me to do in Tunisia, the Muslim world is vast…’ The recovery process can therefore involve several elements of PC construction, and even the whole predication; however, such a link can also be established more locally, and the support for the progression of information can even be a single element of the cleft constituent only. As in the case of the link to the left co-text, this can take place directly or indirectly. In most cases, the element is recovered through reformulation, or through a term indirectly linked to it: (40)
Nel dettaglio, uno che si è prestato a tutto, ma ora s’è pentito, è il costruttore Tommaso Di Lernia. Devastanti i suoi racconti. (lastampa.it) ‘In more detail, one who has consented to everything, but now has repented, is the builder Tommaso Di Lernia. His stories are ravaging [Lit.: Ravaging, his stories].’
The linked elements are almost always adjacent, but this is not a necessary condition as examples of long-distance recovery are indeed to be found in our corpus; in some cases the segments are separated by recovery and development or other relevant information, which is then “jumped over” so as to move back to the information within the PC sentence: (41)
Tutta questa indagine, come detto all’inizio, riguarda “il percepito” sul razzismo in Italia. Quello che non è percepito ma è invece un dato tout court è la (non) conoscenza della legge italiana sulla cittadinanza, proprio quella che il Capo dello Stato ha invitato a rivedere una settimana fa. A oggi, infatti, se sei nato in Italia da genitori stranieri non puoi avere la cittadinanza fino a i 18 anni, e poi hai un solo annodi tempo per farne richiesta, altrimenti toccherà aspettarne altri tre. Domanda finale: “Lo sapevi?”. Risposta corale: “No”. (lastampa.it) ‘This whole enquiry, as mentioned in the beginning, concerns “the perceived” on racism in Italy. What is not perceived but is actually a data tout
172
Laura Baranzini
court is the (non-) knowledge of the Italian law on citizenship, precisely the one that the Italian President has urged to review a week ago. Today, in effect, if you are born in Italy from foreign parents you cannot have Italian citizenship until 18 years of age, and then you only have a year to request it, after which you will have to wait three more years. The final question: “Were you aware of that?”. The unanimous answer: “No”.’ On an argumentative level, the strong cohesion with the preceding co-text takes two different directions, both cognitively basic: a link is reconstructed, which might be dictated by continuity (a sum of co-oriented pieces of information, see [43]) or by opposition (see [42]): (42)
“È il motore che uso anch’io, e apprezzo che ci sia una continua innovazione, una notevole vitalità. Non è più solo motore ma anche tante altre cose. Quello che non mi piace è l’aspetto business. (repubblica.it) ‘That’s the search engine I also use, and I appreciate that there is ongoing innovation, remarkable vitality. It is no longer just an engine, it is many other things. What I do not like is the business aspect.’
(43)
La première solution serait de donner à la Cour de justice européenne le pouvoir d’invalider un budget. La seconde, de donner ce pouvoir à une instance parlementaire. Ceux qui payent la garantie in fine, ce sont les contribuables, et ils doivent avoir leur mot à dire via leurs élus. (lemonde.fr) ‘The first solution would be to give the European Justice Court the power to invalidate a budget plan. The second, to give this power to a parliamentary body. Those who pay the warranty, in the end, are the taxpayers, and they must have a say through their elected representatives.’
In (43), the PC sentence is interpreted as an argument proving the good sense of the second solution proposed (donner ce pouvoir à une instance parlementaire ‘give this power to a parliamentary body’). Since this solution occurs in second position and is the only one which is explicitly motivated, it is easy to interpret the PC sentence not only as a justification of the plausibility of the second solution, but indeed as a dialectic argument in favor of it. This stronger interpretation is then confirmed by the development of the subsequent text, where the relation of motivation (ils doivent avoir leur mot à dire ‘they must have a say’) is the object of explicit further elaboration. Conversely, (42) illustrates a case of opposition –
Pseudo-cleft sentences. Italian-French in contrast
173
much more rare in our corpus —, which can be reconstructed on a lexical level on the basis of the contrast apprezzo / non mi piace ‘I appreciate / I do not like’. Opposition, whether explicit or not, is one of the two possible effects of the cohesive feature described above and it can be assumed that the PC sentence is well suited for the expression of opposition because it integrates it easily into the argumentative and textual structure; although a break is created, a strong link is kept with the preceding context, and at the same time a new element is introduced. This element is most suitable for further developing the relation of opposition and making it productive. In cases where the PC sentence has a textual function of the second type (metalinguistic, as in [44] and [45]), a definition of its argumentative effects in terms of co-orientation or opposition is not as relevant. In this case, the construction functions outside the discourse content; it is used to make explicit an element involved in textual organization, in a neutral way ([44]), or to express a positive or negative evaluation. In this case, it acquires relevance on an argumentative level ([45]): (44)
[…] “Ho chiesto ai miei legali – ha detto il giornalista ed ex governatore del Lazio – di tutelarmi in tutte le sedi nei confronti di chi ha diffuso questa notizia con un intento evidente di delegittimazione, visto che non c’era nessun fondo di verità, nemmeno una proposta. Dico questo con tutto il rispetto del programma e per chi ci lavora.” “L’altra cosa che vorrei dire – conclude Marrazzo – è che attualmente lavoro in Rai come assistente del direttore di Rai3 Antonio Di Bella e mi occupo di documentari”. (leggo.it) ‘[…] “I asked my lawyers – said the journalist and former governor of Lazio – to safeguard me on all levels from whoever has spread this information, obviously aimed at delegitimizing me, as there was not a shadow of truth in it, not even a proposal. This is said with all due respect to the show and to those who work in it.” “The other thing that I would like to say – concludes Marrazzo – is that I am currently working at Rai as an assistant to Rai3 director Antonio di Bella and I am in charge of documentaries”.’
(45)
“On est content d’être là, ce fut une très longue étape. Mais on est ‘safe’ et ce n’est pas si mal, on est mieux servi que certains”, a relativisé Cammas. “Ce qui est bien, c’est qu’on a toujours poussé le bateau au maximum. Ces trois derniers jours en arrière du front nous a contraints à partir vers le
174
Laura Baranzini
sud, et l’on pu s’entraîner pour les 40es rugissants. Une bonne chose. On est plus forts aujourd’hui qu’il y a trois semaines. […] (lefigaro.fr) ‘“We are happy to be here, it was quite a long leg. But we are ‘safe’ and that’s not so bad, we are in a better position than some others”, said Cammas, putting things into perspective. “What good is that we always pushed the boat to the maximum. These three last days behind the front forced us to head south, we could train for the Roaring Forties. That was a good thing. We are stronger now than three weeks ago. […].’
6 Conclusions In this paper we have presented the results of an analysis of Pseudo-cleft sentences based on a bilingual (Italian-French) corpus of newspaper texts. After a theoretical introduction of the object of study, we proposed a multi-level qualitative and quantitative description of the types of Pseudo-cleft constructions retrieved from the corpus. Particular attention has been paid to the description of the main formal types of Pseudo-clefts as well as the illustration of their textual functions (including a description of some of their argumentative functions) from a comparative perspective on the two languages. The analysis of real data highlights discrepancies between the description of the construction in the literature and the empirical evidence drawn from our data: some of the features prototypically associated with PC sentences are not to be found in our corpus. Furthermore, our analysis has provided comparable quantitative results on a particular text type. Naturally, all our observations are typologically marked. Our data allow for a systematic comparison with data drawn from the analysis of other types of text. Italian data were compared with corresponding French data; the comparison has shown that the constructions in the two languages are, essentially, equivalent from a functional point of view. Conversely, differences were observed in terms of frequency, as well as in the distribution of the different formal types. Within our corpus, the Pseudo-cleft sentence occurs more frequently in Italian texts. In its prototypical manifestation, it has the following features: 1. it is introduced by a demonstrative pronoun or, secondarily, by a noun; 2. in Italian, its cleft constituent is mainly a nominal subject; in French it is mainly a clausal subject; 3. it is both preceded and followed by linguistic material, and it relates to both its preceding and its following co-text in a number of ways on a textual level.
Pseudo-cleft sentences. Italian-French in contrast
175
In the texts that we analyzed, Pseudo-cleft sentences are associated with two textual functions, in varying percentages (in particular, the second function is significantly more frequent in French than in Italian); these two functions share the property of cohesively linking the content of the PC with the preceding co-text, as well as the potential ability to link forward to the next portion of text. The first function consists in the resumption of information already expressed or recoverable from the context. Such information evokes a semantic variable that needs to be saturated. This results in a syntactic-semantic tension, thanks to which the recovery of information from the left co-text can be related to the potential development to the right of the information given in the cleft constituent, with the PC structure acting as a “hinge”. The second function concerns the possibility of introducing into the text, through the PC construction, a meta-textual comment, often of an evaluative nature, which exploits the preceding co-text to draw the relevant discourse topic; in this case too, the evaluative remark takes the form of a generic variable, which finds context-appropriate concretization in the cleft constituent and in its possible development within the following co-text.
References Agar Marco, Rocío. This vol. Pseudo-cleft Sentences. Italian-Spanish in contrast. Akmajian, Adrian. 1970. On deriving Cleft sentences from Pseudo-cleft sentences. Linguistic Inquiry 1. 149–168. Amary-Coudreau, Valérie. 2010. Qu’est-ce que “c’ ” est?. Le Français Moderne 78 (2). 259–277. Apotheloz, Denis. 2008. À l’interface du système linguistique et du discours: l’exemple des constructions identificatives (e.g. pseudo-clivées). In Olivier Bertrand, Sophie Prevost, Michel Charolles, Jacques François & Catherine Schnedecker (eds.), Discours, diachronie, stylistique du français. Etudes en hommage à Bernard Combettes, 75–92. Bern: Peter Lang. Baranzini, Laura. 2014. Le frasi pseudoscisse nei testi giornalistici online: italiano e francese a confronto. In Enrico Garavelli & Elina Suomela-Härmä (ed.), Dal monoscritto al web: canali e modalità di trasmissione dell’italiano. Tecniche, materiali e usi nella storia della lingua. Atti del XII Congresso SILFI Società Internazionale di Linguistica e Filologia Italiana (Helsinki, 18–20 giugno 2012), 567–574. Firenze: Franco Cesati. Berretta, Monica. 2002. Quello che voglio dire è che: le scisse da strutture topicalizzanti a connettivi testuali. In Gian Luigi Beccaria & Carla Marello (eds.), La parola al testo. Scritti per Bice Mortara Garavelli, 15–31. Torino: Edizioni dell’Orso. Blanche-Benveniste, Claire. 1986. “Une chose” dans la syntaxe verbale. Recherches sur le français parlé 7. 141–168. Blanche-Benveniste, Claire. 2010. Les pseudo-clivées et l’effet deux points. In Marie-José Béguelin, Mathieu Avanzi & Gilles Corminboeuf (eds.), La parataxe 2. Structures, marquages et exploitations discursives, 185–218. Bern: Peter Lang. Collins, Peter C. 1991. Cleft and Pseudo-cleft Constructions in English. London: Routledge.
176
Laura Baranzini
De Cesare, Anna-Maria. 2005. La frase pseudoscissa in italiano contemporaneo. Aspetti semantici, pragmatici e testuali. Studi di grammatica italiana XXIV. 293–322. De Cesare, Anna-Maria. This vol. Cleft Constructions in a contrastive perspective. Towards an operational taxonomy. Furukawa, Naoyo. 1994. Ce que je crois, c’est que...: séquence thématique et ses deux aspects, cohésion et rupture. Travaux de linguistique 29. 21–37. Molinier, Christian. 1996. Constructions en “C’est”: une classification générale. Cahiers de Grammaire 21. 75–94. Müller, Gabriele. 2007. La construction pseudo-clivée dans l’organisation d’activités complexes: questions de portée. Cahiers de praxématique 48. 231–238. Panunzi, Alessandro. 2008. Strutture scisse e pseudoscisse: valori d’uso del verbo essere e articolazione dell’informazione nell’italiano parlato. In Angela Ferrari (ed.), Atti del X Congresso SILFI. Sintassi storica e sincronica, Subordinazione, coordinazione, giustapposizione (Basilea, 30 giugno–3 luglio 2008), vol. 2, 1121–1137. Firenze: Franco Cesati. Păunescu, Marina-Oltea. 2002. Les énoncés pseudo-clivés comme énoncés introducteurs de thème de discours. Revue Roumaine de Linguistique XLVII (1–4). 121–134. Roggia, Carlo Enrico. 2009. Le frasi scisse in italiano. Struttura informativa e funzioni discorsive. Genève: Slatkine. Roubaud, Marie-Noëlle. 1998. Constructions en c’est: les pseudo-clivées. Cahiers de grammaire 23. 81–94. Roubaud, Marie-Noëlle. 2000. Les constructions pseudo-clivées en français contemporain. Paris: Honoré Champion. Roubaud, Marie-Noëlle. 2005. Une construction trop peu exploitée, la construction pseudoclivée. In Geoffrey Williams (ed.), La linguistique de corpus, 93–100. Rennes: Presses Universitaires de Rennes. Salvi, Giampaolo. 1991. Le frasi copulative. In Giampaolo Salvi, Lorenzo Renzi & Anna Cardinaletti (eds.), Grande grammatica italiana di consultazione, vol. 2, 163–189. Bologna: Il Mulino. Valli, André. 1981. Note sur les constructions dites “pseudo-clivées” en français. Recherches sur le français parlé 3. 196–211. Vaugelas, Claude Favre de. 2009 [1647]. Remarques sur la langue françoise. Zygmunt Marzys (ed.). Genève: Droz.
Rocío Agar Marco*
Pseudo-cleft sentences. Italian-Spanish in contrast 1 Introduction The aim of this paper is to present the first results of my doctoral research project, which investigates Pseudo-cleft constructions in Italian and Spanish from a contrastive perspective.¹ In the present study, we propose a typology of forms of pseudo-clefts and evaluate the frequency of use of each form in the two languages; we will then analyze the textual functions of the construction in online journalistic writing, focusing our attention on the cleft constituent (in bold in the following Italian example: quello che mi piace è la linguistica ‘what I like is linguistics’). It is a well-established fact that Pseudo-cleft sentences (henceforth, PCs) are focusing devices in many languages, including Italian and Spanish (see for instance Kiese 1993, Berretta 1994 and 1995, De Cesare 2005 etc.). However, as our data will show, if we take into account the co-text of the construction (particularly the subsequent co-text), the discourse and pragmatic properties of the Focus of a PC go far beyond a mere focalization of the cleft constituent (henceforth, CC). Indeed, it will be observed in Section 4 that the PC plays a crucial role in the thematic organization of texts.
* The issues developed here are part of a corpus-based research study conducted within a Swiss National Science Foundation project based at the University of Basle and directed by Anna-Maria De Cesare, whom I would like to thank for her helpful discussion on this paper. This is an ongoing work, and consequently many of the issues dealt with are still open. I am also grateful to Alberto Gil for his suggestions and critical observations on a previous draft of this paper as well as to Claudia Ricci for her help in the translation of this paper. 1 The aim of my doctoral dissertation is to provide a contrastive study of Pseudo-cleft sentences in Italian, Spanish and German; particular attention is paid to the construction’s frequency, form and functions in different types of written texts, as well as to the identification of equivalents to Italian Pseudo-cleft sentences in Spanish and German. In the latter language, Pseudo-cleft sentences occur relatively seldom (compared to the other aforementioned languages and based on the quantitative data from our corpus, i.e. written texts from online newspapers. Conversely, the frequency of the construction in spoken German seems to differ slightly, as we can see for instance in Birkner 2008). Rocío Agar Marco, University of Basle
178
Rocío Agar Marco
In the present work, we have chosen to concentrate on the Focus of PCs because most of the literature focuses on the status of the relative clause of the construction, on the fact that such a clause contains information which is presupposed and given, and on how it connects to the prior co-text (see for instance Sornicola 1988 and Berretta 1995: 161). Conversely, much less research can be found on the functions of the CC, which is the item coinciding with the most salient portion of the construction. An exception to this relative lack of research on the CC is De Cesare (2005), which is the starting point for this study. This paper thus aims to provide a typology of the structures taken into account (Section 2) as well as to present quantitative data concerning the frequency of PCs (Section 3). The behavior of the CC within the text will then be analyzed (Section 4), with particular attention to a) the recovery, or the non-recovery, of the CC in the subsequent co-text, b) the ways in which the CC is recovered and c) its discourse functions. Our study is based on a corpus composed of two parts: a subcorpus of texts in Italian (approximately 520,000 words) and a subcorpus of texts in Spanish (approximately 350,000 words). Both subcorpora are part of the ICOCP-corpus,² which was created specifically for the project of the same name. All texts in the corpus are original (non-translated) texts, drawn from different newspaper sections (politics, economy, sports, etc.) and taken from online newspapers only. Most articles in Italian date back to 2011; they represent different text types (articles similar to corresponding paper articles, press releases and free press articles). Our subcorpus of texts in Spanish was elaborated upon within the research project in a subsequent phase, applying similar criteria to those used in the creation of the ICOCP-corpus, in order for it to be comparable to the other subcorpora. The texts in Spanish (only the first text type described above: articles similar to corresponding paper articles) were collected in early 2012.³
2 A typology of Pseudo-cleft sentences PCs, which have been the object of several studies in different languages, have been widely described for their syntactic and pragmatic properties.⁴ Given the
2 For further details, see De Cesare et al. (in this volume). 3 There is, thus, some thematic continuity between Italian and Spanish texts because certain topics are still of interest; Italian and Spanish texts are therefore partly comparable also on a content level. 4 For a specific discussion of PC syntactic and pragmatic properties and other parameters, see the works of Sornicola (1988), Berretta (1994, 1995, 1996 and 2002), De Cesare (2005 and 2011), Panunzi (2009), and Baranzini (in press as well as in this volume) for Italian; Moreno Cabrera
Pseudo-cleft sentences. Italian-Spanish in contrast
179
breadth of literature existing on the subject, I have chosen here not to discuss every feature of this construction thoroughly. I will limit myself to providing a definition of the construction and to describing the typology on which the corpus search and the subsequent analysis were based. I will then discuss some formal properties which could be directly related to the quantitative results of my research within the reference corpus. For further clarifications on the definition of PCs adopted in the ICOCP project, see Baranzini and De Cesare (both in this volume).
2.1 Definition of Pseudo-cleft sentences The PCs traditionally subsume the structures introduced by a fused relative pronoun (also called free relative pronoun), as in (1) and (2), as well as those in which the relative clause is headed by a complex pronoun, as in (3) and (4): (1)
It. Chi mi sta più simpatica è tua sorella.⁵ who me-dat is the.most likeable is your sister ‘The one I like the most is your sister.’
(2)
Sp. Quien me cae mejor es tu hermana. who me-dat seems the.best is your sister ‘The one I like the most is your sister.’
(3)
It. Quello che voglio è venire con te. what I.want is come-inf with you ‘What I want is to go with you.’
(4)
Sp. Lo que quiero es ir contigo. what I.want is go-inf with.you ‘What I want is to go with you.’
Our definition of PCs is broader than the description usually found in the literature, which is based on the relative clause introducer. In contrast to the classic definition of PCs, our definition also includes cases in which the antecedent of
(1999), Pinedo (2000), Helfrich (2003) and Van den Steen (2005) for Spanish; Metzeltin (2010) and Agar Marco (2014) for both languages. 5 In examples (1)–(9) and (20)–(32) the subordinate clause introducer is highlighted in italics.
180
Rocío Agar Marco
the relative clause is a generic noun (see [5] and [6]). This type of PC is taken into account only in some works (see for instance Berretta 2002: 17). (5)
It. La persona che gli manca di più è the person that him-dat misses the.most is sua madre. his mother ‘The person he misses the most is his mother.’
(6)
Sp. La persona a la que más echa de menos the person to the that most he.misses a su madre. to his mother ‘The person he misses the most is his mother.’
es is
In the present paper, we also include structures in which the antecedent of the relative clause is a less generic noun (see Salvi 1991: 177 and Panunzi 2009: 1128). Example (7), proposed by Salvi, and example (8) illustrate this type of PC in both languages:⁶ (7)
It. Il mese in cui verrà è marzo. the month in which s/he.will.come is March ‘The month in which s/he will come is March.’ (Salvi 1991: 177)
(8)
Sp. El mes en que vendrá es marzo. the month in that s/he.will.come is March ‘The month in which s/he will come is March.’
In our view, the PC is equivalent to a specificational structure whose pre-copular segment either coincides with a relative clause or contains it. Furthermore, our definition of PCs includes constructions that are impossible to reduce to a single clause⁷ (which is a syntactic criterion often used in the literature as a means of excluding from the class of PCs some cases which, by contrast, are here consid-
6 All these different types of PC will be discussed in more detail in Section 2.3. 7 For instance, reduction of the PC to a single clause in (3) and (4) would give, respectively: Voglio venire con te and Quiero ir contigo (‘I want to go with you’). The reduced utterance is, thus, a syntactic structure containing the same information but expressed differently.
Pseudo-cleft sentences. Italian-Spanish in contrast
181
ered as having the same textual functions).⁸ The only syntactic requirement considered here to be mandatory in defining PCs is that the pre-copular segment of the construction must contain a relative clause (see examples [1] to [8]). Our definition thus does not include syntactic structures in which the relative clause is found after the copula, as is considered for instance by Salvi (1991). An example of this case is (9): (9)
It. È tua sorella chi mi sta più simpatica. is your sister who me-dat is the.most likeable ‘It is your sister the one I like the most’
In summary, and to provide a unified account of the syntactic properties shared by all the types of PCs taken into account in our study, we can describe the structure of PCs by means of the abstract representation given in (10), which has been applied to examples (5) and (6), now (11a) and (11b): (10)
‘[(NP +) relative clause]NP + copula + CC’⁹
(11)
a.
It. La persona
b. Sp. La persona
che gli manca di più
è
a la que más echa de menos es
head of the rel. clause relative clause
sua madre. a su madre.
copula cleft constituent
According to the representation given in (10), the main parts into which the PC structure is divided are, in order, the relative clause, possibly headed by an NP, the copula and the cleft constituent. As has been shown in previous studies, the CC can have different syntactic functions (subject, object, adjunct etc.) which will not be discussed here.¹⁰ The PC structure as defined in this paragraph belongs to the Type C of De Cesare’s taxonomy (see De Cesare in this volume). However, since Type C includes forms that are not considered in this paper, we will continue using the more traditional denomination, that is, Pseudo-cleft (PC).
8 See Moreno Cabrera (1999: 4249), contracción de la perífrasis; and Collins (1991: 31), non-cleft version. 9 Scheme adapted from Panunzi (2009: 1123). 10 On this issue, see Panunzi (2009: 1123) for Italian and Moreno Cabrera (1999: 4252–4281) for Spanish.
182
Rocío Agar Marco
2.2 Specificational function vs. predicative function In order to delimit more precisely the class of constructions dealt with in this study, it is useful to recall that there are syntactic and semantic structures which are superficially similar to PCs, but cannot be included in the same category since they are predicative.¹¹ These two examples from Salvi (1991) show the difference between a predicative structure in (12a) and a specificational structure in (12b). These two syntactic structures are similar in form, but differ in meaning: (12)
a.
Chi ha ucciso il maggiordomo è stupido. who has killed the butler is stupid ‘Who killed the butler is stupid.’ (Salvi 1991: 177)
b. Chi ha ucciso il maggiordomo è il who has killed the butler is the ‘Who killed the butler is the earl.’ (Salvi 1991: 177)
conte. earl
2.2.1. In Spanish, predicative structures may in some cases be distinguished from specificational structures thanks to the use of the subjunctive mood.¹² One function of the Spanish subjunctive is to express a conjecture, as in (13):¹³ (13)
Sp. “Es una realidad innegable que, o bien la soberanía is a fact undeniable that either the sovereignty popular […] se pone de acuerdo en este tema, o todo of the people agrees on this topic or everything lo que hagamos será agua de borrajas […]”. (elpais.com) what do-sbjv.1pl fut.3sg-be borage.water ‘“It is undeniable that, either the sovereignty of the people […] agrees on this issue, or everything we (will) do will be for nothing […]”’
11 This aspect has already been discussed in depth in many works, see for instance Moreno Cabrera (1999: 4262–4263) and Collins (1991: 37). 12 Besides the use of subjunctive, the Spanish language has other mechanisms to resolve ambiguities: note that, while Italian example (11a) can also be interpreted as predicative, the Spanish example (11b) can only be interpreted as having a specificational function. This is due to the repetition of the preposition a in the CC. 13 In examples (13) to (15), the verbal nucleus of the relative clause is highlighted in italics. The PCs or PC-like structures of examples (13) to (42) will be highlighted in bold when the structure appears within a broader context.
Pseudo-cleft sentences. Italian-Spanish in contrast
183
Todo lo que hagamos corresponds here to all we might hypothetically do. This use leads directly to a predicative reading of the utterance, the specificational reading consequently being ruled out (in other words, the utterance in [13] cannot be a PC). However, not all cases in which the Spanish subjunctive is used express a supposition. See for instance the utterance in (14a) below from the Spanish subcorpus: (14)
a.
Sp. “Es inaceptable” que quien se vaya a sentar en el is unacceptable that who go-sbjv.3sg to sit on the banquillo de los acusados sea el juez. (elpais.com) dock be-sbjv.3sg the judge ‘“It is unacceptable” that the one who is going to be in the dock is the judge.’
b. Sp. Quien se va a sentar en el banquillo de los acusados who go-ind.3sg to sit on the dock es el juez. be-ind.3sg the judge ‘Who is going to be in the dock is the judge.’ In (14a), the verbal nucleus of the relative clause in the PC, vaya, is in the subjunctive mood because it depends, in this case, on the subordinating conjunction que,¹⁴ not because it expresses a hypothesis, as in other cases.¹⁵ The rule allowing us to draw a distinction between cases which are to be interpreted as specificational sentences and cases which are to be interpreted as predicative sentences (i.e. PCs vs. simple copular sentences) appears to have an exception, i.e. the case in which the PC is located within a subordinate clause taking a subjunctive – such as the subordinate following the expression es inaceptable que… ‘it is unacceptable that…’ in (14a). The difference between the two uses of the Spanish subjunctive is clear in the pairs of sentences given in (14) and (15). There is no difference between the meaning of the PC in (14a), where an assumption is made using a subjunctive, and the meaning of the variant with the indicative mood, (14b). In (14a) there is an evaluation on the part of the speaker – es inaceptable – to the statement expressed in (14b), without modification of its meaning (the person who is going
14 Evidence for this is that both vaya and sea are in the subjunctive mood – not only the verbal nucleus of the relative clause of the PC, as it would be the case if the subjunctive did not depend on the expression es inaceptable que. 15 As can be observed, for instance, in (13) or in (15a).
184
Rocío Agar Marco
to be judged is the judge, both in [14a] and in [14b], regardless of the moral acceptability of the matter). Conversely, if we replace the subjunctive by the indicative mood in one of the other examples in which the subjunctive is employed to express an assumption, the PC resulting from this manipulation acquires a different meaning compared to the original sentence, as can be noted in (15): (15)
a.
Sp. “Lo que quiera expresar cada uno es lo what want-sbjv.3sg express-inf each one is the más importante”. (elpais.com) most important ‘Whatever one wants to express is the most important thing.’
b. Sp. Lo que quiere expresar cada uno es lo más what want-ind.3sg express each one is the most importante. important ‘What each one wants to express is the most important thing’. While in (15a) what is referred to as being most important is the fact that each one can express himself / herself freely, in (15b) what is ‘most important’ is the content to which lo refers cataphorically. As well as this difference in meaning, it should also be noted that, unlike (15a), (15b) can be read as a specificational sentence and, thus, as a PC (which, by contrast, is the case for both variants of [14]). Given these possible different uses of the subjunctive, we must be cautious in claiming that the verb in the relative clause of a PC can never be a subjunctive (see Moreno Cabrera 1999: 4292). As noted for (14a), there is indeed an exception to this rule, i.e. the cases in which the PC is embedded in a sentence structure that requires the subjunctive (such as es inaceptable que). We found some of these cases in our data. The above examples thus show that the use of the subjunctive in the precopular relative clause does not necessarily exclude that the sentence we are dealing with is a PC. 2.2.2. In the case of Italian, however, the criterion distinguishing specificational utterances from predicative utterances is solely semantic. In Italian, as well as in the case of Spanish sentences which do not follow the aforementioned rule, it is difficult to identify whether an utterance is to be interpreted as being predicative or specificational in the cases where we don’t have the possibility of considering
Pseudo-cleft sentences. Italian-Spanish in contrast
185
both the meaning of linguistic material surrounding the PC (as in [16] and [17]) and the meaning of the PC components, in particular of the CC (as in [18]):¹⁶ (16)
It. “Quello che dirò è niente rispetto a quello che vorrei dire” (repubblica.it) ‘“What I will say is nothing compared to what I would like to say”’
(17)
It. “Sicuramente quello che è arrivato dal referendum è un segnale .” (repubblica.it) ‘“Certainly what has come from the referendum is a signal ”’
(18)
Sp. Lo que hizo Matas, dijo, “es amoral aunque no sea delito”. (elpais. com)¹⁷ ‘What Matas did, he said, “is immoral although it is not a crime”.’
As well as the above-mentioned cases, in which only the predicative reading is possible, there are ambiguous constructions allowing for both a predicative and a specificational reading.¹⁸ In such cases, it is difficult to disambiguate the two readings even when considering the material surrounding the PC, as in (19): (19)
It. “[…] Va bene che chi lo ha messo lì è il presidente della Repubblica , che si piglierà la responsabilità di averlo fatto…”. (leggo.it) ‘[…] “It’s true that who put it there is the president of the Republic , who will take responsibility for having done so…”’
2.3 Types of Pseudo-cleft sentences Let us now take a closer look at the classification of PCs that has been taken into account in the present paper.
16 In examples (16)–(19) and (34)–(42), the postcopular element (whether it is a CC or not) is highlighted by a frame. 17 The structures in which the CC is an adjective (like in [18]) are rarely part of a specificational structure, because the most frequent role of the adjective is precisely that of assigning a quality to the noun. Exceptions to this tendency are utterances such as Como llegó Juan fue cansado, lit. ‘How Juan arrived was tired’ where cansado does not have a predicative value, but a specificational one (the example is taken from Moreno Cabrera 1999: 4264). 18 See Moreno Cabrera (1999: 4262–4263) and Birkner (2008: 319–324).
186
Rocío Agar Marco
As Table 1 shows, at a first level of analysis, PCs can be divided into two groups: wh-PCs and th-PCs. In the following, these two PC types will be analyzed in more detail, and an example will be given for each language under scrutiny in the present study. Table 1: Types of Pseudo-cleft sentences: introducers¹⁹ PC types wh-PC
th-PC
fused relative pronoun
(i)
(ii)²⁰
(iii)
pronoun
noun [+ generic]
noun [– generic]
It.
Sp.
It.
Sp.
It.
Sp.
It.
Sp.
chi ‘who’
quien²¹ ‘who’
persona ‘person’
donde ‘where’
lo ‘it’ el ‘the’ etc.
persona ‘person’
dove ‘where’
quello ‘that’ ciò ‘that’ colui ‘he’ etc.
luogo posto ‘place’
lugar sitio ‘place’
momento ‘moment’
momento ‘moment’
paese ‘country’ settore ‘sector’ problema ‘problem’ tema ‘topic’ lezione ‘lesson’ mezzo ‘means’
país ‘country’ sector ‘sector’ cuestión ‘question’ aspecto ‘aspect’ condición ‘condition’ movimiento ‘movement’
quanto cuanto ‘how much, ‘how much’ what’ quando ‘when’
cuando ‘when’
periodo ‘period’ volta ‘time’
vez ‘time’
19 In each case, we have searched the corpus for the noun only, without any determiners or modifiers. In the case of words like persona in Italian, whose inflected form for the grammatical number is obtained by modifying the end vowel only, we have searched the corpus for the form person-. When this kind of search was not possible, such as in the case of cuestión-cuestiones (with different accents), the corpus has been searched for each inflected form of the noun. 20 The list of nouns belonging to this group is based on Collins (1991: 26–30): it is the translation of the nouns that he considers equivalent to fused relative pronouns. As the frequency of such nouns is not always the same in Italian and Spanish, the lists of words proposed for each language are not fully equivalent. 21 In order to avoid overloading the content of Table 1, for all instances we have omitted the inflected gender and number forms, leaving it to the reader to take them into account here and in the rest of this paper.
Pseudo-cleft sentences. Italian-Spanish in contrast
187
PC types wh-PC
fused relative pronoun
th-PC (i)
(ii)²⁰
(iii)
pronoun
noun [+ generic]
noun [– generic]
come ‘how’
como ‘how’
modo ‘way’
forma modo manera ‘way’
perché ‘why’
por [*] que²² ‘why’
ragione ‘reason’
razón ‘reason’
motivo ‘cause’
motivo ‘cause’
cosa ‘thing’
cosa ‘thing’
*²³
*
opción decisione ‘option’ ‘decision’ elemento interés ‘element’ ‘interest’ messaggio imagen ‘message’ ‘image’ borsa ‘stock factura market’ ‘invoice’ partito ‘party’ gravamen etc. ‘tax’ etc.
wh-PCs PCs called wh-PCs in English coincide with classic pseudo-clefts,²⁴ namely those opened by fused relative pronouns, like chi ‘who’ or dove ‘where’ in Italian and like quien ‘who’ or cuando ‘when’ in Spanish: (20)
It. Ma dove Monti ha rivelato sapienza politica oltre ogni previsione, è stato nell’affrontare i dossier esplosivi Ici, patrimoniale, regole sui licenziamenti, pensioni – su ognuno dei quali ha alluso, senza mai esplicitarla, la soluzione finale. (lastampa.it) ‘But where Monti revealed political wisdom beyond all predictions, (it) was in handling critical files: town council tax, wealth tax, regula-
22 In this case, the equivalent of Italian perché in Spanish would not be por que, but por lo que (with all its possible forms: por el que, por la que etc). See also Moreno Cabrera (1999: 4273) for the cases in which the relative clause is preceded by a preposition. 23 Neither Italian nor Spanish have a fused relative pronoun corresponding to the English fused pronoun what. 24 As we have seen in Section 2.1, these are the prototypical structures, possibly accepted crosslinguistically as such on the basis of the English literature (see Section 5); see for instance Collins (1991: 26), who speaks about canonical pseudo-clefts.
188
Rocío Agar Marco
tion on dismissal, pensions – for each of which he mentioned, without ever making it explicit, the final solution’. (21)
Sp. Según dijo, quien garantiza el derecho de las partes es el juez, que es quien tiene que decidir qué parte de las conversaciones afectan al derecho de defensa. (elpais.com) ‘According to what he said, who grants the right of the parties is the judge, who is the one who must decide which parts of the conversations affect the rights of the defense.’
th-PCs The so-called th-PCs, on the contrary, are PCs opened by an NP followed by a relative clause. Within the group of th-PCs, three different subtypes can be distinguished: (i) th-PCs in which the head of the NP is a pronoun; in Italian, see quello, ciò ‘that’ etc., followed by relative pronouns che ‘that’ or cui / quale ‘whose / which’;²⁵ in Spanish, see lo, la, el ‘the’ etc., followed by the relative pronoun que ‘that’. (22)
It. Quello che non mi piace è l’aspetto business. (repubblica.it) ‘What I don’t like is the business side’
(23)
Sp. “Lo que procedía era declararlos nulos”, señaló el abogado, pero el instructor optó por concederles un nuevo trámite para que presentaran otro escrito. (elpais.com) ‘“What was appropriate was to declare them null”, the lawyer pointed out, however the examining magistrate chose to assign them a new procedure in order for them to present another file.’
(ii) th-PCs opened by nouns (shown in Table 1, column [ii]) that are semantically equivalent to the fused relative pronouns what, where, who, why, when and how (and aligned with them in Table 1) according to the typology proposed by Collins (1991: 27–31). According to Collins, the head of the NP in this type of PC can possibly be modified by numeral adjectives or quantifiers such as il primo ‘the first’, il solo ‘the only’ in Italian, or el primer ‘the first’, el único ‘the only’ in Spanish. Furthermore, we have included here some modifiers
25 With the whole paradigm of forms: per cui ‘for which / through which’, in cui ‘in which’ etc.
Pseudo-cleft sentences. Italian-Spanish in contrast
189
which were not taken into account in Collins’ work, such as superlatives, for instance Italian il più ‘the most’, la meno ‘the least’, or modifiers defined by Collins as being more complex, such as Spanish la mayoría de ‘the majority of’, uno de los… ‘one of the…’ etc. (24)
It. Le prime persone che ha voluto rivedere sono le sue cuginette (repubblica.it) ‘The first persons that she wanted to meet again were her little cousins’
(25)
It. “E l’altra cosa che avevo era la mia reputazione, il senso di dovere verso la giustizia”. (repubblica.it) ‘“And the other thing that I had was my reputation, the sense of duty towards justice”’
(26)
Sp. La última vez que el desempleo se situó en marcas similares fue en mayo de 2011, con un porcentaje del 84.1%. (elpais.com) ‘The last time that unemployment was at a comparable rate was in May 2011, with a rate of 84.1%.’
(27)
Sp. Otra de las cosas que están siendo estudiadas son las imágenes captadas por las cámaras de vigilancia del parque infantil al que el progenitor acudió en busca de ayuda. (elpais.com) ‘Another of the things that are being studied are the images captured by the surveillance cameras of the playground to which the parent went looking for help.’
The relative pronouns and the prepositions which follow the noun heads of the th-ii-PCs are, in Italian, che ‘that’, cui / quale ‘whose / which’ or a / ad ‘to’ (+ infinitive) and, more seldomly, di / per ‘of / in order to’ (+ infinitive). In Spanish, we find the relative pronoun que ‘that’ or the prepositions de / para ‘of / in order to’ (+ infinitive). (iii) th-PCs opened by nouns that are semantically richer than (ii), for instance Italian il settore ‘the sector’, la lezione ‘the lesson’ etc. and Spanish el sector ‘the sector’, el aspecto ‘the aspect’ etc. The relative pronouns and the prepositions following these NPs are the same as those in the th-ii-PCs. Below is an example for each language:
190
(28)
Rocío Agar Marco
It. “[…] Il settore che in agosto registra la diminuzione tendenziale più ampia è quello delle industrie tessili, abbigliamento, pelli e accessori (–10.1%)”. (ilsole24ore.com) ‘“The sector that has been recording the largest decreasing trend in August is that of textile, clothing, leather and accessories industries (–10.1%)”’
(29)
Sp. La única certeza que tienen los venezolanos cada vez que Chávez toma las pantallas de televisión es cuándo comienza su programa. (elpais.com) ‘The only certainty that the Venezuelans have each time Chávez takes possession of the television screens is when his program is going to start.’
3 Quantitative analysis of Pseudo-cleft sentences 3.1 Frequency of use of Pseudo-cleft sentences and their subtypes Based on the idea that Italian and Spanish are very similar languages (as we have seen in Section 2), we want to verify whether the same structure, i.e. the PC, is found with the same frequency in both languages. The quantitative data drawn from the ICOCP-corpus (cf. Section 1) shows surprising results concerning the PC’s relative frequency calculated for each language, both in terms of its overall distribution and of the distribution of each subtype. Interesting quantitative – and qualitative – differences emerge within one language, as well as crosslinguistically. Here are our results, which we will comment below. As can be observed on the basis of Table 2, we do not take into account PCs of type th-iii.²⁶
26 We haven’t searched for this type of PC mainly because they are difficult to find (semi)automatically. Although they are not taken into account in our quantitative analysis, we still take them into account in the qualitative analysis (see Section 4) because they have the same functions as the other PCs and because they allow us to refine our understanding of the textual functions of “classical” PCs.
Pseudo-cleft sentences. Italian-Spanish in contrast
191
Table 2: Frequency of use of Pseudo-cleft sentences per 100,000 words wh
th-i
th-ii
TOTAL
Italian
0.5
6
6
12.5
Spanish
3
22
10
35
3.1.1. Starting with the Italian sample, it can be noted that the total number of PC occurrences is not very high. Based on the literature (see for instance Berretta 1994: 88), in which PCs are generally considered to be more typical of spoken language, a frequency of 12.5 PCs per 100,000 words can be qualified as expected data. Our figure is indeed lower than the one found in a similar study on spoken Italian, that of Panunzi (2009), where the frequency of the construction is 20 per 100,000 words.²⁷ As far as the single types of PCs are concerned, it can be noted that, while the two th-types are comparable in terms of frequency, significant differences can be observed between the two th-types and the wh-PCs. This leads to a general observation concerning the literature: globally, in the works we have examined (cf. Berretta 1995: 161), the examples of PCs given are introduced indistinctly by chi ‘who’ or quello che ‘what’, especially when such examples are invented. These two PC introducers are thus considered to be equivalent, although their relative frequencies are not at all similar. 3.1.2. The situation is very different in Spanish. The total number of PCs is fairly high, in contrast with what is sometimes stated in the literature, where the construction is stated not to be very frequent in the written language. Most scholars, however, do not back their claims on corpus-based quantitative data nor base their quantitative analysis on data that is comparable to that collected in the present paper. Helfrich (2001), for instance, claims that PCs are not very frequent in literary writing. This claim cannot be verified here because of the nature of the data we take into account, which are based on journalistic writing. As regards the use of the various types of PC, Table 2 shows very different relative frequencies. The extremely high frequency of th-i-PCs (It. quello che, Sp. lo que ‘what’ etc.), for instance, suggests a possible lexicalization of the construction²⁸ whereas, as in Italian, very few occurrences of wh-PCs are to be found.
27 Panunzi (2009: 1130–1131) finds 61 occurrences of PCs in the Italian section of the C-ORALROM oral corpus, which contains approximately 300,000 words. 28 See Berretta (1995: 159) for the interrogative cleft.
192
Rocío Agar Marco
In Spanish too, no major distinctions are drawn in the literature between quien ‘who’ (wh-PC) and el que ‘the one who’ (th-PC) as introducers, which can even be found occurring together in the same example. In our data these two types show a very different distribution. 3.1.3. When comparing the two languages, it can first be observed that the total number of occurrences of PCs found in the Spanish subcorpus is almost three times higher than the total number of occurrences in the Italian subcorpus. The qualitative analysis provided in the next section will attempt to account for this difference. As regards the strategies used by each of the two languages, it can be noted that all types of PC are more frequently used in Spanish than in Italian. This shows that the PC is more frequent in Spanish regardless of the specific type of PC taken into account. Discrepancies between Italian and Spanish can be observed, concerning the distribution of the different subtypes of PCs within each language. For instance, in Italian the quantitative data about th-i-PCs is fairly similar to those about thii-PCs, whereas in Spanish the number of occurrences of th-i-PCs is more than twice as high as the number of th-ii-PCs. Conversely, the occurrences of wh-PCs both in Italian and in Spanish are few, and even less if we compare them to the th-i-PCs (we will come back to this point in Sections 3.2.1 and 5.1). Through a further fine-grained distinction between PCs of different types, i.e. introduced by different forms, in Section 3.2 we will be able to identify structures which are specific to each language.
3.2 Formal differences between Pseudo-cleft sentences and their subtypes The quantitative results that we have obtained suggest that there are clear-cut differences as regards the frequency of the PC in Italian and Spanish. In the following sections we will further analyze this data on the basis of the classification of PCs presented in Section 2.3. We aim to find a relationship between the differences in the form of the construction and the differences in its frequency. More precisely, we will focus our attention on the elements that introduce PCs in both languages. 3.2.1. Quantitative data drawn from the Italian subcorpus shows the existence of richer and more diverse forms than those described in previous research. This is due to the fact that our definition of PC is broader than the traditional definition.
Pseudo-cleft sentences. Italian-Spanish in contrast
193
Indeed, as well as the expected forms quello che and ciò che²⁹ ‘what’ (that are part of the th-i-PCs), we find the forms le prime persone che ‘the first persons who’, un’altra cosa che ‘another thing that’ etc. (that we added as th-ii-PCs). The most unexpected finding concerns wh-PCs, which are generally considered to occur frequently but appear only in two cases in our corpus (recall [19] and [20]). 3.2.2. Although Italian stands out for the large variety of its constructions, in an Italian-Spanish contrastive perspective it can be seen that Spanish has a wider variety of forms, some of which have no equivalent in Italian. Consider for instance: (30)
Sp. “Lo único que da miedo es que nos quedemos como estamos”, ha afirmado. (elpais.com) ‘“The only thing that scares me is that we stay as we are”, he stated.’
The form lo único que… ‘the only thing that…’ corresponds to a PC introduced by an article³⁰ followed by an adjective which in turn modifies an implicit noun (la única [cosa] que ‘the only [thing] that’). In these cases, Italian would need a full NP (l’unica cosa che ‘the only thing that’).³¹ In our data, this difference results in a much greater presence of th-ii-PCs with the noun cosa as head of the NP in Italian compared to Spanish. This indicates a partial internal balancing mechanism between different forms of th-ii-PCs in the two languages. In other words, each language has a different form to express the same meaning.³² 3.2.3. In the cases in which Italian and Spanish have identical forms, these forms do not necessarily appear with the same frequency. Thus, for instance, among th-ii-PCs, some are recurrent in Spanish but rare – although possible – in Italian.
29 See Berretta (1995: 161–162). 30 The status of lo is debated in the literature. According to the Real Academia Española (see Leonetti 1999: 829), lo belongs to the paradigm of definite articles, although it can never occur with a noun, since in Spanish there is no category of neutral nouns; following this assumption, lo should perhaps best be considered as a pronoun. For a discussion on the status of lo in Spanish, see for instance Leonetti (1999: 829–834) and Eguren (1999: 945–950). 31 A different case is that of the masculine singular article, which can be used within an elliptical construction where the noun modified by the adjective is implicit, both in Spanish (El único que tiene miedo es… ‘The only one who is afraid is…’) and in Italian (L’unico che ha paura è…). 32 Concretely, in our data concerning the PCs of the th-ii type, 36% of all occurrences in Spanish follow the scheme ‘the only thing + relative clause’ (see [30]). PCs in which the head of the NP is cosa ‘thing’ represent only 3% of the examples of th-ii-PCs found for Spanish, whereas in Italian this form is found in 30% of the occurrences of th-ii-PCs.
194
Rocío Agar Marco
That is the case for the structure of the type ‘NP + of / in order to + inf.’, which is illustrated in (31).³³ (31)
Sp. […] la única manera de garantizar la supervivencia del euro es que los estados miembros dejen atrás sus recelos a compartir soberanía. (elpais. com) ‘[…] the only way to grant the survival of the euro is that the member nations leave behind all hesitation about sharing sovereignty.’
The only occurrence of a structure of the type ‘NP + of / in order to + inf.’ found in the Italian subcorpus (see example [32] below) shows that the same structure is indeed possible in Italian as well:³⁴ (32)
It. “La mia rielezione sarà dura: l’economia sta uscendo da una recessione mondiale” dice Obama davanti agli invitati che hanno pagato biglietti fra i 10mila e i 35,800 dollari per essere presenti alla festa. “L’unico modo per essere rieletto – ha aggiunto – è fare in modo che tutti voi vi impegnate a questo scopo. […]” (ilsole24ore.com) ‘“It will be hard for me to be reelected: the economic system is just getting out of a period of world global recession” says Obama in front of the attendants who paid a fee ranging from 10,000 to 35,800 dollars to be able to attend the party. “The only way for me to be reelected – he added – is to act in such a way that you all make efforts in this direction. […]”’
A brief targeted search of another corpus of written Italian, the CORIS / CODIS corpus, confirms that the PC introduced by l’unico modo per / di is indeed used.³⁵ Its frequency, though, is similar to that which has been observed within our corpus, that is, very low.
33 33% of instances of th-ii-PCs in Spanish; 3% of instances of th-ii-PCs in Italian. 34 Note that in addition to being the only example of this form in Italian, this is also a translated text. It should also be recalled that th-iii-PCs are not part of our quantitative data. We will not, therefore, take into account the examples found within this subgroup. 35 We have searched the corpus for the cases in which unico ‘only’ and modo ‘way’ occur in close contact among the 20 million words of the subcorpus “Press”, in the CORIS / CODIS corpus. Of 164 total occurrences found, 70 were PCs containing l’unico modo per / di ‘the only way to / of’. That represents a frequency of 0.35 per 100,000 words.
Pseudo-cleft sentences. Italian-Spanish in contrast
195
3.3 A hypothesis on the quantitative differences between Italian and Spanish As observed above, particular forms of PCs might account for some of the differences in frequency which have been noted for the different types of PC in Italian and Spanish.³⁶ However, such accounts do not explain the wide difference of use between PCs in the two languages. In order to find a better account of the quantitative differences between both languages PCs it is necessary to extend our analysis beyond the boundaries of the PC. A second hypothesis that might allow for a more satisfying explanation of the huge quantitative difference between the data for Italian and for Spanish is based on another type of cleft construction which is very similar to the PC (and that is in fact proposed to belong to the same category by De Cesare in this volume). The construction in question is the reversed cleft (It. a dire questo è il presidente ‘who says this is the president’), in which the most salient information is at the end of the utterance, just as in PCs.³⁷ In our reference corpus (see De Cesare et al. in this volume), the reversed cleft, which exists only in Italian, has a frequency of 24 per 100,000 words. This distribution is very similar to the difference between Italian (12.5) and Spanish (35) PCs, as observed in Table 2 (i.e., 23). Our hypothesis, thus, is that a form of rebalancing emerges between the reversed cleft in Italian and PC in Spanish. This tentative hypothesis can be partly confirmed by translation data. In translation, reversed clefts and pseudo-clefts seem interchangeable in many cases.³⁸ In order to support our claims, let us consider example (33), which illustrates the translation into Spanish of an Italian reversed cleft through a PC:
36 Among the examples, other specific forms are found, which nevertheless represent isolated cases. Such is the case for Italian ‘NP + da + inf.’, as in la prima cosa da fare ‘the first thing to do’. 37 The reversed cleft and the PC are indeed considered to be the same structure in some studies (see Gil 2004, De Cesare 2005), as in many cases they are used in an analogous manner. Nevertheless, as these structures differ on a syntactic level, the reversed cleft will be considered here as a distinct construction, in line with Roggia (2009) or De Cesare (2011). 38 Of course, the best way of verifying the equivalence between the two constructions would be to analyze a translation corpus (in line with the methodology used in Brianti in this volume as well as Atayan & Wienen in this volume). In should be noted that translation represents a useful tool for the analysis of the syntactic and functional equivalences between languages, although the communicative spontaneity which only original language corpora can provide is often lost.
196
(33)
Rocío Agar Marco
a.
It.
A parlarmi è Maurizio Prestieri, boss camorrista to speak.inf-me.dat is Maurizio Prestieri camorra boss del Rione Monterosa a Secondigliano. (repubblica.it) of.the Rione Monterosa in Secondigliano ‘Who is talking to me is Maurizio Prestieri, camorra boss of the Rione [neighborhood] Monterosa in Secondigliano.’
b. Sp. Quien me habla es Maurizio Prestieri, jefe camorrista who me.dat speaks is Maurizio Prestieri camorra boss del Rione (barrio) Monterosa, en Secondigliano […]. (elpais.com) of.the Rione (neighborhood) Monterosa in Secondigliano ‘Who is talking to me is Maurizio Prestieri, camorra boss of the Rione (neighborhood) Monterosa in Secondigliano.’ In our view, this example can be interpreted as a sign that, a priori, the hypothesis of rebalancing could be plausible.
4 Functions of the cleft constituent in Pseudo-cleft sentences In the previous section of this study we have seen that PCs can have different frequencies in Italian and Spanish. In this section, we will verify whether the same form, appearing in two similar languages, also has the same functions across these languages.³⁹ Specifically, our aim is to discuss the textual functions of PCs by focusing on the CC. In PCs, the CC is the element saturating the variable opened by the specificational structure in the first part of the construction. Moreover, it is located at the end of the sentence. These two facts allow the claim that the CC coincides with the utterance Focus.⁴⁰ As such, “just as the Focus of an unmarked utterance, the Focus of the PC tends to carry new information, which is located at the end of the utterance” (De Cesare 2005: 299; the translations is
39 For our functional analysis, we have considered all types of PCs in Table 1 (including the PCs introduced by less generic NPs, i.e. th-iii-PCs). 40 In this study, we adopt the definition of Focus of Ferrari et al. (2008), id est: “L’unità nucleare [dell’enunciato] è caratterizzata da un ‘Focus Informativo’, cioè da quella sottocomponente del Nucleo che è specificamente responsabile del valore illocutivo-testuale di questo”, ‘The nuclear Unit [of the Utterance] is characterized by an “Informative Focus”, e.g. by a subpart of the Nucleus that is responsible of its illocutional and textual value’.
Pseudo-cleft sentences. Italian-Spanish in contrast
197
ours). Compared to that of an unmarked sentence, however, the Focus of a PC has specific communicative and textual properties, which we will discuss in the next sections. In order to better understand such properties, we will firstly consider the subsequent co-text of the construction (whether the CC is recovered or not, ways of recovering the CC, discourse functions of the CC, etc.), an aspect on which there has been little research in the literature.
4.1 Ways of recovering the cleft constituent in the subsequent text Firstly, we ought to specify what we will count as “recovered CC”. In our analysis, a CC is recovered within the subsequent co-text either explicitly, through the same phrase or through a pronoun, or implicitly, resumed in the form of an implicit subject (be it within the same utterance or within one or more consecutive utterances). There are of course other forms of CC recovery (for instance in textual movements such as exemplification or specification, which are referentially connected to the CC). However, because they require a further fine-grained description, we will not take these forms of CC recovery into account in our study.⁴¹ We will focus our attention solely on the cases in which referential continuity is explicit or implicit (through an implicit subject). Table 3: Percentages of the cleft constituent in terms of recovery per total PCs⁴² RECOVERED CC
NON-RECOVERED CC
Italian
51.7%
48.3%
Spanish
55.8%
44.2%
The most interesting data emerging from Table 3 is that the CC is recovered in the subsequent co-text in over 50% of all occurrences both in Italian and in Spanish. This data, together with the strong tendency of the CC to carry new information (cf. Berretta 1994: 89 and 2002: 20, as well as De Cesare 2005: 299⁴³), suggests that the PC is likely to have a so-called presentative function, i.e. to introduce a referent which is not only new, but also salient and recovered within the text typically
41 See Baranzini (in this volume, Section 5) for a more thorough investigation of this aspect. 42 As shown in Baranzini (in this volume), the CC can be recovered as a whole or only in part. 43 Gil (2004: 381) also highlights the fact that a PC is particularly suitable for introducing new and salient information into the text.
198
Rocío Agar Marco
as discourse Topic;⁴⁴ such a function has already been noted by Hetzron (1975) for the English PC: “When the focusing is necessary for paving the way for later use of the same element in the discourse or for a pragmatic reaction, the cataphoric construction that moves the focused element to the end of the sentence [PC] is created. This latter type of justification is the presentative function” (1975: 364; the highlighted words are from the author). A PC can be used for other reasons as well. As well as its presentative function, we will see that in some occurrences of PCs the CC allows for reactivating some referents which had been previously abandoned in favor of others. These and other textual functions are studied in depth in De Cesare’s study (2005: 306– 318), which considers different types of Cleft constructions (PCs as well as the so-called reversed clefts). We shall proceed with the same aim in the next section, but we will focus our attention on occurrences of PCs only.
4.2 Informational and textual functions of Pseudo-cleft sentences As mentioned above, our starting point is De Cesare’s (2005) work, in particular where it is claimed that “in many cases […] the textual referent introduced by the Focus of a PC is characterized by semantic and pragmatic saliency, which can be measured mostly by considering the thematic organization level within the text” (2005: 310; the translation is ours).⁴⁵ The role played by the CC of a PC within the thematic structure of the text⁴⁶ is determined by taking into account the different informational and textual functions of the CC. Table 4 below summarizes some of the different functions that the CC can have with respect to the subsequent discourse. Further functions (such as serving as a linking element between paragraphs as a means of textual cohesion) can be identified if one takes into account a different level of analysis. These other
44 For a definition of Topic, in the sense of “aboutness”, see Lambrecht (1994). The presentative function can be related to what has been called in the literature Topic Launching for instance by Johansson 2002, as we see in De Cesare (2005: 311–312): “The Focus of a PC can also be associated with the function that Johansson 2002 has called Topic Launching, according to which the textual referent activated by the CC is recovered in subsequent utterances under the form of a ‘Topical item’” (the translation is ours). 45 See also Hetzron (1975), Furukawa (1994), Gil (2004) and Ferrari (2008). 46 According to Ferrari et al. (2008), text organization has at least two dimensions: the logical dimension and the thematic dimension. The latter one concerns the way in which the text referents of the different Units connect.
Pseudo-cleft sentences. Italian-Spanish in contrast
199
functions, however, can overlap with the functions in Table 4 and will thus not be described in this paper in order to simplify this first proposal of classification. Table 4: Functions and properties of the cleft constituent in a Pseudo-cleft sentence Recovery FUNCTIONS
(Re-)Introduce the Topic Newness
long distance of the whole text new
given
short distance
of the paragraph new
given
of one or two utterances new
In light of the data given in Table 4, it can be stated that the CC of the PC can introduce (or reintroduce) into the text the element representing the central Topic of the discourse in which the PC is inserted. The range of such discourse differs greatly according to the situation: it sometimes coincides with the whole text (4.2.1), whereas in other cases it is limited to one paragraph (4.2.2) or even only to one or two utterances (4.2.3). Let us now focus our attention on each of these cases, and mention their use in the languages we have taken into account in the present study. 4.2.1. The focused element of a PC, which is usually referential, can be used to introduce or reintroduce the central Topic of the whole text (which has often already been announced in the title or the subtitle of the article). The PC introducing this type of CC has different characteristics that depend on their position within the text, as we will see in more detail under a) and b) below. a) When opening the text, the PC, through its CC, introduces the discourse Topic for the first time (not considering the title) in the form of a new referent. The PC can thus be described in this case as having a presentative function. While the Italian section of the ICOCP-corpus contains several occurrences of this case (see [34] below), such a function has not been found in the Spanish corpus: (34)
la prima in Italia Gay e lesbiche, arriva l’agenzia matrimoniale Aiuta gli omosessuali “insospettabili” a trovare l’anima gemella MILANO – Uno le prova tutte per trovare l’anima gemella. Ma quando i metodi tradizionali non funzionano, il tempo stringe e la solitudine aumenta, quello che va ancora per la maggiore è la vecchia agenzia matrimoniale . Ora rinnovata, “moderna”, come racconta la responsabile della prima in Italia solo per gay e lesbiche.
200
Rocío Agar Marco
PER INSOSPETTABILI – Quello che l’agenzia matrimoniale (o meglio “Club per single”, secondo la nuova moda) “Il Delfino” fa, non è per tutti. [ null subject ] Aiuta gli omosessuali a trovare un compagno e, se la ricerca va a buon fine, ad accompagnarli all’altare. L’ agenzia è la prima in Italia diretta solo a lesbiche e a gay […]. L’ agenzia organizza anche sedute di coaching e incontri “SosSolutudine”, per aiutare chi ancora non ha fatto coming out. (corriere.it) ‘The first in Italy Gays and lesbians, a dedicated marriage agency is now available Helps “unsuspected” homosexuals to find soul mate MILAN – People would try anything to find their soul mate. But when traditional methods do not work, time is pressing and loneliness gets overwhelming, what is still most popular is the good old marriage agency . Now updated, “modern”, as tells the director of the first one in Italy for gays and lesbians only. FOR THE UNSUSPECTED – What “Il Delfino” dating agency (or rather “Singles Club”, according to the new fashion) does, is not for everyone. It helps homosexuals find a mate and, if the search is successful, it leads them to the altar. The agency is the first in Italy intended for lesbians and gays only […]. The agency also organizes coaching sessions and “Sos Solitude” meetings to help those who still haven’t come out.’ As can be noted, the referent l’agenzia matrimoniale ‘the dating agency’ is introduced by a PC as early as in the second utterance of the first paragraph; it then becomes the Topic not only of that paragraph, but of the whole text. Evidence of the fact that the focused element of the PC coincides with the most salient information of the discourse is provided by the subsequent co-text (which has not been fully reproduced here for reasons of space), where the CC is resumed nine times, almost in every case through the same NP. That this NP is already present in the title of the article is a further indication that it is indeed the central Topic of the text. The fact that this type of PC has not been detected in any of the occurrences within the Spanish corpus might simply be due to the small size of our corpus. However, since the use of Cleft constructions in general is often said to be a matter
Pseudo-cleft sentences. Italian-Spanish in contrast
201
of stylistic choice,⁴⁷ it is also plausible, in our view, that preferences related to journalistic writing in each language may be at the origin of this difference (see Garassino in this volume for a discussion of Italian and English). b) When it is not located at the beginning of an article, the PC can be used to reintroduce the main Topic of the text after it has been temporarily abandoned in favor of other marginal Topics. This type of PC typically opens a macro-portion of text, such as the paragraph, after a section in which a temporary shift, from the main discourse Topic to a different, less central Topic, had taken place. Let us analyze example (35) below, taken from the Italian subcorpus: (35)
Tabarez, momenti di terrore danno fuoco alla moglie Due sconosciuti aggrediscono la donna sulla porta di casa: le gettano liquido infiammabile e accendono il fuoco con un accendino. Ustioni su tutto il corpo MONTEVIDEO – La moglie del ct della Nazionale di calcio uruguayana, Oscar Tabarez, è stata vittima di una “aggressione selvaggia” – così l’ha definita il quotidiano El Pais – che le ha provocato ustioni al volto e sul 25% del corpo, ieri pomeriggio, sulla porta di casa, nell’elegante quartiere di Carrasco, a Montevideo. […] Non si esclude che all’origine dell’aggressione possa esserci una vendetta dell’ex domesticaTopic, […] Tuttavia il portavoce della polizia di Montevideo, l’ispettore Luis RondanTopic, fa sapere che al momento l’indagine rimane aperta, anche perchè “non ci sono testimoni, nè filmati, nè una descrizione degli aggressori. E l’unica che può rivelare particolari utili è la vittima , che fino ad ora non è stata in condizioni di parlare, e che già in passato avrebbe subito minacce”. Due mesi fa infatti la moglie del ct avrebbe trovato cenere ed i resti di un piccolo incendio sotto una finestra di casa, ed aveva denunciato il fatto alla polizia. (repubblica.it) ‘Tabarez, moments of terror His wife set on fire Two strangers attack the woman on her doorstep: they throw flammable liquid at her and light the fire with a lighter. Burns all over her body. MONTEVIDEO – The wife of Uruguay national soccer team coach, Oscar
47 Basically, Johansson (2001: 551) claims that the use of Cleft constructions is related to parameters such as “persönlich / unpersönlich” (‘personal / impersonal’), “formell / informell” (‘formal / informal’) (this observation and the terms are from Wienen 2006: 239).
202
Rocío Agar Marco
Tabarez, was the victim of “brutal aggression” – as the newspaper El Pais defined it – which caused her to suffer burns on her face and on 25% of her body, yesterday afternoon, on her doorstep, in the elegant neighborhood of Carrasco, Montevideo. […] It is not excluded that the cause of this aggression could be an act of revenge by a former female domestic employeeTopic […] However, Montevideo police spokesman, Inspector Luis RondanTopic, announces that, for now, the investigation remains open, also because “there are no witnesses, nor recordings on camera, nor a description of the attackers. And the only one who can reveal useful details is the victim , who up to the present moment has not been able to talk, and who apparently had previously received threats”. Two months ago the coach’s wife had reportedly found ashes and traces of a small fire underneath one of her house windows, and had reported the matter to the police.’ The referent La moglie di Tabarez, ‘Tabarez’ wife’, which represents the central Topic of the text (and, as such, is resumed many times throughout the article and already appears in the title), is reintroduced within a PC after having been temporarily abandoned in favor of other, less central Topics:⁴⁸ the domestic employee and Inspector Luis Rondan. At the same time, it may be noted that these Topics, less relevant from a pragmatic point of view, are in fact all cognitively related to the central Topic of the text, namely to Tabarez’ wife. Contrary to what we observed in the case of a), in which the CC of the PC introduces the main Topic for the first time, the type of PC described in b), in which the CC of the PC resumes the main Topic after it has been abandoned in favor of other Topics, is found in both languages within our corpus. Let us now examine one of the many examples found in the Spanish corpus: (36)
El Barça se airea Guardiola levanta el veto a sus futbolistas, con la agenda repleta de actos “No hay entrevistas”, dijo Guardiola pocos días antes de ganar la Supercopa de España al Madrid. El motivo por el que [ null subject ] prohibió a sus jugadores charlas individuales con los medios y actos sociales no está claro, pero lo cierto es que los jugadores del Barcelona llevan desde
48 Alternative Topics will be highlighted in (35), (36) and (38) with italics and the label “Topic” in subscript.
Pseudo-cleft sentences. Italian-Spanish in contrast
203
el inicio del curso bajo la ley impuesta por su entrenador , que no da explicaciones: “Lo he decidido así”, [ null subject ] dijo un día. Según las fuentes, varían las razones: “[ null subject ] Pensó que el inicio de la temporada era un momento complicado, [ null subject ] les necesitaba centrados”, admite el entorno. “Jugábamos contra el Madrid, veníamos de un final de temporada muy duro, y [ null subject ] quería protegerles desactivando altavoces”, convienen otros. […] El técnico , aprovechando los primeros días del año sin partido entre semana, abre la veda. Y desde el domingo, al término del partido ante el Valencia, los jugadoresTopic pueden hablar con quien quieran y atender actos sociales, publicitarios o benéficos, que de todo tienen: el lunes XaviTopic acompañará a Isidro Fainé, presidente de La Caixa y de su Fundación, ante una decena de niños beneficiarios del programa de superación de la pobreza infantil; y el miércoles acudirá a un acto con Garmin, empresa de GPS. También el lunes PiquéTopic presenta la colección de verano de Mango e IniestaTopic tiene otra campaña publicitaria […]; en la misma semana MascheranoTopic tiene una grabación con Nike, ThiagoTopic un acto en la tienda del Barça… y así, casi todos. Mientras, a quien espera el club es a Guardiola . Desde que empezó el año raro es el viernes que alguien no anuncia que a la semana siguiente Guardiola comunicará al presidente si acepta la renovación o cierra cuatro años extraordinarios. […] (elpais.com) ‘Barça gets some air Guardiola lifts the veto on his players, whose agenda is filled with events “No interviews”, said Guardiola few days before winning Spain Supercopa against Madrid. The reason why he forbade his players individual talks with the media and social events is not clear, but what is certain is that, since the beginning of the season, the players of Barcelona have been subject to this rule imposed by their coach , who gives no explanation for it: “ I decided it would be that way”, he said one day. The reasons thereof vary according to different sources: “ he thought that the start of the season was a difficult moment, he needed them to be focused”, members of his entourage admit. “We were playing against Madrid, were just back from a very hard end of season, and he wanted to protect them by having all speakers turned off”, others agree. […]
204
Rocío Agar Marco
The trainer , taking advantage of the first days of the year without matches, opens the way. And since last Sunday, after the end of the match against Valencia, his playersTopic have been allowed to talk to whoever they wish and attend social, advertising or charity events of all kinds: next Monday XaviTopic will be accompanying Isidro Fainé, Chairman of La Caixa and its Foundation, visiting a dozen of children who benefit from a program against child poverty; and on Wednesday he will attend an event with Garmin, a GPS company. Also next Monday PiquéTopic is planned to do the opening of Mango summer collection and IniestaTopic is to star in another advertising campaign […]; in the same week, MascheranoTopic will attend a recording session with Nike, ThiagoTopic is to attend an event in the Barça store… and so it goes for almost all of them. In the meantime, the one the club is waiting for is Guardiola . Since the beginning of the year, there has been hardly any Friday in which it hasn’t been announced that the following week Guardiola would inform the president whether he is accepting the renewal or he is going to put an end to four extraordinary years.’ As usual, the subtitle of the article (“Guardiola lifts the veto on his players, whose agenda is filled with events”) already provides the reader with information on the text’s main Topic: Guardiola. The Catalan team coach is the element which the greatest part of the text is about and as such is resumed many times in the form of an implicit or explicit subject (resumed by the same NP or through synonyms; cf. el entrenador ‘the trainer’, el técnico ‘the coach’), also functioning as a source of reported speech. Only in the third paragraph, after the first utterance which is once again about Guardiola, a new referent is introduced: los jugadores ‘the players’. Xavi, Piqué, Iniesta, Mascherano, Thiago are all secondary discourse Topics, which relegate the main Topic to the background. In this case, it is precisely a PC that recovers the temporarily abandoned Topic: the CC reintroduces Guardiola into the text in the form of a proper noun; the reader’s attention is thus focused back on the main Topic. 4.2.2. In most of the cases observed in both languages, the CC of a PC is not used to (re-)introduce the central Topic of the whole text, but rather the central Topic of one or more consecutive paragraphs, as can be seen for instance in (37): (37)
El Madrid se embarca hoy hacia Moscú, donde el martes se medirá al CSK en la ida de los octavos de final de la Liga de Campeones. Mourinho anunció que la competición europea pasa a ser su prioridad. […]
Pseudo-cleft sentences. Italian-Spanish in contrast
205
“Un buen modo de ir a Moscú es con felicidad ”, observó el entrenador. “Y la felicidad viene como consecuencia de los resultados y del trabajo bien hecho. Nos queda un largo camino por delante en la ‘Champions’ pero la felicidad hace milagros y la gente está feliz . Viajaremos a Rusia como un grupo contento . La motivación es alta para enfrentar al CSKA y a las difíciles condiciones que nos esperan allí”. Toño, el portero del Racing, debió hacer unas 10 paradas […] (elpais.com) ‘Today Real Madrid team is embarking on a trip to Moscow, where, next Tuesday, it will be facing CSK in the first leg of the second round of the Champions League. Mourinho announced that the European competition is now his priority. […]’ “A good way of travelling to Moscow is with happiness ” the coach observed. “And happiness comes as a consequence of results and of a well-done job.” We have a long road ahead in the ‘Champions League’ but happiness makes miracles, and people are happy . We will travel to Russia as a happy group. The motivation is high in facing CSKA and the harsh conditions awaiting us there". Toño, the Racing goalkeeper, must have done about 10 stops […]’ Being located at the beginning of a paragraph in which a secondary Topic is to be introduced, the PC in (37) draws our attention to a change of Topic occurring through the referent coded by the CC: con felicidad ‘with happiness’. Such a Topic change only concerns the paragraph in which the construction occurs: the reader’s attention is shifted from the secondary Topic of the preceding paragraph (the trip to Moscow of the Madrid team) to a new secondary Topic (con felicidad). Going back to what had been stated about (35), here too it can be observed that this Topic, less relevant from a pragmatic point of view, is in fact subordinate to the central Topic of the text, the Real Madrid team. Note also a further peculiarity of (37): the referent introduced in the text through the CC of the PC is not a human, concrete referent as in the previous cases, but a non-human, abstract referent, a quality. This example documents the variety of forms that the CC may take when introduced by the PC. Within the group of PCs that introduce the Topic of one or several consecutive paragraphs, there are also examples of another type: the referent introduced in the text by the CC of a PC can be used to create a shift from the central Topic of the text to a secondary Topic of the discourse. This function is clear in example (38): (38)
Nel 1960, all’età di 15 anni, John HepworthTopic entrò nel Seminario di San Francesco Saverio di Adelaide. Gli abusi, [null subject]Topic ha detto,
206
Rocío Agar Marco
sono iniziati un mese dopo il suo ingresso. Allora [null subject]Topic venne minacciato di espulsione se [null subject]Topic avesse parlato di questi incontri sessuali forzati. Colui che abusò di lui all’inizio era John Stockdale , all’epoca un seminarista più grande. Il sacerdote , morto a 57 anni in un club per soli uomini, è stato accusato di abusi sessuali da molti chierichetti che servivano nelle sue parrocchie. HepworthTopic ha detto che dopo due anni […] (zenit.org) ‘In 1960, at the age of 15, John HepworthTopic entered the Seminary of Saint Francis Xavier in Adelaide. The abuse, he Topic said, began a month after he arrived. At the time he Topic was threatened with expulsion if heTopic spoke of these forced sexual encounters. Who abused him in the beginning was John Stockdale , an older seminarist at the time. The priest , who died at 57 in a club for men only, was accused of abuse by many altar boys who were serving Mass in his parish. HepworthTopic said that after two years […]’ In (38), the central Topic of the article is John Hepworth, who has been victim of sexual abuse by John Stockdale. As can be observed, the writer exploits the specificational meaning of the PC to indicate a temporary shift from the Topic of the whole text (John Hepworth) to the Topic of the paragraph in which the PC occurs, namely John Stockdale. Now, it is clear that the referent John Stockdale answers a possible question from the reader (who abused John Hepworth?), which is itself suggested within the text by an utterance mentioning the abuse but not mentioning the abuser, and thus raising the question. The question is resumed in the first part of the relative clause of the PC and is then answered through the referent of the CC. While the Topic is secondary, it is salient, in that it is part of one of the most important phases in the life of John Hepworth. A PC involved in a Topic change of this type – from the main Topic to a secondary one – can be used systematically in a text to introduce new Topics and it thus plays an important role in text organization. The CC of a PC, both in Italian and in Spanish, contributes to the distribution of Topics within different paragraphs. In our corpus it is not uncommon to find several specificational constructions introducing the most salient referent of each paragraph.⁴⁹ As specificational 49 It should be remembered that according to our broad definition, PCs are specificational structures (see Section 2.1). Within the class of “specificational structures”, we also find the type that has no relative clause in its pre-copular segment (another one is Lorenzo Cola vs. another one who is revealing the most secret mechanisms is Lorenzo Cola). In this study, we do not take into account specificational structures lacking a relative clause.
Pseudo-cleft sentences. Italian-Spanish in contrast
207
constructions, PCs participate in the organization of texts, notably in terms of their thematic organization. The example which shows most clearly the importance of the PC in text organization is undoubtedly (39): (39)
Finmeccanica, spuntano i verbali: “Tangenti, Guarguaglini sapeva” L’ex consulente Borgogni accusa il presidente. Spuntano le assunzioni per i figli dei politici Ci sono almeno tre pentiti , dietro l’inchiesta sulle tangenti EnavFinmeccanica. […] Sta collaborando infatti Lorenzo Borgogni , il manager che teneva i contatti con la politica per conto di Finmeccanica. [ null subject ] Racconta di piccoli clientelismi quale l’assunzione della figlia dell’ex deputato di Forza Italia, Ilario Floresta, attuale membro cda di Enav. Ma sa molto di più. “Vi è una conversazione telefonica intercettata dalla quale si evince con solare evidenza che il ruolo di Lorenzo Borgogni dentro Finmeccanica fosse anche quello di occuparsi di contribuzioni illecite al sistema dei partiti” […] Un altro che sta svelando i meccanismi più segreti è Lorenzo Cola , l’ex consulente personale del presidente di Finmeccanica, Pier Lorenzo Guarguaglini. “Guarguaglini – dice Cola in un verbale di importanza capitale – autorizzava tali operazioni. […]”. Cola qui riferisce di un versamento da 300mila euro […] Nel dettaglio, uno che si è prestato a tutto, ma ora s’è pentito, è il costruttore Tommaso Di Lernia . Devastanti i suoi racconti. […] Ma oltre al racconto di Di Lernia , c’è pure quello di Lorenzo Cola , che sa di un incontro carbonaro avvenuto a Roma, presso l’Harry’s Bar di via Veneto, tra Matteoli, Pugliesi e Tulliani (di Optimatica). “Peraltro – dice Cola – Matteoli all’epoca sponsorizzava Martini per la nomina a presidente di Enav perché debitore di un favore […]”. (lastampa.it) ‘Finmeccanica, minutes now available: “Bribery, Guarguaglini knew” Former consultant Borgogni accuses the chairman and reports recruitment of children of politicians There are at least three informants behind the inquiry on the EnavFinmeccanica bribery scandal. […] Lorenzo Borgogni , the manager in charge of political relations on behalf of Finmeccanica, is indeed collaborating. He is talking about patronage on a small scale such as the hiring of the daughter of the former Forza Italia member of parliament, Ilario Floresta,
208
Rocío Agar Marco
current member of the board of Enav. But he knows much more. “There is an intercepted telephone conversation which very clearly leads to the assumption that the role of Lorenzo Borgogni inside Finmeccanica was also that of being in charge of illegal funding to the system of political parties” […] Another one who is revealing the most secret mechanisms is Lorenzo Cola , former personal consultant for Finmeccanica’s chairman, Pier Lorenzo Guarguaglini. “Guarguaglini – says Cola in a statement of paramount importance – allowed these activities.” In his statement Cola reports a payment of 300’000 Euros. In more detail, one who has consented to everything, but now has repented, is the builder Tommaso Di Lernia . Ravaging, his stories. […] But in addition to the account by Di Lernia , there is also that by Lorenzo Cola , who knows about a clandestine meeting in Rome, near Harry's Bar in via Veneto, among Matteoli, Pugliesi and Tulliani (from Optimatica). “Moreover – says Cola – at that time, Matteoli was sponsoring Martini to be appointed as president of Enav because he owned him a favour. […]’ The text opens with the statement “There are at least three informants”. Such informants turn out to be precisely the most important referents, introduced in several passages within the text and subsequently recovered through various devices: the textual referent Lorenzo Borgogni is presented in the first paragraph, Lorenzo Cola is introduced in the second, and Tommaso Di Lernia in the fourth. What should be noted in this article is the fact that for introducing two of the three informants, the same structure is used, and this structure is a PC. After its first occurrence in a PC, the referent Cola is recovered several times as a Topic and as the source of the reported speech. The same holds true for Tommaso Di Lernia. This referent is not entirely new when appearing in the PC.⁵⁰ However, it is highlighted because it occurs in this construction, and it is thus reactivated, so that it is also the source of the quotes in the fourth paragraph. This referent will be further recovered as the central Topic of other paragraphs. Let us have a look at one last example from Spanish and further analyze the characteristics of the function observed for Italian in (39): (40)
En lo que sí están mejor las cajas es en su cobertura . Los procesos de reestructuración y de integración y el dinero público recibido han permi-
50 It has in fact already been mentioned in the third paragraph, which we haven’t quoted here for lack of space.
Pseudo-cleft sentences. Italian-Spanish in contrast
209
tido a las cajas de ahorros dotar provisiones con cargo a fondos propios (es decir, sin incluirlas en la cuenta de resultados) por unos 26,000 millones de euros. Así, las cajas tienen unas coberturas de sus activos problemáticos que se sitúan en torno al 30%, frente al 26% de los bancos, y además han reconstituido sus provisiones genéricas, que cubren el conjunto de los riesgos y no solo el inmobiliario. Con todo, incluso al tomar solo en cuenta el riesgo no cubierto , la conclusión sigue siendo la misma: el ladrillo pesa más en las cajas. (elpais.com) ‘Where savings banks do perform better is in their hedging . Restructuring and integration processes as well as the public money received allowed savings banks to make provisions granted on their own funds (that is to say, without including them in their profit and loss accounts) for about 26.000 million Euros. Thus, savings banks dispose of coverage for their troubled assets for around 30%, as opposed to 26% of banks, and in addition they have reconstituted their generic provisions, which cover the totality of risks and not only real estate risk. Nevertheless, even considering only the uncovered risk, the conclusion is always the same: brick weighs more in savings banks.’ In the article from which this excerpt is taken,⁵¹ there is a discussion of the features of Spanish saving banks as compared to private banks, weighing up the pros and cons of both. Just as in the Italian example, the journalist shows a tendency to introduce each salient referent (in this case, each feature of savings banks) in a separate paragraph, which is often opened by a specificational construction.⁵² Furthermore, in this case as well as in the Italian one, it is precisely a PC that introduces one of these referents, namely the hedging of banking institutions, which will then become the main Topic of the whole paragraph. Based on the examples found in our corpus (cf. [39] and [40]), it can be concluded that there is a certain systematicity in the way the PC is used to organize information within the text, especially as noted above as regards the introduction of the most salient discourse referents. As it can be observed from Table 4, the CC of a PC that introduces the central Topic of one or more consecutive paragraphs can also be discourse-given. For reasons of space, we will not provide a detailed description of these cases. Example (41) shows that the function of a PC with a discourse-given CC is similar to the functions described above for PCs with discourse-new CC:
51 For reasons of conciseness, only this passage is quoted here. 52 See note 49.
210
(41)
Rocío Agar Marco
In altri Paesi avanzati, invece, i finanziamenti alla ricerca sono erogati da un’agenzia indipendente […]. Se vogliamo davvero rendere più efficaci le procedure di erogazione dei finanziamenti alla ricerca, la prima cosa da fare è attribuire questo compito a un’agenzia indipendente […]. (ilsole24ore.com) ‘In other advanced countries, instead, research funding is granted by an independent agency […]. If we really want to make research funding procedures more efficient, the first thing to do is assign this task to an independent agency […].’
4.2.3. In both Italian and Spanish, the most frequent function of the CC of a PC is to introduce a referent which is salient in one or two following utterances of the text. Consider this example from the Italian subcorpus: (42)
LE ASSOCIAZIONI – L’insolito black-out sulla rete ha messo subito sull’attenti le associazioni di consumatori. La più lesta a muoversi è stata Adiconsum , che ha preannunciato un’iniziativa in tutela dei clienti interessati. Dice Pietro Giordano, segretario generale: […]. Dello stesso avviso Adoc, come conferma il presidente Carlo Pileri: […]. (corriere.it) ‘CONSUMER ASSOCIATIONS – The unusual blackout on the web immediately drew the attention of consumer associations. The quickest to react was Adiconsum , which announced an action plan to protect the consumers involved. Pietro Giordano, Adiconsum’s Secretary General, states: “[…]”. The same view is held by Adoc, as confirmed by Adoc president Carlo Pileri: […].’
The paragraph reproduced in (42) describes the black-out effects in consumer associations. The CC of the PC introduces one of these associations, Adiconsum, which becomes the Topic of the next proposition (“whichTopic announced an action plan to protect the consumers involved”). On the contrary, the utterance beginning with Dello stesso avviso ‘the same view is held by’ introduces another consumer association, Adoc. This new referent functions as a new Topic, instead of Adiconsum, which is abandoned. This example shows how the CC of a PC can introduce an important discourse referent in spite of the fact that it is not resumed throughout the text or in one paragraph: Adiconsum is important because it is one of the associations, which is the Topic of the paragraph as can be seen from the title (see the beginning of the text given in [42]). The saliency of this kind of CC is based on the fact that it is a derived Topic, i.e. a Topic derived from the main paragraph Topic.
Pseudo-cleft sentences. Italian-Spanish in contrast
211
5 Conclusions Let us conclude by summarizing the most significant results obtained in this study and discussing some open questions to be dealt with in future research. 5.1. The data presented in this paper has shown that Italian and Spanish, two very similar languages, display significant similarities as regards to the form, the frequency and functions of the PC. However, we have also observed important differences between the two languages.⁵³ Firstly, on a formal level, we have chosen to divide PCs into two subtypes, depending on the element introducing the relative clause: wh-PCs (introduced by a fused relative pronoun, as Italian chi, dove, quando and Spanish quien, donde, cuando ‘who, where, when’ etc.) and th-PCs (introduced by an NP followed by a relative pronoun, as Italian quello / ciò / la cosa / la persona / la ragazza che and Spanish lo / la cosa / la persona / la chica que ‘what / the thing / the person / the girl that’); the latter can in turn be divided into three groups (cf. Table 1), according to the form of the NP preceding the relative clause. All the types of PCs described above have different frequencies of use, both in the same language and in an Italian-Spanish contrastive perspective. The analysis of our corpus has shown that the frequency of use of all the types of PCs in Spanish is by far higher than the frequency observed for Italian. On the basis of this observation, we have looked for possible explanations, focusing in particular on the different elements introducing the relative clause. We have indeed been able to single out forms which are specific to Spanish, such as lo único que ‘the only [thing] that’, and forms that, although present in both languages, are preferred in one of them. This is the case for PCs headed by the NP cosa ‘thing’, preferred in Italian. Another case is the sequence ‘NP + of / in order to + inf.’, which is preferred in Spanish. Moreover, our corpus shows a clear difference in use between wh-PCs and th-i-PCs. The fact that these two types of PCs are usually considered equivalent in the literature could result from the reference point of most analyses, namely the research conducted on English: in this language, PCs introduced by what, where, who, etc. are much more frequently used (Berretta 1994: 87). It is nevertheless to be remembered that this lack of homogeneity between the frequency of wh-PCs and th-i-PCs in the two languages must be considered within the limits of the
53 The similarities and differences between Italian and Spanish have been observed by several authors for other Cleft constructions, notably by Guil (1994: 111–126) who focuses on the socalled inferential clefts (It. è che, non è che; Sp. es que, no es que ‘it is that’, ‘it’s not that’). For an analysis of this structure in different languages, see Atayan and Wienen (in this volume).
212
Rocío Agar Marco
present corpus and must not be generalized to other text types. This is true only as far as the language of online newspapers is concerned. As regards text organization, after having made explicit what is intended here as “recovery” of the CC (namely, explicit recovery or recovery under the form of an implicit subject), we have focused our attention on quantitative data concerning the rate of CC recovery in our corpus, which turns out to be over 50% in both languages. Moreover, through the analysis of the co-text following the PC, as well as of the type of CC recovery, it has been observed that the referent of the CC in a PC plays a very important role in the thematic organization of the text, since it introduces elements which are relevant to the text development. Globally, our analysis has shown that the CC in a PC tends to introduce a referent which becomes the central Topic of the discourse in which the PC is found. The semantic and pragmatic importance of this referent, verified on the basis of the analysis of some textual functions, together with the strong tendency of the CC to be discourse-new, confirms the hypothesis according to which PCs have a presentative function. In contrast to what we observed about the form and frequency of PCs in Italian and Spanish, no relevant differences have emerged, in this phase of our research, in the functions of the PCs in both languages. Indeed, most of the textual functions carried out by PCs are found in both Italian and in Spanish. Specifically, the introduction of the Topic of one or more utterances through the CC turned out to be the most frequent function in both languages. Our analysis has also highlighted the fact that the Topics introduced by the CC of a PC also often coincide with the main Topic of the paragraph or of the whole text. There are however also some functional differences between Italian and Spanish PCs. For instance, some functions of the construction are attested in Italian only: cf. the case in which the CC of the PC introduces the most salient referent in the text for the first time (see [34]). We may wonder if such a difference could be related to a different journalistic style in Italian and Spanish online news writing. 5.2. Many issues remain open. A question arises, for instance, about the extremely high frequency of único ‘only’ in Spanish PCs. In the present work, a brief mention has been made of the possible different forms of PCs containing this adjective to explain the differences in frequency in Italian and Spanish. It would be extremely useful, however, to carry out research in this direction so as to determine whether there are other aspects related to this phenomenon, for instance a tendency to grammaticalization and the loss of markedness of the construction (cf., among others, Berretta 2002 for Cleft sentences in Italian).
Pseudo-cleft sentences. Italian-Spanish in contrast
213
Another open question concerns the form of data rebalancing between Spanish PCs and Italian Reversed clefts, which are used exclusively in the latter language. Translation corpora, as briefly mentioned in this paper, could be an interesting means of investigating the hypothesis that Reverse clefts might be used in Italian in many cases where Spanish would use PCs. A further aspect to be investigated is the relationship between the use of PCs and spoken discourse. For a better understanding of the different functions of this construction, it would be useful to investigate, for instance, whether or not PCs are found within reported speech, and, if they are, how frequently they occur. Data gathered on the subject appears to confirm a certain correlation between the PC structure and reported speech. In our corpus of newspaper texts, where quoting is rather frequent, nearly half the examples appear within reported speech. In light of our research, it can be concluded that PCs are much more than simple focusing devices: in the journalistic texts observed in this study, PCs are an important mechanism of text organization. In this paper, we have concentrated on the role played by PCs in the thematic dimension of text organization. Further investigation is required as regards its logical dimension.
References Agar Marco, Rocío. 2014. Le frasi pseudoscisse nei testi giornalistici online: Italiano e spagnolo a confronto. In Enrico Garavelli & Elina Suomela Härmä (eds.), Atti del XII Congresso SILFI. Dal manoscritto al web: canali e modalità di trasmissione dell’italiano. Tecniche, materiali e usi nella storia della lingua (Helsinki, 18–20 giugno 2012), 557–565. Firenze: Franco Cesati. Atayan, Vahram & Ursula Wienen. This vol. Inferential cleft constructions in translation. French c’est que in political texts. Baranzini, Laura. In press. Le frasi pseudoscisse nei testi giornalistici online: italiano e francese a confronto. In Elina Suomela-Härmä (ed.), Atti del XI Congresso SILFI. Dal manoscritto al web: canali e modalità di trasmissione dell’italiano. Tecniche, materiali e usi nella storia della lingua (Helsinki, 18–20 giugno 2012). Baranzini, Laura. This vol. Pseudo-cleft sentences. Italian-French in contrast. Berretta, Monica. 1994. Ordini marcati dei costituenti di frase in italiano. Vox romanica 53. 79–105. Berretta, Monica. 1995. Ordini marcati dei costituenti maggiori di frase: una rassegna. In Silvia Dal Negro & Bice Mortara Garavelli (eds.), Temi e percorsi della linguistica. Scritti scelti, 125–169. Vercelli: Edizioni Mercurio. Berretta, Monica. 1996. Come inseriamo elementi nuovi nel discorso/3: ‘Che mi fa paura è la nebbia’. Italiano e Oltre 2. 116–122.
214
Rocío Agar Marco
Berretta, Monica. 2002. Quello che voglio dire è che: le scisse da strutture topicalizzanti a connettivi testuali. In Gian Luigi Beccaria & Carla Marello (eds.), La parola al testo. Scritti per Bice Mortara Garavelli, 15–31. Torino: Edizioni dell’Orso. Birkner, Karin. 2008. Relativ(satz)konstruktionen im gesprochenen Deutsch. Syntaktische, prosodische, semantische und pragmatische Aspekte. Berlin & New York: Mouton de Gruyter. Brianti, Giovanna. This vol. Cleft sentences. A translation perspective on Italian and French. Collins, Peter C. 1991. Cleft and Pseudo-cleft Constructions in English. London: Routledge. CORIS / CODIS. http://dslo.unibo.it/CODIS/ (accessed 24 January 2013). De Cesare, Anna-Maria. 2005. La frase pseudoscissa in italiano contemporaneo. Aspetti semantici, pragmatici e testuali. Studi di grammatica italiana 24. 293–322. De Cesare, Anna-Maria. 2011. L’ordine dei costituenti in italiano e in prospettiva contrastiva con il tedesco. Tra sintassi, tipologia e pragmatica, Habilitationsschrift, Istituto d’italianistica, Università di Basilea, MS. De Cesare, Anna-Maria. This vol. Cleft Constructions in a contrastive perspective. Towards an operational taxonomy. De Cesare, Anna-Maria, Davide Garassino, Rocío Agar Marco & Laura Baranzini. This vol. Form and frequency of Italian Cleft constructions in a corpus of electronic news. A contrastive perspective with French, Spanish, German and English. Eguren, Luis J. 1999. Pronombres y adverbios demostrativos. Las relaciones deícticas. In Ignacio Bosque & Violeta Demonte (eds.), Gramática descriptiva de la lengua española 3, 929–972. Madrid: Espasa Calpe. Ferrari, Angela, Luca Cignetti, Anna-Maria De Cesare, Letizia Lala, Madga Mandelli, Claudia Ricci & Carlo Enrico Roggia. 2008. L’interfaccia lingua-testo. Natura e funzioni dell’articolazione informativa dell’enunciato. Alessandria: Edizioni dell’Orso. Furukawa, Naoyo. 1994. Ce que je crois, c’est que…: séquence thématique et ses deux aspects, cohésion et rupture. Travaux de linguistique 29. 21–37. Gil, Alberto. 2004. Textstrukturelle Dimensionen der Satzspaltung im Italienischen. In Alberto Gil, Christian Schmitt, Dietmar Osthus & Claudia Polzin-Haumann (eds.), Romanische Sprachwissenschaft: Zeugnisse für Vielfalt und Profil eines Faches: Festschrift für Christian Schmitt zum 60. Geburtstag, 369–385. Bern: Peter Lang. Guil, Pura. 1994. Es que… in italiano. In Anna Giacalone Ramat & Massimo Vedovelli (eds.), Italiano lingua seconda / lingua straniera, Atti del XXVI Congresso internazionale di studi della SLI (Siena, 5–7 novembre 1992), 111–126. Roma: Bulzoni. Helfrich, Uta. 2003. Hendidas y seudo-hendidas. Un análisis empírico-diacrónico. In Fernando Sánchez Miret (ed.), Actas del XXIII Congreso Internacional de Lingüística y Filología Románica (Salamanca, 24–30 septiembre 2001), vol. 2 (1), 439–452. Tübingen: Niemeyer. Hetzron, Robert. 1975. The presentative movement or why the ideal word order is V.S.O.P. In Charles N. Li (ed.), Word Order and Word Order Change, 345–388. Austin: University of Texas Press. Johansson, Mats. 2001. Clefts in contrast: A contrastive study of it clefts and wh clefts in English and Swedish texts and translations. Linguistics 39 (3). 547–582. Kiese, Jörn. 1993. Fokussierende Sätze im Deutschen und Englischen. Bern: Peter Lang. Lambrecht, Knud. 1994. Information Structure and Sentence Form. Topic, Focus, and the Mental Representations of Discourse Referents (Cambridge Studies in Linguistics 71). Cambridge: Cambridge University Press.
Pseudo-cleft sentences. Italian-Spanish in contrast
215
Leonetti, Manuel. 1999. El artículo. In Ignacio Bosque & Violeta Demonte (eds.), Gramática descriptiva de la lengua española 3, 787–890. Madrid: Espasa Calpe. Metzeltin, Michael. 2010. Erklärende Grammatik der romanischen Sprachen: Satzkonstruktion und Satzinterpretation. Wien: Praesens Verlag. Moreno Cabrera, Juan Carlos. 1999. Las funciones informativas. Las perífrasis de relativo y otras construcciones perifrásticas. In Ignacio Bosque & Violeta Demonte (eds.), Gramática descriptiva de la lengua española 3, 4245–4302. Madrid: Espasa Calpe. Panunzi, Alessandro. 2009. Strutture scisse e pseudoscisse. Valori d’uso del verbo essere e articolazione dell’informazione nell’italiano parlato. In Angela Ferrari (ed.), Sintassi storica e sincronica dell’italiano: subordinazione, coordinazione, giustapposizione. Atti del X Congresso della SILFI (Basilea, 30 giugno–3 luglio 2008), vol. 2, 1121–1137. Firenze: Franco Cesati. Pinedo, Alicia. 2000. English clefts as discourse-pragmatic equivalents of Spanish postverbal subjects. Estudios Ingleses de la Universidad Complutense 8. 127–151. Roggia, Carlo Enrico. 2009. Le frasi scisse in italiano. Struttura informativa e funzioni discorsive. Genève: Slatkine. Salvi, Giampaolo. 1991. Le frasi copulative. In Lorenzo Renzi & Giampaolo Salvi (eds.), Grande grammatica italiana di consultazione, 163–189. Bologna: Il Mulino. Sornicola, Rosanna. 1988. It-Clefts and Wh-clefts: two awkward sentence types. Journal of Linguistics. 343–379. Van den Steen, Katleen. 2005. Cleft constructions in French and Spanish. In Nicole Delbecque, Johan van der Auwera & Dirk Geeraerts (eds.), Perspectives on Variation: Sociolinguistic, Historical, Comparative, 275–290. Berlin & New York: Mouton de Gruyter. Wienen, Ursula. 2006. Zur Übersetzbarkeit markierter Kohäsionsformen: Eine funktionale Studie zum Kontinuum von Spaltadverbialen und Spaltkonnektoren im Spanischen, Französischen und Deutschen. Bern: Peter Lang.
Iørn Korzen*
Cleft sentences. Italian-Danish in contrast 1 Introduction The aim of this paper is to compare the use of Cleft sentences (henceforth CSs) in Italian and Danish. The comparison is primarily based on the Italian and Danish part of the EUROPARL-corpus, which contains the proceedings of the European Parliament dating back to 1996 (Koehn 2005; http://statmt.org/europarl/). The approach chosen is synchronic, non-generative and limited to it-clefts, and it will give relatively more focus to the situation in Danish and to the differences encountered between Danish and Italian. The paper is structured in the following way: Section 2 gives an account of CSs in Danish, with § 2.1 focusing on the linguistic literature and § 2.2 on the structural and syntactic features of Danish CSs. Section 3 describes the equivalent situation in Italian, similarly focusing on the literature (§ 3.1) and on structure and syntax (§ 3.2.). Based on some of the scholars cited in Sections 2.1 and 3.1, I suggest, in Section 4, a typology of CSs, which I use, in Section 5, to compare the linguistic material and the textual function of the Italian and Danish CSs in the EUROPARL-corpus investigated. Before the concluding remarks in Section 7, I briefly refer to another corpus of comparable Italian and Danish texts, the “MR. BEAN-corpus”, and look at the CS occurrences found there. As it turned out, Danish CSs were more than three times more frequent than the Italian counterparts in the EUROPARL-corpus, so in order to arrive at an empirical basis of a reasonable size in both languages, I found it necessary to examine a much larger Italian than Danish text corpus, as Table 1 shows:
* I am grateful to Anna-Maria De Cesare, Michael Herslund and Hanne Korzen for very fruitful discussions and valuable comments on an earlier draft of this paper.
Iørn Korzen, Copenhagen Business School
218
Iørn Korzen
Table 1: The EUROPARL-corpus investigated and Cleft sentences found Italian
Danish
Text corpus investigated (number of words)
251,284
128,082
Total number of CSs found
119
203
equivalent to 1 CS per
2,112 words
631 words
The same cross-linguistic difference in CS frequency is found in the other comparable corpus that I shall refer to in Section 6, a phenomenon that would seem to speak in favour of the so-called compensation mechanism principle, i.e. the hypothesis, first launched by Jespersen (1937: 76) and later discussed e.g. by Lambrecht (2001: 488) and Dufter (2009: 93–94), that the CS may compensate for a relatively more rigid constituent order in a particular language. Constituent order is much more rigid in Danish than in Italian,¹ and we shall return briefly to this discussion in Section 7. All Europarl texts investigated in this paper are L1 texts, i.e. texts originally produced in the given language and not translated from other languages. Whereas parallel texts, i.e. L1–L2 texts, are clearly best suited e.g. for improving machine translation (one of the original and main purposes of the EUROPARLcorpus, cf. Koehn 2005), comparable (L1–L1) texts are generally more suitable for descriptive, possibly typological comparisons. In the case of parallel texts, the “filter” of the translator and possibly specific translation strategies risk confusing the picture, and/or the L2 texts risk ending up with a text structure too similar to that of the L1, a phenomenon Baroni and Bernardini (2006) refer to as “translationese”.²
1 On the rigid constituent order in Danish and on the classification of Danish as a V2 language, see e.g. Heltoft (1992), Hansen & Heltoft (2011: 311–334) and other references cited therein. On the relatively much less fixed word order in Italian and comparisons with Danish, see e.g. Bach & Schmitt Jensen (1990: 648–660). For similar comparisons between Danish and French, see e.g. Herslund (2006) and (In press). 2 “It is common, when reading translations, to feel that they are written in their own peculiar style. Translation scholars even speak of the language of translation as a separate ‘dialect’ within a language, which they call third code […] or translationese […]. Translationese has been originally described […] as the set of “fingerprints” that one language leaves on another when a text is translated between the two”. (Baroni and Bernardini 2006: 260). Cf. also McEnery et al. (2006: 49), who state that “source and translated texts […] alone serve as a poor basis for cross-linguistic contrasts, because translations (i.e. L2 texts) cannot avoid the effect of translationese”. See also Korzen and Gylling (2012).
Cleft sentences. Italian-Danish in contrast
219
2 Cleft sentences in Danish As in other languages, Danish it-clefts are binary structures with a matrix consisting of an impersonal, non-anaphoric and non-deictic pronoun subject (det ‘it’), a copula verb and the cleft constituent (henceforth ClC) followed by a subordinate clause (henceforth subclause). The choice of subclause connective depends on the syntactic relation between the ClC and the subclause, i.e. on the syntactic function that the ClC would have in the equivalent non Cleft sentence (see Table 2 in Section 2.2.1). For this reason, Danish CS descriptions are often based on syntactic criteria.
2.1 A short literature survey Before the 1970s and e.g. Dyhr’s (1978) volume on Danish and German CSs, the literature on Danish CSs was relatively limited. The most common Danish term for CS is “sætningskløvning” (‘sentence clefting’), and it appears for the first time – rather briefly – in Wiwel (1901: 343–344), who considers CSs to be very “usable” (“brugelige”) for “putting a special emphasis on a word”. According to Wiwel, the close relationship between the two parts of the CS is revealed by the fact that personal pronouns in the third person singular in the matrix clause must be in the reflexive form if they refer to the subject of the subclause, see example (15) in Section 2.2.2. After Wiwel, CSs are treated – equally briefly – by Byskov (1910: 122–123) under the term “Tveformede Sætninger” (Lit. “twofold sentences”). Byskov adds that generally the subclause has no initial connective but is considered as a relative or completive subclause depending on the possibility to insert som / der ‘that’ (relative) or at ‘that’ (completive), respectively. Mikkelsen (1911: 569–572) treats CSs in a chapter entitled “Extension of one sentence into two” (“Udvidelse af én sætning til to”), and elsewhere in his book (Mikkelsen 1911: 21 and 507, respectively) he defines the CS as a “unique” (“særegen”) and “remarkable” (“ejendommelig”) kind of construction. The subclause is relative if the “emphasised” word in the original sentence is subject or related to the subject (in written language also if direct or indirect object, part of a constituent consisting of a prepositional phrase, or related to such constituents), and a completive clause in all other cases (as well as if direct or indirect object or part of a prepositional phrase in spoken language); see examples in the following Sections. Mikkelsen (1911: 570) also mentions the particular temporal CS type:
220
(1)
Iørn Korzen
For ti år siden rejste han herfra → Det er ti år siden, (at) han rejste herfra (without the preposition for). Lit. ‘For 10 years ago he left here → It is ten years ago (that) he left here.’
See an authentic example in (7) below. In modern Danish the structure in (1) is much more frequent than the “full” CS with the preposition for in (1’): (1’)
Det er for ti år siden, (at) han rejste herfra. Lit. ‘It is for ten years ago (that) he left here.’
Rehling (1932: 49–50) uses the term “sentence splitting” (“sætningsspaltning”), whereas A. Hansen (1933: 39–49), who dedicates more detail to the phenomenon, prefers the terms “cleft” and “split sentences” (“kløvede / spaltede sætninger”), although he admits not being totally content with these terms since the sentence is not really divided into two but reflects one and only one utterance (A. Hansen 1933: 42). The CSs play a very important role in modern spoken Danish, having had a very “rich evolution” (A. Hansen 1933: 47). CSs are treated briefly by Diderichsen ([1946] 1971: 196), who states that the structure is used “when an entity is to be particularly emphasised or indicated”. The subclause is vaguely defined as a “loosely attached subordinate clause dislocated to the right” (“en løst tilknyttet Bisætning i Ekstraposition”), which has “no normal syntactic function” in relation to the matrix clause, and of which it is generally “preferred to omit an initial conjunction”. Dyhr’s (1978) study is a systematic and thorough investigation of Danish and German it-clefts and Pseudo-clefts based on a corpus of novels and newspaper texts. He dedicates most of his volume to the syntactically based analysis, but includes also minor chapters on generative and text linguistic aspects, the latter treating Theme-Rheme phenomena and old vs. new and focus information in relation to CSs (Dyhr 1978: 155–162). Nølke (1984) investigates pragmatic and derivational aspects of CSs and claims that “a study of Danish clefts presents some arguments against a transformational approach” and that a “uniform non-transformational treatment seems to be required.” (Nølke 1984: 72). According to Nølke, CSs consist of a focused argument and an open sentence describing this argument, and CSs “always (i) presuppose strongly that the open sentence is true of a specific number of (unknown) individuals, (ii) presuppose weakly that it is false of some other individuals, and (iii) assert the identification of the presupposed “true” individuals” (Nølke 1984: 75). The focal constituent must be “presented as having been selected among similar possibilities, which implies that it is associated with a paradigm of alternatives.” (Nølke 1984: 80).
Cleft sentences. Italian-Danish in contrast
221
Gettrup et al. (1986: 72–77) see the function of the CS as the rhematisation of one constituent, often the subject, and the thematisation of all other constituents. The rhematised constituent is emphasised and the Rheme-Theme structure of the CS distinguishes it radically from the non-cleft structure. The CS can also indicate a topic shift and thereby function as a text structuring device. E. Hansen’s (1995) lengthy description of it-clefts is partly repeated in Hansen and Heltoft’s (2011) recent and comprehensive Danish grammar, where the chapters on it-clefts take up 24 pages. The CS is said to “correspond” to a synonymous non-cleft sentence; the initial subject pronoun det ‘it’ is always unstressed and the subordinate clause is “formed like a relative clause” headed by der, at (corresponding to the relative ‘that’ and the completive ‘that’, respectively) or without a conjunction (Hansen and Heltoft’s 2011: 127). The CS structure is defined as “the most important grammaticalisation of focusing in Danish” (Hansen and Heltoft 2011: 129). The focus is new information whereas the rest is “old information”, which “the speaker or writer presumes to be known by the addressee”, either due to general background knowledge or to information given in the preceding text. If the presupposed content is part of the immediately preceding text, the subclause is normally omitted, and we get an elliptic CS, such as: (2)
Someone took my sunglasses. Who was it [that took my sunglasses]?
According to Hansen and Heltoft (2011), the ClC identifies the entity/ties of which the subordinate clause expresses truthful information. Consequently, to every correct CS must correspond an “identification negation” or “exclusion sentence” concerning entities for which the subordinate clause does not hold. The CS type: (3)
It is with regret that the Government must note that […]
is called a “weak cleft”, which is said to be rarer and more peripheral (Hansen and Heltoft 2011: 139–140). The subclause verb is a verbum dicendi and the subordinate clause is not presupposed as in the “genuine cleft”. Götzsche (1998) rejects E. Hansen’s (1995) conception of the CS structure as a particular grammatical construction. Depending on the “reference” of the matrix subject, the same structure can have different “pragmatic functions”, Götzsche talks about “identification” (the pragmatic content of a CS), or “description” or “explanation” (when the structure ‘It is + predicative argument + subclause’ is not interpreted as a CS but It has a deictic function). Thus, according to Götzsche, talking about a particular grammatical CS structure would be erroneous. For
222
Iørn Korzen
more discussion on the grammatical status of CSs, see Sections 3 and 4 and footnotes 12 and 21. Herslund (2005) considers the focused constituent as the Theme of the utterance, typically endowed with an anaphoric function, including gerunds or constituents with an anaphoric meaning “derived” by contrast or other relations to the preceding co-text. Also in other cases, the CS presents the constituent as somehow connected with the preceding co-text, thus it acquires a text connective function. In some cases, CSs come close to grammaticalisations, e.g. (in French) C’est ainsi que (Herslund 2005: 133). Conversely, Togeby (2009) sees the presupposed subclause of the CS as the Theme of the utterance cataphorically referred to by the initial it.³ The identification predication of the matrix is the rhematic part which functions as the dissolving of a dilemma on the addressee’s part (but not on the speaker / writer’s part), namely choosing for which of a given set of alternatives the subordinate clause is valid. Togeby does not consider the cleft entity to be highlighted or opposed to other entities, but merely as pointed out, and he finds the meaning of the it-clefts best described as “conclusive pointing out” (Togeby 2009: 447).
2.2 Cleft sentences in Danish: Structural and syntactic features As already stated, the Danish matrix clause consists of non-anaphoric and nondeictic subject det ‘it’ + verb + ClC. The verb is a finite form of the copula verb være ‘be’ or blive ‘be / become’ possibly combined with a modal verb; see Section 2.2.3.
2.2.1 Syntax and subclause connectives The subclause may have an initial connective, der / som ‘that’ (relative) or at ‘that’ (completive), or it may lack a connective, [Ø], depending on the syntactic relation between the ClC and the subclause. The correlations are shown in Table 2, of which the right column indicates the connectives occurring in the Danish EUROPARL-corpus investigated. The corpus was searched for constructions with the pronoun det ‘it’ and the relevant cases, i.e. the CSs, were picked out manually. For more statistic details, see Section 5. 3 A similar analysis of it in CSs is found elsewhere as well, cf. for English e.g. Bache and Davidsen-Nielsen (1997: 262): “The pronoun It is simply a form word representing the information expressed by the relative clause”. For a brief discussion, cf. also Nølke (1984: 107–108).
Cleft sentences. Italian-Danish in contrast
223
Table 2: Cleft constituent and subclause connective in Danish ClC: syntactic function
Subclause connective
Corpus occurrences
a. subject b. direct object c. indirect object d. prepositional complement e. genitive determinative f. lexical verb in analytic forms g. predicative h. adverbials (incl. adverb. clauses)
der / som ‘that’ (relative) ø / som ø ø / som hvis ‘whose’ ø / (in rare cases at ‘that’) ø ø / at ‘that’ (completive)
106 der; 9 som 22 ø; 5 som none 25 ø; 4 som none none none 16 ø; 6 at
As Table 2 shows, the connective som is generally much rarer than the alternative possibilities; it seems to occur particularly in written Danish (Hansen and Heltoft 2011: 1793) and often in cases of complex matrix clauses, as in (4):⁴ (4)
subject (a) Det er denne tvangstanke om Europas grænseløst forenede stater, som nu i praksis lægger også nyttigt internationalt samvirke for had. (ep-96-10-23. txt:28) ‘It is this obsession about Europe’s borderless United States that now in fact makes unpopular even useful international collaboration.’
In the following two cases, som could easily have been omitted, and it gives a somewhat complex, possibly archaic, feeling to the text: (5)
direct object (b) Derfor er det også en vigtig betænkning, som min gruppefælle har skrevet, og af den fremgår det, […] (ep-03-02-11.txt:66) ‘Therefore it is also an important report that my colleague has written, and it shows […]’
4 In this and the many following Europarl citations, the English translations are literal and made by this author, hence they are not the official translations of the European Parliament. The aim of the translations has generally been to preserve the original cleft construction (and possibly other structures as well) and not to produce English sentences in their own right. CSs are written in bold and phenomena particularly investigated in bold and/or italics. References to the EUROPARL-corpus (“ep”) consist of numbers indicating year-month-day and speaker identity (“txt”).
224
(6)
Iørn Korzen
prepositional complement (d) Det han drømte om, var det han kaldte et lille, praktisk sekretariat. Det er ikke det, som Prodi er formand for i dag. (ep-00-02-15.txt:25) ‘What he dreamed of was what he called a small, practical secretariat. It’s not that which Prodi is chairman of today.’
Of the 22 adverbial ClCs mentioned under (h) in Table 2, five were cases of the particular temporal type cited in (1) above and in (7). Of these five cases, only one occurred with an at subclause. Similarly to som, at occurs often, but not necessarily, in cases of complex matrix clauses; see examples with adverbial ClCs + at in (21) and (39). An example without at: (7)
adverbial (temporal) (h) Hr. formand, det er over 50 år siden, [ø] vi vedtog menneskerettighedserklæringen, men vi må nok spørge os selv, om menneskerettighedssituationen faktisk er blevet bedre. (ep-01-01-17.txt:142) ‘Mr President, it is more than 50 years ago, [ø] we adopted the Declaration of Human Rights, but we may well ask ourselves whether the human rights situation has improved.’
Of the 106 der subclauses linked to a subject ClC and the 22 [ø] subclauses linked to an object ClC, two subject and one object ClC were found in interrogative constructions. In (8), I cite a case with an interrogative subclause, the parallel main interrogative clause being Hvad var det egentlig der skete i Berlin? ‘What was it actually that happened in Berlin?’: (8)
interrogative / subject Det er jo vanskeligt at behandle disse sager uden også at se dem i lyset af, hvad det egentlig var, der skete i Berlin. (ep-99-04-13.txt:31) ‘It’s difficult to deal with these matters without seeing them also in the light of what it actually was that happened in Berlin.’
In the Italian EUROPARL-corpus, only one case of such an interrogative construction was found, and there may be a diaphasic difference between the two languages, i.e. that the construction is felt as more informal in Italian⁵ than in Danish – even though also the number of Danish occurrences was very low.
5 See Section 3.1 and Sabatini (1985: 163).
Cleft sentences. Italian-Danish in contrast
225
As indicated in Table 2, no cases of (c), (e), (f), and (g) were found in the Danish EUROPARL-corpus (nor in the MR. BEAN-corpus” that I cite in Section 6), and in both written and spoken Danish, these structures are generally rare – especially the (f) type. The following examples come from Hansen and Heltoft (2011: 1790–1791):⁶ (9)
indirect object (c) Det blev Jens hun sendte checken. ‘It became Jens [whom] she mailed the check.’
(10)
genitive (e) Det er din onkel hvis hus er brændt. ‘It is your uncle whose house burned down.’
(11)
lexical verb (f) Det er rejst han er. Det må være spist østers hun har. ‘It is left [that] he has. It must be eaten oysters [that] she has.’⁷
(12)
predicative (g) Det er ikke ret meget fed du er. ‘It is not very cool [that] you are.’
Not included in Table 2 are 10 occurrences of the type Det er (ikke) sådan at…, Det er ikke fordi… ‘It is (not) so [the case] that…, It is not because…’, a type which is semantically very similar to the Italian structures è che… / non è che… that many scholars include among CSs (see footnote 15). (13)
Det er stadig sådan, at det er kvinderne, der tjener mindst, […] (ep-03-0312.txt:258) Lit. ‘It is still so [i.e. the case] that it is women who earn less […]’
In this Danish case, the subclause is, strictly speaking, a consecutive clause headed by sådan at ‘so that’, but the consecutive meaning is very weak and the clause can be analysed as completive, as in the similar Italian cases. However, this construction shall be left out of the investigation of this paper, see Section 5 and Table 5.
6 Since no source is indicated for examples (9) to (11), these are probably created by the authors. Example (12) is attributed to the Danish author Klaus Rifbjerg. 7 These types are treated fairly extensively by Dyhr (1978: 58–68).
226
Iørn Korzen
2.2.2 Pronominal cleft constituents If the ClC is a personal pronoun, the oblique form is used as in other cases of personal pronouns in non-subject positions; in the polite third person singular, De ‘you’ is nominative and Dem is oblique: (14)
Jeg mener, at det var Dem, hr. formand, der hævede mødet, og det er også, hvad der står i den franske version. (ep-02-01-16.txt:3) ‘I think it was you, Mr President, who adjourned the meeting, and that is also what the French version says.’
If the personal pronoun refers to the subclause subject, the reflexive form is used (no examples found in Europarl): (15)
Det er sig selv han kan takke for sin succes. Det er af sine venner man skal høre sandheden. (Hansen and Heltoft 2011: 1792). ‘It is himself he can thank for his success. It is from one’s friends one shall hear the truth.’
In such cases the oblique non-reflexive form may however appear in spoken language (cf. Hansen and Heltoft 2011: 1793).
2.2.3 The matrix verb By far the most common matrix verb in the Danish EUROPARL-corpus is være ‘be’ in the present tense. The corpus contains two occurrences of være in the preterite tense, one is cited in (16), and three examples with blive ‘become’ indicating future time, of which one is cited in (17): (16)
Det var den praksis, der førte til Kommissionens fald. (ep-00-01-18.txt:259) It was that practice that led to the fall of the Commission.
(17)
Men den manglende handlekraft fra Rådet har faktisk også den konsekvens, at det i stedet for de ansvarlige politikere bliver domstolene, der træffer beslutningerne. (ep-01-01-16.txt:113)
Cleft sentences. Italian-Danish in contrast
227
‘But the lack of action from the Council has in fact also the consequence that instead of the responsible politicians it will be the courts that make the decisions.’ Future time can also be expressed by the modal verb vil ‘will’ + infinitive, which is found three times in the corpus, one is cited in (18). Moreover, the corpus contains two cases in which the modal verb burde ‘shall’ occurs, cf. (19), one with modal skulle ‘must’ and one with modal kunne ‘can’ in the matrix clause. Such cases of “raising” are a particular feature of Danish and concern epistemic and deontic modals, cf. Herslund (2003: 876–877): (18)
Så vil det være folkevalgte fra medlemslandene, der afgør, om beslutninger skal flyttes fra vælgere og folkevalgte i medlemslandene til EU. (ep-01-0314.txt:17) ‘So it will be people elected from the member states who decide whether decisions shall be transferred from voters and elected representatives of the Member States to the EU.’
(19)
Det bør være de kompetente udvalg i Parlamentet, der foretager undersøgelser og stiller forslag til reformer. (ep-00-01-19.txt:77) ‘It should be the competent parliamentary committees that conduct investigations and make proposals for reforms.’
2.2.4 Other details Danish has a special particle jo, equivalent to German ja,⁸ which underlines that the information conveyed by the surrounding sentence or clause is supposedly shared among the interlocutors and not under discussion. Therefore, jo is fairly frequent in the matrix clause of Danish CSs. In the EUROPARL-corpus it occurred in 17 of the 203 CSs, and cases were found in all the CS types represented in the corpus, distributed in the following way: Subject: 4 cases; object: 1; prepositional complement: 7, cf. (20); adverbials: 4, cf. (21); 1 case was found in a construction like the one cited in (13).
8 On German ja in CSs, cf. Dufter (2009: 110–111).
228
(20)
Iørn Korzen
[…] og det er jo beskæftigelsen, vi taler om. (ep-99-01-13.txt:238) ‘[…] and it is [as we all know] employment we are talking about.’
(21)
Men det er jo ikke i denne betænkning, at vi beslutter det, […] (ep-0005-16.txt:41) ‘But it is [as we all know] not in this report, we decide it […]’
As in other languages, CSs or parts thereof can be repeated with a particular rhetorical effect to which we shall return in Section 5.2.1.5; in (22) a subclause is repeated and in (23) the first two CSs have identical matrix clauses whereas the third is elliptic and contrastive: (22)
Det er prisen på pengene, som ikke må stige, det er det offentliges forbrug, som ikke må stige. (ep-00-03-13.txt:27) ‘It is the price of money that must not increase, it is public consumption that must not increase.’
(23)
Det er ikke EU, der skal skaffe os af med en regering. Det er ikke EU, der skal gennemføre en human udlændingepolitik, det er danskerne selv – we, the people. (ep-02-02-04.txt:53) ‘It is not the EU that should rid us of a government. It is not the EU that should implement a humane foreign policy, it is the Danes themselves – we, the people.’
When, as in the last case cited, the content of a possible subclause is co- or contextually given, i.e. mentioned in a previous co-text or obvious given the communication situation, it is normally omitted, cf. Section 2.1, example (2). For instance, the typical Danish way of initiating a phone call is a case of contextual givenness and subclause omission: (24)
The person answering the phone: Det er Peter [der taler / tager telefonen]. ‘It is Peter [who speaks / answers the telephone].’ The person calling: Det er Lisa [der ringer]. ‘It is Lisa [who is calling].’
The subclauses between the brackets in (24) are always implicit.
Cleft sentences. Italian-Danish in contrast
229
3 Cleft sentences in Italian (and other languages) 3.1 A short literature survey The Italian literature on CSs is considerably vaster than the Danish literature, and this Section merely intends to give a few examples of viewpoints that have been particularly inspiring for this paper. For extensive references, see e.g. De Cesare (in this volume) and Garassino (in this volume). Not only are linguistic studies more numerous, their methodologies are also more varied. For instance, generative studies are fairly frequent, e.g. Graffi (1978), Benincà (1978) (on the structure cited in [1] and [7] above), Frison (1982, 2001), and Grewendorf and Poletto (1989), and CSs have been investigated diachronically e.g. by Metzeltin (1989) (who cites examples from 14th Century Italian and compares CSs in the five major Romance languages), D’Achille et al. (2005) (who compare modern Italian with English), and Sornicola (1991). CSs in special diamesic variants or text types have also been studied, cf. e.g. De Stefani (2009) on functional aspects of CSs in conversational Italian and French and De Cesare (2012) on CSs in Italian and English news releases produced by news agencies. Most Italian grammars refer to CSs, albeit some quite briefly. CSs were mentioned as early as in 1881 by Fornaciari ([1881] 1974: 383), who states: “The French make widespread use of this reinforcement, extending it not only to the subject but also to the object and all other sentence complements […]. But this is very different from the disposition of our language”.⁹ Regarding CS frequency in particular diamesic and diaphasic variants, Dardano (2005: 206) and Dardano and Trifone (2001: 82, 448–449) place CSs among “oral phenomena”, whereas Serianni (1991: 569) believes the construction to be very frequent in modern both written and spoken Italian of all linguistic levels and the purist reservations to be extinct. In a vast survey on Italian language in newspapers from the beginning of the 20th century to the present, Bonomi (2002: 44) finds CSs to be sporadic at the beginning of the century, less uncommon around 1930, and very common since the Second World War. Sabatini (1985: 163) mentions CSs as one of 35 phonological, morphological and syntactic features of spoken and written Italian characterised by low to medium formality. Not only does the structure put maximum emphasis on the
9 “I Francesi fanno larghissimo uso di questo rinforzamento, estendendolo non solo al soggetto, ma anche all’oggetto e a’ complementi tutti quanti della proposizione […]. Ma ciò è assai disforme dall’indole della nostra lingua”.
230
Iørn Korzen
new information, it also expands the length of the utterance, which facilitates the addressee’s reception of it. Sabatini mentions the cleft interrogative structures, e.g. (25)
Dov’è che hai comprato questa borsa? Quand’è che parti? Chi è che deve arrivare? ‘Where is it that you bought this bag? When is it that you leave? Who is it that is coming?’
as particularly frequent at this level of formality.¹⁰ Based on a systematic corpus enquiry, Roggia (2006) comes to the opposite conclusion of Dardano and Trifone as well as Sabatini and finds CSs less frequent in oral than in written language – according to the author due to competing marking possibilities in oral language – and more frequent in texts of a certain formality, where more attention is given to the elaboration and thereby variation of the text. Many Italian (as well as non-Italian) researchers on CSs are inspired by American and English scholars such as Prince, Delin and Lambrecht, for instance by Prince’s (1978) distinction between Known, Given and New information. Prince defines Given information as the “information which the coöperative speaker may assume is appropriately in the hearer’s consciousness” due to the specific co- and context, and Known information as the “information that the speaker represents as being factual and already known to certain persons (often not including the hearer)” (Prince 1978: 903). Given information is relevant to WHclefts, whereas Known information is carried by it-CSs as the speaker’s choice of a particular validity-level to ascribe to the utterance (Prince 1978: 903). An it-CS may have a stressed focus representing New information, and the subclause will then typically, but not necessarily, convey Known information and have a low communicative dynamism in the sense of the Prague School.¹¹ In other cases, the subclause contains the central message of the communication, as in:
10 Berretta (2002: 28) calls these structures “strongly marked and redundant with a semi lexical status” (“forme ipermarcate, ridondanti, con statuto semilessicale”). In Italian, the interrogative perché ‘why’ cannot be cleft (D’Achille et al. 2005: 264), unlike the Danish equivalent hvorfor, which has no problem of this kind. 11 See e.g. Firbas (1966), who develops the “functional sentence perspective” of the Prague school, defined as the various degrees of “communicative dynamism” in the sentence. The “communicative dynamism” of a sentence element is defined as “the extent to which the sentence element contributes to the development of the communication” (Firbas 1966: 270). I have discussed the “communicative dynamism” in connection with noun phrases with the partitive (and indefinite) article in Italian in Korzen (1996: 452–454).
Cleft sentences. Italian-Danish in contrast
(26)
231
It was just about 50 years ago that Henry Ford gave us the weekend. (cit. Prince 1978: 898).
In such cases, the subclause contains New information and is therefore high on communicative dynamism, even if the information is structurally marked as a known fact, not as the speaker’s assessment, and the syntactic subordination may convey a degree of “backgroundness” (Prince 1978: 904). This CS type generally has a short and anaphoric focus or an adverbial focus expressing time, place, or reason, and Prince terms it “informative-presupposition it-cleft” (Prince 1978: 898–899).¹² For a more fine-grained distinction between the notions Given - Inferable New, Roggia (2009: 140) suggests Given Semiactive / Inactive (between Given and Inferable) for the entity that has been introduced but at a co-textual distance, and Anchored (between Inferable and New) for the entity that is new in itself but represented by an NP which is linked to the co-text by means of an embedded (e.g. attributive) anaphoric constituent.¹³ For further discussion, see also Garassino (in this volume). The distinction between shared knowledge and logical presupposition is discussed e.g. by Delin (1995). The former may be defined as “those beliefs that conversational participants come to believe to be shared among them as the direct result of the conversational interaction”, whereas the latter is signalled by presupposition-triggering syntactic structures such as CS subclauses and is “immune from operators that express negation, possibility, questioning” (Delin 1995: 99). The presupposition is in itself independent of the speaker / writer’s assumption of whether the hearer / reader is aware of the presupposed content or not. New information is generally indicated by prosodic nuclear accents, and such accents are by no means excluded from the presupposition part of the CS (Delin 1995: 113). In-depth investigation and discussion of pragmatic / informative and text structuring roles of CSs are frequent also in the Italian literature. Based on Dik’s 12 Lambrecht (2001) considers CSs as grammatical constructions in the sense of “form-function pairings whose structural and semantic properties cannot, or not entirely, be accounted for in terms of other properties of the grammar […] and which therefore require independent explanation” (Lambrecht 2001: 466). Thus Lambrecht objects e.g. to treating Prince’s “informative-presupposition it-cleft” as a separate category since no difference is reflected by formal or semantic criteria in grammar, instead he suggests the term pragmatic accommodation for such cases. 13 Based on Prince’s distinction between Given and Known, Roggia (2009: 139) also distinguishes between cognitive Activation (attivazione) of a referent or propositional content, which means “presence in the short-term memory”, and Knowledge (conoscenza), which implies “presence in the encyclopedic long-term memory”.
232
Iørn Korzen
(1989: 277) definition of (information) focus as “that information which is relatively the most important or salient in the given communicative setting, and considered by S[peaker] to be most essential for A[dressee] to integrate into his pragmatic information”,¹⁴ Berretta (1994) distinguishes between three different CS types depending on the function of the ClC: (i) the ClC has the full “original” focus marking, and thus also identifying, function; (ii) the ClC summarises elements of the preceding co-text and functions as a strong anaphor; (iii) the ClC summarises the preceding co-text and marks the transition from one text sequence to another. In (ii) the CS expresses, according to Berretta (1994: 99–100), “perhaps” (Italian “forse” in brackets) focus marking, but with a specific textual function. In (iii), the function of the CS is purely textual; the focalizing role is very feeble (“esile”), and the subclause, which according to the “classical analysis” conveys presupposed information, here brings new information. In many cases, these CSs have become more or less lexicalised formulas signalling transition between sentences or text sequences, see also Berretta (2002) and Bazzanella (1994: 128–134).¹⁵ Also definitions of focus different from that of Dik (1989: 277) just cited have inspired CS analyses, e.g. Krifka (2007: 18): “Focus indicates the presence of alternatives that are relevant for the interpretation of linguistic expressions”; similar definitions are found e.g. in Rooth (1992) and von Heusinger (1999).¹⁶ Based on this tradition as well as on Prince’s typology, Dufter (2009) proposes a more finegrained distinction, in which Prince’s “informative-presupposition clefts” are divided into two subtypes, one corresponding to Prince’s description, in which the ClC is background and the subclause conveys new information (for which
14 Similarly Halliday (1967: 204): “Information focus reflects the speaker’s decision as to where the main burden of the message lies.”. Cf. also the notion of fuoco informativo in De Cesare (2011: 78–83). For a discussion and other references, see e.g. Roggia (2009: 95–107). 15 Like most other Italian scholars, both Berretta and Bazzanella include the structures è che […] / non è che […], ‘it is (not) [the case] that […]’ in the category of CSs, of which the focal element is the polarity (positive vs. negative). This structure is investigated particularly for Italian and Spanish by Guil (1994). 16 Cf. Rooth (1992: 76): “[T]he focus semantic value for the sentence [S [Mary]F likes Sue] is the set of propositions of the form ‘x likes Sue’, while the focus semantic value for [S Mary likes [Sue]F] is the set of propositions of the form ‘Mary likes y’.” Similarly von Heusinger (1999: 125): “In the simple assertion Sam talked to FredF, the focus on Fred is said to induce a presupposition that Sam talked to someone and that there are reasonable alternatives that would have been good choices for this someone, as well.”.
Cleft sentences. Italian-Danish in contrast
233
reason Dufter defines the subclause as focus), and one where both the ClC and the subclause contain new information, i.e. an “all focus” type, as exemplified in: (27)
Mr President, Commissioner, it is with great sorrow that we follow the events in Nigeria. (example from Dufter 2009: 105).
Dufter also proposes a particular subtype of Prince’s “stressed it-cleft” in which also the subclause is said to contain a focus: (28)
[…] it is not the individual’s disability that needs to be corrected, it is the restraining environment that must adapt. (example from Dufter 2009: 101).
However, strictly speaking the second focus is not part of the subclause of the first CS in such cases. Rather, we are dealing with two parallel and contrasted CSs, in which both ClCs are foci of their respective CSs, a sort of explicitly stressed and contrasted focus structure. With the omission of this CS as a special type (type 1b in Dufter’s system), Dufter’s typology (Dufter 2009: 100), to which we shall return in Section 4, can be described as in Table 3: Table 3: Three information-structural types of Cleft sentences (based on Dufter 2009)
Type (i) Type (ii) Type (iii)
Cleft constituent
Embedded clause
…XFocus… …Background
… Background …XFocus… All focus
In relation to focus marking, the distinction between identifying and non-identifying CSs plays an important role, cf. e.g. Roggia (2009: 23–40) and De Cesare (2011: 253–269). As an example of an identifying CS, De Cesare cites: (29)
It is Stella that talked about Eva
which precisely presupposes a proposition with an open variable, x talked about Eva. This proposition is saturated by the ClC, Stella, which becomes focus due to its position in the matrix clause. One of De Cesare’s examples of a non-identifying CS is a structure similar to (27): (30)
It is with great pleasure that Stella talked about Eva
234
Iørn Korzen
This CS type, which Roggia (2009: 146) and De Cesare (2011: 281) term “metatextual”, does not presuppose an open proposition saturated by an identifying mechanism, but a closed proposition without any identification: Stella talked about Eva.
3.2 Cleft sentences in Italian: Structural and syntactic features Before returning to the discussion of a CS typology, we shall have a quick look at some important structural and syntactic features of Italian CSs with examples taken from the Italian EUROPARL-corpus. Due to the status of Italian as a prodrop language, the CS matrix clause has no explicit subject equivalent to the Danish det ‘it’. Therefore, the Italian corpus was searched manually for 3rd person singular and plural forms of the copula verb essere ‘be’ in all tenses and moods, a fairly time consuming task.¹⁷ By and large, modern Italian is as “permissive” as Danish regarding the syntactic categories and linguistic material of possible ClCs. For thorough surveys see e.g. Frison (2001), D’Achille et al. (2005) and De Cesare (2011: 250–253). One exception is the prepositional complement, which cannot be separated from the preposition in Italian as in Danish (and English), cf. the Danish examples (6), (41), (57), (62) and (64). Instead, the whole prepositional phrase (e.g. prepositional object) can be focalized as ClC, as in: (31)
[…] e l’autore non si è lasciato adulare da dichiarazioni aperte sulla liberalizzazione e la legalizzazione, perché non è di questo che si tratta. (ep96-04-18.txt:181) ‘[…] and the author has not let himself be flattered by open statements on liberalisation and legalisation, because it is not what this is about. [Lit. it is not with this that it has to do].’
A similar construction may not be completely ruled out in very formal and archaic Danish, but it is certainly very rare, and there are no examples in the EUROPARLcorpus. On the other hand, Italian adverbial preposition phrases occur very frequently as ClCs, see Section 5.2.3.
17 For a study on the copula essere ‘be’ in Italian CSs, see Panunzi (2009).
Cleft sentences. Italian-Danish in contrast
235
3.2.1 Explicit and implicit Cleft sentences Unlike Danish, in all cases of Italian (explicit) CSs, the subclause connective is the same: the generic subordinator che.¹⁸ However, Italian CSs have a particular characteristic feature in that the subclause can be explicit or implicit. In the first case, the verb is in a finite form and the subclause connective is the generic che; in the second case, which is limited to subject ClCs, the verb appears in the infinitive form preceded by the preposition a. This structure is by some deemed “better” than the finite version, cf. e.g. Frison (2001: 218)¹⁹ and D’Achille et al. (2005: 265), and in her own corpus of texts of mixed formality, Berretta (1994: 92) found this structure in 84.3% of the subject ClCs. However, the EUROPARL-corpus differs in that respect. Out of a mere 27 cases of subject ClC (see Section 5 and Table 4), only 8 are implicit, of which two are cited in (32) and (33): (32)
Io ho ricordato questi passaggi duri della relazione perché sono proprio questi a rendere credibile la proposta. (ep-96-05-09.txt:36) Lit. ‘I remembered these hard passages of the report because it is precisely them to give [i.e. that give] credibility to the proposal.’
(33)
Signora Presidente, nella passata legislatura fu soprattutto la commissione per l’agricoltura ad occuparsi di questo problema, mentre ora è la commissione per la protezione dell’ambiente. (ep-02-02-05.txt:258) ‘Madam President, in the previous government it was especially the Committee on Agriculture to deal [i.e. that dealt] with this problem, but now it is the commission for the protection of the environment.’
The implicit subclause may precede the ClC, e.g.
18 I.e., a che without a semantic content which may introduce a subclause of any grammatical type (completive, interrogative, relative, adverbial). Roggia (2009: 29) advances the hypothesis that this che derives from a unification of two different constructions, one with a relative subclause and one with a completive subclause, i.e. something very close to the present situation in Danish. 19 “La forma ‘a + infinito’ è senz’altro preferibile a quella con subordinata esplicita in tutti i possibili tempi della coniugazione che la copula può assumere nelle diverse situazioni contestuali.” (Frison 2001: 218). ‘The form ‘a + infinitive’ is undoubtedly preferable to the one with an explicit subordinate clause in all possible tenses of conjugation that the copula can assume in different contextual situations.’
236
(32’)
Iørn Korzen
A rendere credibile la proposta sono proprio questi passaggi. Lit. ‘To give credibility to the proposal it is precisely those passages.’
However no such cases were found in the EUROPARL-corpus.
3.2.2 The matrix verb Regarding the matrix verb in the Italian Europarl examples, and similarly to the Danish cases, the present tense is by far the most common. As sole exceptions, the corpus contains two cases of passato remoto, (33) and (34), and one of passato prossimo (35). The same tenses are found in the (explicit) subclauses: (34)
Peraltro, fu proprio a Barcellona che questo dialogo venne lanciato. (ep-02-01-16.txt:48) ‘Moreover, it was precisely in Barcelona that this dialogue was launched.’²⁰
(35)
Signor Presidente, un anno dopo Helsinki è stato con pena ed apprensione che abbiamo seguito gli eventi in Turchia, […] (ep-01-01-18. txt:123) ‘Mr President, one year after Helsinki it has been with pain and apprehension that we have been following the events in Turkey, […]’
4 A Cleft sentence typology. General observations At this point, and on the basis of the scholars mentioned in Sections 2 and 3, especially the work of Prince, Berretta, Dufter, and De Cesare, I shall make some general observations and thereafter suggest a CS typology that can serve as a foundation for a comparison between the Italian and Danish CSs in the EUROPARL-corpus. The grammatical structure of CSs gives particular pragmatic values to ClC and subclause. Due to its syntactic function as a predicative argument of the matrix clause, the ClC obtains a role as information-structural focus (in Lambrecht’s 20 For a detailed syntactic analysis of proprio and suggestions of possible translations into English, see De Cesare (2001).
Cleft sentences. Italian-Danish in contrast
237
terms)²¹, and the subclause expresses information about this focus formed as a logical presupposition, viz. information that is “immune from operators that express negation, possibility, questioning” (Delin 1995: 99). However, it is evident that not all CSs are functionally alike, and that two core factors are at play regarding the role of a given CS: the syntactic status of the ClC and the relation between the CS and the preceding co-text. Based on information-structural parameters, it would seem that we can talk about two main CS categories: (i) An identifying CS with a ClC associated with a paradigm of alternatives (Prince 1978: “stressed focus clefts”). This CS presupposes a proposition with an open variable linked to the verb, i.e. a valency argument, typically but not necessarily denoting a first-order entity in Lyons’ (1977: 438–452) terminology²², an entity in (potential) contrast to another.²³ The information conveyed by the subclause is generally known or given / inferable from the co- or context and low on communicative dynamism in the Prague School definition of this term (cf. footnote 11). The low communicative dynamism also shows in the tendency to omit the subclause of this CS type. (ii) A CS type of which the subclause conveys new information and is high on communicative dynamism, even if the information conveyed is marked as presupposed and a fact known to (some) people, although possibly not (yet) to the addressee (Prince 1978: “informative-presupposition clefts”). This CS is not (necessarily at least) identifying; the ClC does not (necessarily) presuppose a paradigm of alternatives, it may be anaphoric and/or a scene-setting
21 Cf. Lambrecht (2001: 470): “I suggest that the copula does in fact assign a functional role to its object. However, this role is pragmatic (information-structural), not semantic. Since the object of a copula normally functions as a predicate, and since predicates, unlike subjects, normally function as foci […], the occurrence of the focus phrase in the object position of the copula is naturally interpreted as a signal that its denotatum has a focus relation to the proposition in which it plays a semantic argument role […] [T]he copula, together with its empty subject, serves as a kind of focus marker for the argument of another predicator.”. Cf. also the notion of fuoco linguistico in De Cesare (2011: 73–83). 22 Lyons (1977: 438–452) distinguishes between first-order entities, which are physical objects that are located and observable in the dimension space, second-order entities, which are events, processes, states-of-affairs, etc. that occur or take place in the dimension time, and third-order entities, which are abstract entities outside of space and time such as propositions, ideas, theorems, reasons, feelings, etc., that can be predicated as ‘true’ rather than as ‘real’. 23 “In realtà il contrasto non è che un effetto della natura identificativa (cioè di identificazione esaustiva) del focus delle FS [frasi scisse], e può essere attivato o meno dallo specifico contesto in cui la costruzione è inserita.” (Roggia 2009: 100). ‘In fact, contrast is nothing other than an effect of the kind of identification (i.e. exhaustive identification) of the focus of the CS, and it can be activated or not activated by the specific context in which the construction is found’. We return briefly to the question of exhaustiveness in Section 5.2.1.3.
238
Iørn Korzen
adverbial (see below); a special subtype is the “emotion setting” ClC type: with deep regret, with great pleasure, etc. This main bipartition may be further developed by the following observations: (a) A first-order nominal ClC denotes a physical entity bound, i.e. individualised, in the dimension space, for which reason it may be more likely to appear in co(n)texts with a possible paradigm of alternatives,²⁴ pointing to a type (i) CS; (b) However, an anaphoric ClC will normally combine with a subclause that conveys new information about the anaphorised referent, which points to a type (ii) CS; cf. also Berretta (1994, 2002); (c) An adverbial scene-setting ClC is less likely to entail a paradigm of alternatives; typically, a background scenery is more rarely seen in (potential) contrast to another scenery than specific, especially first-order, entities located in such scenery.²⁵ However, also in the cases of (b) and (c) a paradigm of alternatives – and thus an interpretation of a type (i) CS – is far from ruled out, but it will generally require a special co-text that somehow expresses a (potential) contrast. Such contrast can be expressed in different ways: – by the presence of a second and parallel CS and/or ClC, cf. example (36); – by a focalizing adverb, in Nølke’s (1983) terminology un adverbial paradigmatisant,²⁶ that has the ClC in its scope, e.g. in Italian proprio ‘exactly, precisely’, soprattutto ‘especially’, precisamente ‘precisely’ etc., cf. (37), also cited as (34) above; – by a negation that has the ClC in its scope; typically a negated / rejected ClC will imply a positive alternative, cf. (38);
24 For the distinction between entities “bound” or “continuous” in a certain dimension, see also Zemach (1970). 25 Here, the notion of text pragmatic prominence, or topicality (i.e. the likelihood that a discourse referent will appear as topic), plays an important role, cf. Korzen (2000, 2001, 2008). See also Section 5.2.2 and the hierarchies cited in (69) and (70). Below, and in Section 5.2.3, we return to the notion of scene setting, which in this context is conceived in a broad sense including both spatiotemporal and argumentative “settings”. 26 “[U]n adverbial paradigmatisant introduit en tant que présupposé un paradigme d’éléments semblables à l’élément auquel il est attaché dans la phrase actuelle. […] Par ce fait, ils font partie d’un groupe de phénomènes de la nature auquel appartient par exemple le clivage, car une fonction du clivage semble justement être d’introduire un ensemble d’éléments semblables au foyer du clivage.” (Nølke 1983: 19). See also Roggia (2009: 105) and De Cesare (2011: 254–269).
Cleft sentences. Italian-Danish in contrast
–
239
the ClC may contain a numeral (e.g. date or year), whereby the naturally paradigmatic characteristics of numerals may point to potential contrasts, cf. (39):
(36)
Signora Presidente, avremmo potuto discutere della necessità di dare più forza ai principi fondatori dell’Unione, per poter difenderli tutti insieme; […]. È con la diplomazia, la persuasione, la cultura, la solidarietà, e non con le minacce, che le Istituzioni possono farsi rispettare e far rispettare i principi, […] (ep-00-02-02.txt:20) ‘Madam President, we could have discussed the need to give more strength to the founding principles of the Union, in order to defend them all together; […] It is with diplomacy, persuasion, culture, solidarity, and not through threats, that the institutions can gain respect and ensure respect for the principles, […]’
(37)
Noi abbiamo il dovere di assicurare che il Consiglio di Barcellona sia coronato da un pieno successo. […] Peraltro, fu proprio a Barcellona che questo dialogo venne lanciato. (ep-02-01-16.txt:48) ‘We have a duty to ensure that the Council of Barcelona is crowned with a complete success. […] Moreover, it was precisely in Barcelona that this dialogue was launched.’
(38)
È molto triste che a Monterrey non arriveremo con una posizione unitaria da parte dell’Unione europea e siamo preoccupati – e non contenti – del fatto che siamo ancora fermi allo 0,4 percento, forse, del prodotto nazionale lordo per quanto riguarda l’aiuto allo sviluppo. Non è così che si rafforza la governance globale! (ep-02-03-11.txt:52) ‘It is very sad that we will not go to Monterrey with a unified position from the European Union and we are concerned – and not happy – with the fact that we are still stuck at 0.4 percent, perhaps, of the gross national product for development aid. It is not this way we strengthen global governance!’
(39)
Det var i 1993 under det danske formandskab, at betingelserne for udvidelsen blev fastlagt, […] (ep-02-12-18.txt:31) ‘It was in 1993 during the Danish Presidency, that the conditions for the expansion were set, […]’
240
Iørn Korzen
We shall return to the cited focalizing possibilities in Section 5.2.1, but with these observations in mind, we can now specify CSs (i) and (ii) above and operate with a total of four different CS types: Type 1. An identifying CS with a focus associated with a paradigm of alternatives. The ClC is most often a valency argument, given, inferable or new,²⁷ typically but not necessarily denoting a first-order entity; the ClC may also be an adverbial adjunct, but will then mainly occur in a contrast explicit co-text as just seen. The information conveyed by the subclause is typically known, given or inferable or contains given, inferable or anchored entities. This subtype corresponds to Dufter’s and Berretta’s type (i); see more examples in Section 5.2.1. Type 2. A CS whose ClC is an anaphoric valency argument. The focus marking is “weaker” (Berretta’s type [ii]), the CS serves first and foremost a text cohesive purpose by focalizing a textual (anaphoric) relation. This CS can bring a discourse referent from a narrative or pragmatic background to a foreground (Berretta 1994: 99) or express a relation between different states-of-affairs (Prince 1978: 902). It differs from Type 1 in that the focus is mainly on the subclause, which typically conveys new information; see examples in Section 5.2.2. Type 3. A CS whose ClC is typically an adverbial adjunct denoting a second- or third-order entity and/or a spatio-temporal or argumentative scenesetting. The ClC is generally, but not necessarily, anaphoric, and focus is not on the ClC but on the subclause, which typically conveys new information (Berretta’s type [iii] and Dufter’s type [ii]); see examples in Section 5.2.3. Type 4. A CS whose ClC is an adverbial adjunct denoting a third-order entity: an emotional “setting” (pride, happiness, joy, sadness, regret, etc.) for a communicative or similar act. Both ClC and subclause convey new information and the CS mainly occurs in the beginning of a communication as a fixed “cliché” structure (Dufter’s type [iii]); see examples in Section 5.2.4. There is a gradual loss of focus marking function from the “classic” CS Type 1 and “downwards”. As we shall see, and as already noted e.g. by Berretta (1994, 2002) and Herslund (2005), in the case of Type 3 many matrix clauses have become (more or less) lexicalised units that serve as text cohesive elements, anaphors / connectives, e.g. 27 Including Roggia’s (2009: 140) two intermediate variants Given semiactive / inactive and anchored, see Section 3.1.
Cleft sentences. Italian-Danish in contrast
(40)
241
It was at that point that… → Then… It was for this reason that… → Therefore…
5 The Italian and Danish Europarl Cleft sentences in comparison: linguistic material and textual function One of the differences between the four CS types just described concerns the syntactic relation between the ClC and the subclause: In types 1 and 2, the ClC is typically a valency argument and in types 3 and 4, an adverbial adjunct. A comparison between the Italian and Danish CSs in the EUROPARL-corpus could therefore take its starting point in the distinction between valency and adverbial ClCs and investigate the distribution and the linguistic material of the two ClC types, and thereafter focus on their textual function based on the information-structural characteristics of ClCs and subclauses, as outlined in Section 4. The valency arguments represented in the EUROPARL-corpus are subject, object and the Italian prepositional object. In the same group I include the complement of Danish valency prepositional phrases, since these structures are typically translational equivalents of Italian prepositional objects, although strictly syntactically speaking the prepositional complement is a secondary constituent, see examples (41), (57), (62) and (64) below. Table 4 shows the distribution between valency and adverbial ClCs: number of occurrences and percentage of all occurrences. Table 4: Distribution between valency and adverbial cleft constituents in Europarl Italian ClCs
Danish ClCs
Valency arguments total
39
36.4%
168
90.8%
Subject
27
25.2%
113
61.1%
Object
10
9.3%
26
14.1%
Prepositional object
2
1.9%
0
0
Prepositional complement
–²⁸
29
15.7%
Adverbial adjuncts total
68
63.6%
17
9.2%
Total occurrences
107
100%
185
100%
28 Cf. Section 3.2.
242
Iørn Korzen
As the table shows, there is a clear difference in the distribution of valency and adverbial ClCs in the two languages. Whereas Danish valency ClCs account for more than 90% of the total number of ClCs, the Italian valency ClCs amount to 36.4%. The Italian adverbial ClCs cover the rest of the CSs, i.e. as many as 63.6% of all cases, while in the Danish corpus, the equivalent number is a modest 9.2%. We return to this phenomenon and to possible explanations for it in Section 7. In these figures, as well as in the following comparisons and analyses, the constructions mentioned in Table 5 are left out, the (a) cases because the ClC may be both a valency argument and an adverbial (and anyway the occurrences are very few), the (b) cases because their status as CSs is questionable, and the (c) cases because they are not directly derived from non-cleft structures. Table 5: Constructions not included in the comparisons Italian cases
Danish cases
(a) the interrogative CSs cited in (8) and (25) above:
1
3
(b) the completive structure cited in (13) and footnote 15:
5
10
(c) the adverbial / temporal type cited in (1) and (7):
6
5
Adding the numbers of Table 5 to the total number of occurrences in Table 4, we obtain the totals given in Section 1, Table 1.
5.1 The linguistic material of valency and adverbial cleft constituents The two (Italian) ClCs functioning as prepositional objects (cited as examples [31] and [73]) consist of prepositional phrases, whereas the other valency arguments consist of nominal material: NPs (with common or proper nouns) and pronouns. The adverbial ClCs consist of adverbs, prepositional phrases or (in Italian) the gerund. Tables 6 and 7 show the distribution of linguistic material in the valency arguments and in the adverbial adjuncts, respectively.
Cleft sentences. Italian-Danish in contrast
243
Table 6: Linguistic material of valency cleft constituents in the EUROPARL-corpus Italian CSs
Danish CSs
Subject NP with a definite article NP with a demonstrative article NP with an indefinite article proper noun personal pronoun demonstrative pronoun
27 15 5 0 4 1 2
100.0% 55.6% 18.5% 0% 14.8% 3.7% 7.4%
113 46 10 26 13 4 14
100.0% 40.7% 8.8% 23.0% 11.5% 3.5% 12.4%
Object NP with a definite article NP with a demonstrative article NP with an indefinite article proper noun personal pronoun demonstrative pronoun
10 1 1 1 0 0 7
100.0% 10.0% 10.0% 10.0% 0.0% 0.0% 70.0%
26 0 3 13 0 0 10
100.0% 0.0% 11.5% 50.0% 0.0% 0.0% 38.5%
Prepositional object Prepositional phrase
2 2
100.0% 100.0%
0 0
Prepositional complement NP with a definite article NP with a demonstrative article NP with an indefinite article proper noun personal pronoun demonstrative pronoun
29 4 3 11 1 0 10
100.0% 13.8% 10.3% 37.9% 3.4% 0.0% 34.5%
Table 7: Linguistic material of adjunct cleft constituents in the EUROPARL-corpus Italian CSs
Danish CSs
Adverbial adjuncts total
68
100.0%
17
100.0%
adverbs
4
5.9%
11
64.7%
prepositional phrases
62
91.2%
6
35.3%
gerunds (inexistent in Danish)
2
2.9%
0
0.0%
Table 6 shows that in both languages the subject is the most frequent valency ClC, and the NP with a definite article is the most frequent subject ClC.²⁹ NPs with 29 Probably the definite article is the most frequent determiner of subject NPs in general and not just of subject ClCs, due to the typical Theme-Rheme distribution.
244
Iørn Korzen
a definite or a demonstrative article indicate that the entity referred to by the NP is identifiable (Korzen 1996, 2008), and like the personal and demonstrative pronouns they refer anaphorically or cataphorically to a co-textual unit or exophorically to an extratextual entity that is either present in the context (deixis) or generally known.³⁰ Examples of NPs with definite articles referring to extratextual generally known entities can be seen in (36) above and (44) below, anaphoric demonstrative pronouns in (31) and (73), and deictic pronouns in (45) and (46). Cataphoric reference is generally rare and no cases were found in the Italian corpus. The Danish corpus contains one example of a cataphoric pronoun: (41)
Ja, men er det ikke netop det, åbenheden går ud på – at nogen skal kunne få en afgørende indflydelse på beslutningsprocessen? (ep-0105-02.txt:82) ‘Yes, but is it not precisely that which transparency is all about – that anyone may have a decisive influence on the decision-making process?’
In (41), the pronoun det ‘that’ refers to the following completive phrase at nogen skal kunne få… ‘that anyone may have…’. A definite article may refer cataphorically to a postmodifying relative clause; the Danish corpus has the following example of such reference: the definite article de [fællesskabsinstrumenter] ‘the [Community instruments]’ refers to the relative clause der er blevet indført… ‘that have been introduced…’ in (42): (42)
Det er især de fællesskabsinstrumenter, der er blevet indført med henblik på gennemførelsen af det indre marked, der tjener til at afskaffe kontrollen med varer og bagage ved de indre grænser. (ep-96-04-16.txt:159) ‘It is especially the Community instruments that have been introduced for the implementation of the internal market that serve to abolish the control of goods and luggage at internal borders.’
NPs with an indefinite article indicate co- and contextually new referents, and as Table 6 shows, these are fairly frequent in Danish (amounting to precisely 50% of the object cases), whereas the Italian corpus contains just one case of an indefinite article in an object NP: 30 In Korzen (2006), I discuss the three deictic valencies in these NPs and pronouns: anaphoric, cataphoric and exophoric, the exophoric being limited to specific entities in the case of demonstrative NPs and pronouns.
Cleft sentences. Italian-Danish in contrast
(43)
245
È un esame di coscienza che, giorno per giorno, sui fatti operativi dovremmo fare, perché è chiaro che la globalizzazione sta divaricando la nostra società, […] (ep-00-02-15.txt:65) ‘We should examen our conscience [Lit. It is an examination of conscience that we should do] day by day on what we are actually doing because it is clear that globalisation is increasing the divisions in our society, […]’
Table 7 shows the distribution of linguistic material of adverbial ClCs, and not only are the adverbial ClCs much more frequent in Italian than in Danish, the linguistic material also shows significant cross-linguistic differences; especially regarding the prepositional phrases, which in Italian cover more than 90% of all adverbial ClCs, and in Danish a mere 35% (of an already low number of cases). This gives us a first impression of fundamental differences in the use of CSs in the two languages, an impression that is confirmed when we look more closely into the single Italian and Danish cases in Section 5.2.3.
5.2 The textual function of valency and adverbial cleft constituents In order to categorise the Europarl CSs between the four CS types outlined in Section 4, it is necessary to combine the subdivision of the CSs according to the syntactic and, possibly, anaphoric function of the ClC with an analysis of the information-structural features of ClCs and subclauses.
5.2.1 The (“classic”) Type 1 Cleft sentence The “classic” identifying Type 1 CS contains a (typically valency) ClC associated with a potential or explicit paradigm of alternatives, and the subclause conveys information which is mainly (co- or contextually) known, given or at least inferable. In oral speech the ClC is typically stressed. This CS category is quite heterogeneous, and a full description of all subtypes is lengthier than the descriptions of the other three CS types. As stated in Section 4, this type is found in all cases in which a ClC contrast is rendered explicit and/or a focalizing adverb or a negation has the ClC in its scope, see Sections
246
Iørn Korzen
5.2.1.1–3. Also repeated ClCs or subclauses and ClCs introducing co- and contextually new discourse referents will typically belong to Type 1, see Sections 5.2.1.4–5.
5.2.1.1 Explicit contrast Cases with an explicit contrast are quite frequent. The contrast can be conveyed by a parallel ClC (often in an elliptic structure), as in (44), or in other ways by the surrounding co-text, cf. (45) and (46), where the ClCs are deictic pronouns: (44)
Signor Presidente, avrei voluto dire al relatore che si sbaglia. We are on trial: non è il Consiglio ad essere messo alla prova oggi, bensì il Parlamento. (ep-02-02-06.txt:43) ‘Mr President, I would have wanted to tell the speaker that he is wrong. We are on trial: it is not the Council to be put [that is put] to the test today, but the Parliament.’
(45)
Hr. formand, jeg mener, at der er en oversættelsesfejl i den danske tekst, idet der under punkt 7, “Hævelse af mødet”, står: “Aldersformanden hævede mødet kl. 19.30”. Jeg mener, at det var Dem, hr. formand, der hævede mødet, og det er også, hvad der står i den franske version. (ep02-01-16.txt:3) ‘Mr President, I believe that there is a translation error in the Danish text, since item 7, “Closing of the meeting,” reads: “The oldest Representative adjourned the meeting at 19.30.”. I think it was you, Mr President, who adjourned the meeting, and that is also what the French version says.’
(46)
Det er især uforståeligt, at den britiske regering, som i første omgang traf de rigtige beslutninger, lod 5 år gå, før den implementerede dem. Det er årsagen til, at BSE-sygdommen fik det omfang, den fik, og blev overført til Frankrig, Schweiz og Portugal. Men det er ikke os, der skal vælte den britiske regering, det må andre gøre. (ep-97-02-18.txt:119) ‘It is especially difficult to understand why the British government, that initially took the right decisions, allowed 5 years to pass before it implemented them. This is the reason why BSE gained such proportions and
Cleft sentences. Italian-Danish in contrast
247
spread to France, Switzerland and Portugal. But it is not us who must overthrow the British government, others must do that.’ Very often in such cases, the CS and/or the contrasted construction will entail a textual topic shift.
5.2.1.2 Focalizing adverbs In other cases, a focalizing adverb ensures the Type 1 reading of the CS. In the following Danish case, the ClC is the adverb her ‘here’ used anaphorically; thus, without the focalizer, we would probably have interpreted a Type 3 CS: (47)
Men det er netop her, vi finder det egentlige rationale i mantraet om sikkerhed, frihed og retfærdighed. (ep-00-01-19.txt:113) ‘But it is precisely here that we find the true rationale of the mantra of freedom, security and justice.’
The same is true of the following Italian case, where the ClC is adverbial, but not used anaphorically: (48)
È proprio con spirito costruttivo che vorrei sottoporle alcuni dubbi e critiche che il gruppo Verts / ALE solleva sulla prospettiva del programma da lei appena esposto. (ep-02-01-16.txt:54) ‘It is precisely in a constructive spirit that I would like to submit some doubts and criticism that the Verts / ALE Group raises regarding the prospect of the programme you have just laid out.’
Without proprio ‘precisely’, we would probably read a Type 4 CS in this case. The focalizing adverbs that occur in the Italian and Danish corpus are cited in Table 8 together with the number of tokens in the corpus:
248
Iørn Korzen
Table 8: Focalizing adverbs in the Italian and Danish Cleft sentences Italian focalizers
Danish focalizers
proprio ‘precisely’ appunto ‘exactly’ solo ‘only’ soltanto ‘only’ soprattutto ‘especially’ almeno ‘at least’ anche ‘also’
13 2 3 1 2 1 1
netop ‘precisely’ udelukkende ‘exclusively’ kun ‘only’ alene ‘only’ især ‘especially’
5 1 8 1 2
Total occurrences percentage of all CSs occurrences with adverbial ClCs
23 19.3% 9
Total occurrences percentage of all CSs occurrences with adverbial ClCs
17 8.4% 1
As the table shows, the focalizers are proportionally more than twice as frequent in Italian as in Danish, occurring in 19.3% of all Italian CSs against 8.4% of the Danish CSs. Furthermore, they occur more frequently with adverbial ClCs in Italian (9 cases) than in Danish, where the only occurrence with an adverbial ClC is cited in (47). The table also shows that the most frequent Italian focalizer is proprio ‘precisely’. Of the 13 cases of proprio, four occur with subject ClCs, see an example in (32), one with an object ClC and eight with adverbial ClCs, see an example in (37). The focalizer almeno ‘at least’ occurs with an adverbial ClC, solo ‘only’, soltanto ‘only’, soprattutto ‘especially’and anche ‘also’ only with subject ClCs, and the two cases of appunto ‘exactly’ have an object ClC in their scope; one case is the following: (49)
Qualsiasi organizzazione, infatti, di qualunque genere ha il dovere di adattarsi costantemente ai cambiamenti del mondo che la circonda. È appunto quello che stiamo cercando di fare in questo momento. (ep01-02-13.txt:8) ‘Any organization, in fact, of any kind has a duty to constantly adapt to the changes of the world around her. It is exactly what [Lit. that which] we’re trying to do at this time.’
The Danish focalizers all occur with subject ClCs, except for one case of kun ‘only’ with an object ClC, three cases of netop ‘precisely’ with a prepositional complement, and the case of netop with an adverbial cited in (47). The comparative expressions in quanto tale (Italian) / som sådan (Danish) ‘as such’, which are not generally considered as focalizers, may follow a ClC and –
Cleft sentences. Italian-Danish in contrast
249
som sådan typically together with a negation – create a paradigm; this happens in one Italian and one Danish case of the corpus: (50)
È la costante imparzialità della Commissione, in quanto tale, che va garantita. (ep-97-02-18.txt:137) ‘It is the continued impartiality of the Commission, as such, that needs to be ensured.’
(51)
Men jeg lytter også til de irske politikere, både ja- og nejsiden, som siger, at det ikke var udvidelsen som sådan, der var problemet i Irland, […] (ep-01-06-13.txt:40) ‘But I also listen to Irish politicians, both on the “yes” and the “no” side, who say that it was not the expansion as such that was the problem in Ireland, […]’
With very few exceptions, the valency ClCs that occur with a focalizer are anaphoric, and the focalizer ensures a Type 1 interpretation rather than a Type 2 reading (see Section 5.2.2). The following case is one of the exceptions where a subject ClC refers to a co- and contextually new entity, thus making a Type 1 CS the most probable reading even without the focalizer især ‘especially’ (see also Section 5.2.1.6): (52)
Nu er steningerne ganske vist sat i bero, men alle ved jo, at der i efteråret 2002 var en række store demonstrationer i Iran for menneskerettigheder og demokrati og imod henrettelser og steninger. Det var især en række studerende, der stod i spidsen for demonstrationerne, og når jeg rejser spørgsmålet nu, så er det fordi, at fire af disse studerende er blevet dømt ved en domstol i Shiraz til at få hugget arme og ben af. (ep-03-01-13. txt:44) ‘It is true that the stonings have now been suspended, but everyone knows that in the autumn of 2002, there were a number of large demonstrations in Iran in favour of human rights and democracy and against executions and stonings. It was especially a number of students who spearheaded the demonstrations, and when I raise the issue now, it is because that four of these students have been sentenced by a court in Shiraz to have their arms and legs cut off.’
250
Iørn Korzen
5.2.1.3 Combined explicit contrast and focalizer Very often an explicit contrast is combined with a focalizing adverb, in (53) solo, anche ‘only, also’, in (54) kun ‘only’, respectively: (53)
Questo Libro bianco è una risposta alle sfide che l’ampliamento ci impone, ampliamento che ci obbliga a ripensare al funzionamento di tutte le nostre Istituzioni, a ripensare alle nostre stesse politiche – a tutte le nostre politiche – a ripensare soprattutto a cosa dobbiamo continuare a fare a livello di Unione quando i paesi membri saranno venticinque o trenta […]. Ma non è solo l’ampliamento che ci spinge a questo ripensamento: come detto poc’anzi, è anche la stessa globalizzazione dell’economia e della politica. (ep-00-02-15.txt:17) ‘This White Paper is a response to the challenges that the expansion forces upon us, an expansion that forces us to rethink the functioning of all our institutions, to rethink our own policies – all our policies – especially to rethink what we must continue to do at the EU level when member countries will total twenty-five or thirty […]. But it is not only the expansion that forces us to rethink this: as mentioned earlier, it is also the globalization itself of economy and politics.’
(54)
Indtil nu er det kun meninger, De straffer, og ikke handlinger. Hvis der var tale om handlinger, så var det måske Danmark, der skulle straffes for et indvandrerforbud, som FPÖ ikke er kommet igennem med. Det er altså meninger, De har straffet, og ikke handlinger. (ep-00-03-15.txt:215) ‘Until now, it is only opinions you are punishing, and not actions. If it were actions, perhaps it would be Denmark who should be punished for an immigration ban that the FPÖ has not been able to implement. It is thus opinions you have punished and not actions.’
A restrictive relative clause can play a delimiting role similar to that of a contrasted CS. This is the case in example (42), repeated here, which nevertheless has been supplied with a focalizer: (55)
Det er især de fællesskabsinstrumenter, der er blevet indført med henblik på gennemførelsen af det indre marked, der tjener til at afskaffe kontrollen med varer og bagage ved de indre grænser. (ep-96-04-16.txt:159)
Cleft sentences. Italian-Danish in contrast
251
‘It is especially the Community instruments that have been introduced for the implementation of the internal market that serve to abolish the control of goods and luggage at internal borders.’ The presence of focalizers such as those mentioned in Table 8 is a strong argument against the hypothesis of a presupposition of exhaustiveness in CSs. If identifying CSs contained a semantic content of exhaustiveness, such focalizers would neither be necessary nor, in reality, compatible with these CSs, as argued also by e.g. Dufter (2009: 95–98) and De Cesare (2011: 254–263). On the basis of an analysis of the focalizers only / also and their Italian counterparts solo / anche in English and Italian CSs, De Cesare and Garassino (In press) conclude that exhaustiveness is best accounted for in terms of a conversational implicature, an inherently pragmatic – and not semantic – phenomenon. Moreover, it is “a discrete notion that can have different degrees of prominence depending on the context” (De Cesare and Garassino [In press], Section 5.3). According to Roggia (2009: 99), the content of exhaustiveness is present in the identifying CSs but deleted in the presence of focalizing adverbs.
5.2.1.4 Negated cleft constituents A negated ClC is normally in explicit or implicit contrast to a parallel positive alternative and will therefore signal a Type 1 CS. In (56) (part of which is cited in [31]), the following co-text contains such an explicitation, whereas this is not the case in (57): (56)
Contrariamente a diversi pareri espressi da altre commissioni, è una relazione equilibrata negli intenti, e l’autore non si è lasciato adulare da dichiarazioni aperte sulla liberalizzazione e la legalizzazione, perché non è di questo che si tratta. La Commissione, con la sua proposta, mira a mettere i paesi in via di sviluppo nelle condizioni di contrastare efficacemente la tossicodipendenza […] (ep-96-04-18.txt:181) ‘Contrary to many opinions expressed by other committees, it is a report which is balanced in its intent, and the author has not let himself be flattered by open statements on liberalisation and legalisation, because it is not what this is about. The Commission, in its proposal, aims to put developing countries in a position to effectively combat drug addiction […]’
(57)
Hr. formand, den liberale gruppe vil også give udtryk for, at vi bakker meget stærkt op om de initiativer, som De, hr. Kinnock, har iværksat på perso-
252
Iørn Korzen
naleområdet. Det er ikke nogen let opgave, De er i gang med, og derfor er det vigtigt, at Parlamentet giver sin fulde opbakning. (ep-01-09-03.txt:118) ‘Mr President, the liberal group will also like to express that we very strongly support the initiatives that you, Mr Kinnock, have launched regarding staff. It is not an easy task you have embarked on, and therefore it is important that Parliament gives its full support.’
5.2.1.5 Repetitions The repetition of a matrix clause normally implies a highlighting of the ClC constituent, and it seems reasonable to include such cases among the Type 1 CSs: (58)
[…] og det bliver også stadigvæk tydeligere, at skal der opbygges et internationalt retssystem, som er holdbart, og som gælder for store såvel som for små, er det EU, der skal spille den afgørende rolle, det er EU, der skal være garanten for, at store lande såvel som små lande følger et sæt internationale spilleregler. (ep-03-02-12.txt:219) ‘[…] and it also becomes more and more clear that if we are to build an international legal system that is durable and applicable to both large and small countries, it is the EU that must play the crucial role, it is the EU that must guarantee that large countries as well as small countries follow a set of international rules.’
Also a repeated ClC can be negated and contrasted: (59)
Det er ikke EU, der skal skaffe os af med en regering. Det er ikke EU, der skal gennemføre en human udlændingepolitik, det er danskerne selv – we, the people. (ep-02-02-04.txt:53) ‘It is not the EU that should rid us of a government. It is not the EU that should implement a humane foreign policy, it is the Danes themselves – we, the people.’
In the following example an anaphoric reference, alle 20 ‘all 20’ is changed into a deictic reference jer 20 ‘you 20’ in the CS structure. Therefore, and because there is a very clear stress on the ClC, activated also by the repetition of the matrix clause, the structure should be read as a Type 1 CS:
Cleft sentences. Italian-Danish in contrast
(60)
253
Hr. Prodis Kommission har ikke skyld for fortidens synder, men alle 20 bærer nu sammen på ansvaret for, at der ikke bliver ryddet op, gjort rent og startet på en frisk. Det er jer 20, der har ansvaret for, at Kommissionens forslag til forordning om åbenhed vil lukke dokumenter, som i dag er åbne. Det er jer 20, der har ansvaret for, at vi som parlamentsmedlemmer fortsat ikke får, hvad vi beder om. I kører efter enevældens principper. (ep-00-04-11.txt:36) ‘Mr Prodi’s Commission is not responsible for past sins, but all 20 are collectively responsible for not having tidied up, cleaned up and started afresh. It is you 20, who are responsible for the fact that the Commission’s proposal for a regulation on transparency will close documents that are currently open. It is you 20 who are responsible for still not ensuring that we as MEPs get what we ask for. You are following the principles of absolutism.’
In (61), the ClC contrast emerges from the preceding co-text and from the repetition of the subclause: (61)
Stabilitetspagten siger, at underskuddet på finansloven i almindelige år højst må være 1%. Der er ingen krav om, at arbejdsløsheden skal være under 5% eller – endnu bedre – under 2 eller 1%. Det er prisen på pengene, som ikke må stige, det er det offentliges forbrug, som ikke må stige. (ep-00-03-13.txt:27) ‘The Stability Pact says that the budget deficit in normal years must not exceed 1%. There is no requirement that unemployment should be below 5% or – even better – less than 2 or 1%. It is the price of money that must not increase, it is government consumption that must not increase.’
5.2.1.6 The cleft constituent introduces a new entity or concept Finally, a valency ClC introducing a co- and contextually new and non-inferable entity will normally be focused (stressed in oral speech) and license a potential presence of alternatives; this is the case in (43) above, the only case of an Italian ClC NP with an indefinite article. In fact, as we have seen, cases of completely new ClCs are very rare in the Italian corpus; typically, Italian Europarl ClCs are somehow inferable, semiactive or anchored in the preceding co-text or situation, whereby the CS structures (different from the ones cited in Sections 5.2.1.1–5) acquire the particular text cohesive function that is typical of Types 2 and 3.
254
Iørn Korzen
In this respect, the Danish corpus is very different. CSs are often used to introduce and at the same time highlight new discourse referents, which may become topics of the following co-text. Like the CS subtype described in 5.2.1.1, this subtype typically entails a textual topic shift: (62)
EU’s kompetencer skal afgrænses, så det er grænseoverskridende emner, man beskæftiger sig med. (ep-96-06-19.txt:17) ‘EU’s competences should be delimited, so it is cross-border issues we are dealing with.’
(63)
[…] jeg vil gerne udtrykke min glæde over, at der er sat en offensiv dagsorden for beskæftigelsen og den økonomiske udvikling ved topmødet i Lissabon. Det er en dagsorden, der betoner, at det er en stærk og dynamisk økonomi og en bedre konkurrenceevne, der skal sikre målet om større beskæftigelse i EU. (ep-00-03-13.txt:30) ‘[…] I am very pleased that a proactive agenda for employment and economic development has been set for the Lisbon Summit. It is an agenda which emphasizes that it is a strong and dynamic economy and improved competitiveness that must ensure the goal of increasing employment in the EU.’
Another very frequent function of Danish CSs is to define or evaluate previously introduced referents. Thus the definition or evaluation, i.e. this point of view on the entity in question, is new, and the ClC appears with an indefinite article. This kind of definition or evaluation is often conveyed by very short sentences,³¹ and generally the content of the subclause is contextually known or inferable: (64)
Det er en appel til regeringskonferencen, der er tale om. (ep-97-01-30.txt:19) ‘It is an appeal to the IGC we are talking about.’
(65)
Hr. formand, det er en gammel sag, vi debatterer i dag. (ep-98-02-16. txt:20) ‘Mr President, it is an old case we are debating today.’
31 On sentence length in Danish and Italian and the typically much shorter sentences in Danish than in Italian, cf. e.g. Korzen (2009), Korzen and Gylling (2012).
Cleft sentences. Italian-Danish in contrast
(66)
255
Hr. formand, det er en tragedie, vi behandler her. (ep-01-06-14.txt:102) ‘Mr President, it is a tragedy we are treating here.’
See also example (5).
5.2.1.7 Type 1 Cleft sentences. Conclusion and a few statistics In all the cases cited in Sections 5.2.1.1–6, the ClC is explicitly or implicitly associated with a paradigm of alternative referents, and the CS has an identifying function. In a small number of cases, the ClC appears without any explicit indications of possible alternatives, nor is the ClC anaphoric, as in Type 2 below. In many of such cases, the ClC indicates a generally well-known EU institution such as the Commission or the Parliament, and given the context, there are, I believe, good reasons for categorising these cases as Type 1 cases, since they express a precise identification of the institution in question with specific, even if implicit, alternatives, and the subclause typically expresses co- or contextually known, given or inferable information: (67)
Jeg har opfordret Kommissionen til at stille forslag om udfasning af fremtidige EU-midler til energisektoren i Kosovo, hvis den nuværende situation fortsætter. Det er yderst beklageligt, at agenturet har indgået flere underhåndsaftaler om ansættelse uden om udliciteringsreglerne. Det er imidlertid Kommissionen, som har tvunget agenturerne til at underskrive kontrakterne. (ep-03-04-08.txt:9) ‘I have asked the Commission to propose phasing out future EU funding of the energy sector in Kosovo if the present situation continues. It is extremely regrettable that the Agency has concluded several private agreements for employment outside of the outsourcing rules. However, it is the Commission that has forced the agencies to sign the contracts.’
(68)
[…] se questa funzione informativa non è espletata dall’autorità giudiziaria del luogo in cui avvengono i fatti, hanno il dovere e il diritto di espletarla i membri direttamente, in quanto è questo Parlamento che deve decidere quali disposizioni si debbano applicare, […] (ep-01-02-12.txt:50) ‘[…] if this information function is not performed by the judicial authority of the place where the events occur, the members have the duty and the right to perform it directly, as it is this Parliament that must decide which regulations are to be applied, […]’
256
Iørn Korzen
However, in such cases, as well as in the cases cited in Section 5.2.1.6, where there is no explicit contrast, the focus marking can be considered as less strong than in the cases mentioned in Sections 5.2.1.1–5. Counting all cases of CS types mentioned in Sections 5.2.1.1–6 as well as cases of the kind cited in (67) and (68), we arrive at a total of 33 Italian cases of Type 1 CSs, i.e. out of the total of 107 Italian CSs included in the survey, a percentage of 30.8%. Performing the same operation on the Danish corpus, we obtain 95 CSs out of a total of 185 cases, i.e. 51.4%. The considerable difference is especially caused by the frequency of the indefinite structures cited in Section 5.2.1.6. As for the other Type 1 CSs mentioned, the Danish corpus totalled 60 cases, i.e. 32.4%, indicating a much smaller difference between the two languages, cf. Table 9. Table 9: Occurrences of the Type 1 Cleft sentences in Europarl Italian ClCs
Danish ClCs
Type 1 CSs total
33
30.8%
95
51.4%
ClCs cited in sect. 5.2.1.1–5
32
29.9%
60
32.4%
ClCs cited in sect. 5.2.1.6
1
0.9%
35
18.9%
5.2.2 The Type 2 Cleft sentence The Type 2 CS has an anaphoric valency argument as ClC, and like the Type 1, this CS is identifying in that the ClC is seen in (at least potential) opposition to alternatives. However, unlike Type 1 the subclause conveys new information about a previously mentioned entity, and what is focalized here is primarily a textual, anaphoric relation. This CS is particularly suitable to anaphorise “weak” antecedents, i.e. antecedents that are placed in low positions in syntactic, semantic and/or text pragmatic hierarchies (Korzen 2001, 2008). Two important hierarchies regard syntactic function / semantic role and semantic individuation, respectively; the further to the right on the scales, the lower the position: (69)
syntactic function / semantic role subject / Agent > dative / Experiencer > object / Patient > non-primary constituents
Cleft sentences. Italian-Danish in contrast
(70)
257
semantic individuation Countable first-order entities, singular > Countable first-order entities, plural > Uncountable entities > Second-order entities > Third-order entities³²
Two examples: (71)
La Commissione rimane convinta che il principale responsabile di questa situazione è la sovraccapacità della flotta comunitaria: è quindi la flotta che deve essere ridotta. (ep-97-01-16.txt:9) ‘The Commission remains convinced that the main cause of this situation is the overcapacity of the EU fleet: it is therefore the fleet that should be reduced.’
(72)
Siamo consapevoli dei problemi legati a questa fase di transizione ed è questa transizione che ha creato più preoccupazione nel nostro personale […] (ep-00-03-01.txt:16) ‘We are aware of the problems associated with this transition phase and it is this transition that has created more concern among our personnel […]’
In (71), the antecedent, flotta comunitaria ‘EU fleet’, is the prepositional complement of an attributive prepositional phrase, i.e. a tertiary constituent, which means the lowest position on the scale in (69). In (72), the antecedent transizione ‘transition’ is even lower in the syntactic hierarchy, and furthermore, transizione is a second-order entity, i.e. it is low also on the semantic scale in (70). Due to the anaphorisation both antecedents are topicalised, i.e. moved from a textual background to a foreground. Very often, the ClC is a resumptive anaphor, i.e. an anaphor that encapsulates a whole preceding clause or sentence or a part of a clause or sentence, thereby topicalising the content. The antecedent is thus a third-order entity, cf. the scale in (70), which requires a particularly “strong” anaphor, if a pronoun, then normally a demonstrative pronoun; if an NP, then often an NP with a demonstrative article (Korzen 2000, 2006):
32 Two other scales regard identifiability (+ identifiable > – identifiable) and referentiality (Deictic > Specific > Generic > Non-specific), cf. Korzen (2000, 2001, 2008). The scales are inspired by the topicality hierarchies proposed by Givón (1976) and Herslund (1996: 79), the transitivity scales proposed by Hopper & Thompson (1980), and the scales for object marking discussed by Bossong (1991).
258
(73)
Iørn Korzen
Evidentemente, la questione dell’immigrazione è il fatto principale dietro il quale si nascondono poi tutte le cause di razzismo e di xenofobia, ed è su questo che deve rivolgersi tutta l’attenzione dei governi europei. (ep-97-01-29.txt:55) ‘Clearly, the issue of immigration is the main thing behind which all the causes of racism and xenophobia are hidden, and it is to this that all the attention of European governments must be directed.’
(74)
Der mangler jo en politisk overbygning til at styre den økonomiske politik. Dette er baggrunden for den franske regerings forslag om, at EU-eurosystemet udstyres med en egentlig økonomisk regering. Det er disse politiske konsekvenser, der er så afgørende, når det drejer sig om den danske befolknings tilslutning. (ep-00-05-17.txt:28) ‘We lack a political superstructure to manage the economic policy. This is the reason for the French Government’s proposal that the EU euro system be equipped with a real economic government. It is these political consequences that are so crucial when it comes to the consent of the Danish people.’
The associative anaphoric ClCs, i.e. the ClCs that designate entities associable with the entities designated by the antecedents, are especially frequent in the Danish texts. In the following case, the antecedent is udbud af materiale på Internettet ‘Internet sale of material’, and the anaphor den udbydende ‘the supplier’: (75)
Hvis vi f.eks. vil børnepornografien til livs, er det nødvendigt, at der ved udbud af materiale på Internettet altid afleveres en adresse, en personlig kode eller lignende, således at det er den udbydende, der har ansvaret. (ep-97-01-15.txt:207) ‘If for example we want to get rid of child pornography, it is necessary that in all cases of Internet sale of material there must be delivered an address, a personal code or the like so that it is the supplier who is responsible.’
Table 10 shows the distribution of the different ClC types in the Italian and Danish Type 2 CSs. In this regard, the two languages are not far apart. They are further apart when it comes to the percentage of all Type 2 occurrences of the total number of CSs. Here, Italian Type 2 occurrences amount to 16.8% of the total number of Italian CSs included in the survey, and Danish Type 2 occurrences to 40.5% of all Danish CSs included, see also Table 13 in Section 7.
Cleft sentences. Italian-Danish in contrast
259
Table 10: Distribution of the Type 2 cleft constituents Italian ClCs
Danish ClCs
coreferential anaphors
5
27.8%
25
33.3%
associative anaphors
4
22.2%
17
22.7%
resumptive anaphors
9
50.0%
33
44.0%
Type 2 CS total
18
100%
75
100%
5.2.3 The Type 3 Cleft sentence The Type 3 CS has typically an adverbial ClC denoting a second- or third-order entity, and/or a spatio-temporal or argumentative scene-setting. In these cases, the centre of attention is not the ClC but the subclause, which generally conveys new information. Whereas Type 1 and 2 CSs were more frequent in Danish, we find the opposite situation regarding Type 3 due to the high frequency of one particular ClC, viz. the prepositional phrase per questo ‘for this [reason]’. This phrase, with the resumptive anaphor questo, expresses the third-order entity reason and is very frequent in Italian argumentative texts such as the EUROPARL-corpus: (76)
Queste possibilità risultano tuttavia perlopiù sconosciute e la Commissione ritiene pertanto necessario valorizzarle e chiarirne la portata. È per questo che ci proponiamo di adottare, quanto prima, strumenti finalizzati al perseguimento di questo scopo. (ep-99-01-27.txt:188) ‘These possibilities, however, are mostly unknown and the Commission therefore considers it necessary to enhance them and clarify their scope. It is for this [reason] that we propose to adopt as soon as possible instruments for the pursuit of this goal.’
The structure è per questo (che) is a very good example of a “weakened” and lexicalised CS matrix whose role is reduced to text structuring functions as a text cohesion device (Berretta 1994, 2002; Herslund 2005 on parallel structures in French), in this case an argumentative connective expressing consequence synonymously with therefore. Also in a cross-linguistic comparison, it is noteworthy that such “weakening” of CSs is a much more frequent phenomenon in Italian than in Danish. The ClC per questo occurs as many as 26 times in the Italian
260
Iørn Korzen
corpus corresponding to 21.8% of all the CSs in the Italian corpus. In 11 cases it appears sentence initially after a full stop as in (76), in 13 cases in the initial position of a main clause after the conjunction ed ‘and’, as in (77): (77)
La presente relazione raggiunge questo obiettivo soltanto a metà, ed è per questo che la valutazione finale del nostro gruppo su di essa non sarà unanime. (ep-00-04-12.txt:41) ‘This report achieves this goal only halfway, and it is for this [reason] that our group’s final evaluation of it will not be unanimous.’
In one case è per questo occurs after the conjunction ma ‘but’ and in one case after a colon. The synonymous structures, è per ciò che and è per questo motivo che, each occur once in the corpus. The Danish corpus contains four cases of the equivalent ClC derfor ‘therefore’: (78)
Parlamentet skal give decharge til Kommissionen, og det er derfor, vi primært skal holde Kommissionen ansvarlig. (ep-01-04-03.txt:62) ‘Parliament must grant discharge to the Commission, and it is therefore we primarily hold the Commission accountable.’
Other adverbial ClCs indicate background settings of other kinds, especially time or space (possibly in a metaphorical sense), manner, cause, means or the like. Also in these cases, the ClC very often contains an anaphoric constituent and marks the transition from one text segment to another: (79)
Signor Presidente, la Convenzione che sta per aprirsi non assolverà il suo mandato se non darà risposte convincenti su un punto cruciale: come garantire la democraticità del processo di costruzione europea. È di qui che parte la relazione, approvata dalla commissione per gli affari costituzionali e oggi sottoposta al giudizio del Parlamento in sessione plenaria. (ep-02-02-06.txt:127) ‘Mr President, the Convention that is about to begin does not fulfil its mandate if it fails to give convincing answers on a crucial issue: how to guarantee the democratic nature of the European integration process. It is from here that the report approved by the Committee on Constitutional Affairs and today submitted for the evaluation of Parliament in plenary session takes its starting point.’
Cleft sentences. Italian-Danish in contrast
(80)
261
Nu har den albanske befolkningsgruppe fået sine fulde borgerlige rettigheder. Det er sådan, man løser mindretalsproblemer. Jeg kommer selv fra et grænselandsområde mellem Danmark og Tyskland, […] (ep-0110-04.txt:96) ‘Now the Albanian community has obtained full civil rights. It is this way you solve minority problems. I myself come from a border area between Denmark and Germany, […]’
(81)
La globalizzazione, la Commissione la vuole, la ritiene positiva perché sta portando verso un minimo di decenza miliardi di persone che erano emarginati dal mercato mondiale; perché è grazie a questa globalizzazione che Cina e India e, ripeto, alcuni miliardi di persone si stanno svegliando. (ep-00-02-15.txt:65) ‘The Commission wants globalisation, it is in favour of it because it helps billions of people achieve a minimum of decency, people who were marginalised from the world market, because it is thanks to this globalisation that China and India and, I repeat, several billion people are waking up.’
As shown in Section 5.1, Table 7, the Italian corpus contains two gerund clause ClCs. One of these, cited in (82), is used anaphorically and in comparison with another gerund clause; the two antecedents, aprire i mercati ‘to open / opening up markets’ and concedere aiuti ‘to grant aid’, are written in bold and the anaphoric gerunds in bold italics: (82)
Aprire i mercati significa mettere in discussione le posizioni di vendita di fortissime lobby, quella agricola in particolare, in Europa e negli Stati Uniti, […] È una certa politica difficile, molto più difficile – ripeto – che quella di concedere aiuti, cosa della quale magari nessuno – parlo dei contribuenti – si accorge; ma è aprendo i mercati, molto più che concedendo aiuti, che si possono offrire a centinaia di milioni di uomini reali e durature possibilità di affrancarsi dalla fame e dalla povertà […] (ep-03-01-30.txt:14) ‘Opening up markets means questioning the sales positions of strong lobbies, especially the agricultural one, in Europe and the United States, […] It is a difficult policy, much more difficult – I repeat – than to grant aid, something which maybe no one – I’m thinking of the taxpayers – realizes; but it is [by] opening markets, much more than [by] granting
262
Iørn Korzen
aid, that we can offer hundreds of millions of people real and lasting opportunities to free themselves from hunger and poverty […]’³³ The only other gerund clause ClC in the corpus introduces new information: (83)
È dando individualmente prova di coraggio e non di codardia che si può lottare contro razzismo e xenofobia! (ep-96-05-09.txt:35) ‘It is [by] giving individual proof of courage and not cowardice that we can fight against racism and xenophobia!’
Table 11: Distribution of the cleft constituents in the Type 3 Cleft sentences Italian ClCs
Danish ClCs
48
6
anaphoric
42
3
non anaphoric
6
3
2
9
anaphoric
2
8
non anaphoric
0
1
2
–
anaphoric
1
–
non anaphoric
1
–
Prepositional phrases
Adverbs
Gerunds
Type 3 total
52
48.6%
15
8.1%
anaphoric
45
42.1%
11
5.9%
non anaphoric
7
6.5%
4
2.2%
Of the 52 Italian Type 3 CSs, 48 have a prepositional phrase as ClC, two an anaphoric location adverb, qui ‘here’ and lì ‘there’, and two a gerund clause ClC, one of which is used anaphorically. Of the 48 prepositional phrases, 42 are used anaphorically, and out of these 42 anaphoric ClCs, 28 consist of the three units that can all be defined as lexicalised connectives: è per questo che…, è per ciò che…, è per questo motivo che… 33 Given the contrast, one might argue that this example should have been analysed as a Type 1 CS. However, a decisive point is the syntactic function of the ClCs, adverbial adjuncts, and the semantic content of manner in the ClCs.
Cleft sentences. Italian-Danish in contrast
263
The Danish corpus contains nine adverbial ClC, of which eight are used anaphorically: derfor ‘therefore’ (4 cases), her ‘here’ (3), sådan ‘this way’ (1), and one used deictically: her og nu ‘here and now’. In addition to this, there are six Danish prepositional phrases, of which three have an anaphoric complement and three a generally known complement.
5.2.4 The Type 4 Cleft sentence The Type 4 CS could be viewed as a subtype of the Type 3 CS, considering that the ClC is again an adverbial adjunct expressing a particular setting for the subclause, which conveys new information. It is, however, a very special kind of CS, and it is often mentioned particularly by scholars in the field; as stated in Section 3.1, it is termed “metatextual” by Roggia (2009: 146) and De Cesare (2011: 281). The setting expressed by the ClC of this CS type can be defined as “emotional”, and the subclause typically contains a verbum dicendi or a verb with a similar communicative content. The CS often occurs at the beginning – or not long after the beginning – of an oral speech, and there is a strict correlation between the ClC and the subclause. The ClC so to say “prepares” the addressee(s) for the following text and the ClC paradigm is strongly limited by the content of this text. Two examples: (84)
Signora Presidente, signor Primo ministro, onorevoli parlamentari, è con grande piacere che mi presento oggi dinanzi a voi, all’inizio di una nuova presidenza, […] (ep-01-01-17.txt:151) ‘Madam President, Mr. Prime Minister, ladies and gentlemen, it is with great pleasure that I present myself before you today, at the beginning of a new presidency […]’
(85)
Signor Presidente, vorrei rivolgerle le più sentite congratulazioni per la sua elezione alla Presidenza del Parlamento europeo. Essa corona il percorso politico di un europeista leale e convinto, ed è quindi con grande entusiasmo che mi appresto ad instaurare con lei un rapporto di collaborazione che mi auguro pieno e proficuo; […] (ep-02-01-15.txt:36) ‘Mr President, I would like to extend to you the most heartfelt congratulations on your election as President of the European Parliament. It crowns the political career of a loyal and convinced Europeanist, so it is with great enthusiasm that I establish a relationship with you that I hope will be close and fruitful; […]’
264
Iørn Korzen
Regarding this CS type, the statistical count is very simple: The Italian EUROPARLcorpus contains four cases and the Danish corpus none. I leave it to my readers to decide whether this is enough to back a claim about emotional differences between Italians and Danes.
6 Cleft sentence occurrences in another corpus of comparable texts Before the final conclusions and a complete statistical overview, I shall briefly refer to another comparable Italian-Danish corpus, the so-called MR. BEANcorpus (Skytte et al. 1999; http://blog.cbs.dk/mrbean-korpus/), and look into the CS occurrences found there. This (relatively small) corpus consists of narrative texts, more precisely of a number of retellings of two Mr. Bean episodes produced by 27 Italian and 18 Danish university students in 1995. Each student produced a written and an oral retelling, one of each episode; the oral one was recorded on tape and later transcribed. The corpus therefore consists of 54 Italian and 36 Danish Bean retellings, half of them written and half of them oral, and apart from these texts, each student was asked to produce a brief written argumentative text on the basis of four claims suggested by the organisers regarding the topic “smoking”. For detailed accounts of the creation of the corpus, see Skytte et al. (1999) and Korzen (2007a). Table 12 shows the size of each subcorpus and the number and types of CSs encountered. It is interesting that the CSs found in the Italian Bean retellings are exactly the same number and types in the oral and the written texts, even though the former are more than twice the size of the latter. In the Italian argumentative texts, we find the typically argumentative ClC per questo, supplied with the focalizing adverb proprio. In Korzen (2007a: 211), I discussed the level of formality of the Italian and Danish Bean retellings. Generally, the register of the Italian texts, oral as well as written, is considerably higher than that of the Danish texts. In both cases, the creation of the corpus took place in a university, Torino and Copenhagen Universities, respectively, but there is no doubt that the university scene involves a higher formality level in Italy than in Denmark. Furthermore, many of the Danish students collaborated with university teachers they already knew and were familiar with, while the Italian students met a team of researchers not only unknown to them, but consisting of foreign academics. Not wanting to make a bad impression seemingly led the Italians to adopt a relatively high register, which can explain e.g. the presence of a CS with an implicit subclause in reversed order in an oral text, cf. (86).
Cleft sentences. Italian-Danish in contrast
265
Table 12: CS occurrences in the MR. BEAN-corpus
Bean, oral texts Corpus size: CS occurrences total: CS types:
Bean, written texts Corpus size: CS occurrences total: CS types:
Argumentative texts Corpus size: CS occurrences total: CS types:
(86)
Italian
Danish
8,659 words 2 (~ 1 per 4,330 words) Both occurrences: Type 1, subject ClCs, implicit subclauses (see Section 3.2.1), one of which in reversed order, i.e. the subclause preceding the matrix, cited in (86) below³⁴
9,077 words 9 (~ 1 per 1,009 words) Type 1: 7 occurrences (5 subject ClCs, 1 object, 1 prepositional complement) Type 2: 1 occurrence (prepositional complement) Type 3: 1 occurrence (adverbial: space) (not included: the interrogative type cited in (8) and (25): 6 occurrences)
4,050 words 2 (~ 1 per 2,025 words) Both occurrences: Type 1, subject ClCs, implicit subclauses, one of which in reversed order
4,592 words 6 (~ 1 per 765 words) Type 1: 4, all subject ClCs Type 3: 2 (adverbials: space and manner)
1,798 words 1 (~ 1 per 1798 words) Type 1, CS matrix with focalizing adverb: È proprio per questo
1,320 words 2 (~ 1 per 660 words) Type 1: 2 (1 subject ClC, 1 prepositional complement) (not included: the interrogative type cited in (8) and (25): 1 occurrence)
a questo punto ehm, si introduce invece un’altra figura, un pupazzo di un poliziotto, e a introdurla, è-, quello che, potrebbe essere, un commesso nel negozio che-, guarda- in malo modo l’uomo che si è divertito a giocare-, con il presepe (Skytte et al. 1999: 587 / IMA8)³⁵ ‘at this point ehm another figure is introduced, a policeman puppet, and to introduce it [introducing it] is what could be a shop assistant who looks with disfavour at the man who had fun playing with the nativity scene.’
34 As stated in Section 3.2.1, no such cases were found in the EUROPARL-corpus. 35 In the transcription of the oral texts, commas indicate short breaks and hyphens the prolongation of a vowel, generally as a sign of hesitation.
266
Iørn Korzen
In this case, the implicit subclause a introdurla ‘to introduce it’ also functions as a resumptive anaphor of the preceding sentence with a verb that is lexically identical to that of the antecedent (si introduce). Anaphors lexically identical to their antecedents are particularly typical of oral texts (cf. e.g. Korzen 2007a: 220; 2007b: 102–106), a phenomenon that may – at least in part – explain the presence of the CS cited in (86) in an oral text. The results of Table 12 will be compared with the Europarl results in the following section.
7 Conclusion The CS investigation carried out in this study has brought forward a number of inter- and intra-linguistic differences regarding CS usage. The differences were found by employing a typological distinction between four CS types that differ with respect to the syntactic function of the ClC and the information-structural features of ClC and subclause – and thus with respect to the pragmatic and textual function of the CS. The classification was carried out on 107 Italian and 185 Danish CSs found in the EUROPARL-corpus and on 5 Italian and 17 Danish CSs found in the MR. BEAN-corpus, excluding the structures cited in Section 5, Table 5 (which amounted to a total of 12 Italian and 18 Danish Europarl cases and 7 Danish Bean cases). Since the CS occurrences in the MR. BEAN-corpus were so few, especially in the Italian texts, conclusions on that basis should be taken with a grain of salt. However, the Italian CS frequency proved about twice as high in the written texts as in the oral ones, confirming the findings of Roggia (2006) cited in Section 3.1, even though the corpus of Roggia appears to be of a lower formality level than the BEAN-corpus. The CS frequency in the Italian EUROPARL-corpus, cited in Section 1, Table 1, was close to that found in the written retellings and argumentative texts of the Italian MR. BEAN-corpus, amounting to approximately one CS per 1,800–2,100 words. In the Danish corpora, the CS frequency was between three and four times higher than in the Italian corpora, again highest in the written texts where it amounted to one Danish CS per 630–760 written words against approximately one CS per 1,000 oral words. Since Danish has a much more rigid constituent order than Italian, cf. Section 1 and footnote 1, the cross-linguistic differences would argue in favour of the compensation mechanism principle. It would seem that the relatively frequent CS usage in Danish compensates for the more rigid constituent order. The compensation hypothesis was first launched by Jespersen (1937), later followed
Cleft sentences. Italian-Danish in contrast
267
by Lambrecht (2001),³⁶ but strongly criticised by Dufter (2009) on the basis of a cross-linguistic analysis of the EUROPARL-corpus³⁷. It should be noted, however, that Dufter’s comparisons were performed on a parallel EUROPARL-corpus, i.e. on English L1 texts and their translations into German, French, Italian, Spanish and Portuguese (but not Danish); cf. the brief discussion on parallel corpora in Section 1 and footnote 2. All the CS types investigated in this paper are syntactically similar (with the slight differences described in Sections 2.2 and 3.2), and consequently their fundamental characteristics and typical text functions are related. The distinctions between the four CS types should therefore rather be seen as continua regarding features and functions than as four individual and distinct phenomena. Type 1 is the “classic” focus marking and identifying CS with a ClC associated with a paradigm of alternatives, but even within this type there are gradual differences in focus marking strength between structures with and without explicit focalizing elements, cf. Section 5.2.1.1–7. Some of the Type 1 subtypes signal a text topic shift, i.e. a transition to a new text topic, cf. Sections 5.2.1.1 and 5.2.1.6. The Type 1 ClC is typically, but not necessarily, a valency constituent designating a first-order entity and it may be anaphoric, i.e. the discourse referent it designates may have been mentioned previously in the text, but this is not a distinctive feature of this type. On the other hand, the anaphoric function is a distinctive feature of the Type 2 ClC, which is a valency constituent anaphorising a “weak” antecedent. This type may still be considered an identifying CS, but it is generally weaker in its focus marking than the different Type 1 subtypes. Instead, it plays an important text cohesive function in that the anaphoric ClC expresses a textual topic shift and the subclause conveys new information about the new topic. In the sense of potential alternatives, the focus marking of Types 3 and 4 is very weak, but also these two CS types have specific and important text functions. Both have adverbial ClCs expressing various kinds of background setting for the following text, Type 3 generally a scene-setting, which can be e.g. spatiotemporal or argumentative, Type 4 an emotional background. Type 3 is mainly anaphoric and it thereby functions as a text cohesion device encapsulating a preceding text segment and expressing the transition to a new one. Type 4 typi-
36 “The occurrence of cleft constructions in a language correlates with the degree of positional freedom of prosodic accents and syntactic constituents in that language.” (Lambrecht 2001: 488). 37 “At least for clefted subjects […] it seems safe to assume that a significant portion of cleft occurrences cannot be interpreted as a compensatory strategy employed for want of better alternatives. Our findings therefore shed considerable doubt on the compensatory device theory of clefting.” (Dufter 2009: 95).
268
Iørn Korzen
cally occurs text-initially, or at least text sequence-initially, the ClC indicating the speaker’s emotional state in relation to the specific communicative situation. The subclauses of Types 2, 3 and 4 normally convey new information, for which reason they have a high communicative dynamism in the sense of the Prague School. On the other hand, the Type 1 subclauses generally contain known, given or inferable information and are therefore low on communicative dynamism and may be omitted in certain co-texts. In all cases, the information conveyed by the subclause is formed as a logical presupposition, i.e. as a fact expressed as given and indisputable. Figure 1 sums up the mentioned characteristic features of the four CS types. As the figure shows, apart from their various degrees of focus and identification marking, all four CS types play important roles for the surrounding co-text.
Type 1 CS: Focus marking
Type 2
Type 3
Type 4
strong / varying weak
CS: Identifying function
very weak
identifying
non-identifying
ClC: Syntactic function
(mainly) valency valency
adverbial adjunct
ClC: Link to pre-text
(possibly anaphoric)
strong anaphor (mainly) anaphoric
non-anaphoric
CS: Topic and text continuity
possible topic shift
topic shift
textual transition
text initiation
Subclause: Communicative dynamism
low
high
high
high
Figure 1: The Cleft sentence typology suggested in this paper
The distribution of the four CS types in the EUROPARL-corpus was shown in Tables 9 to 11 and is repeated in Table 13 below. The distribution of all Europarl CSs according to the syntactic function of the ClC was shown in Section 5, Table 4. Of the 68 Italian adverbial ClCs cited there, 12 were categorised as Type 1 CSs due to focalizing elements, the rest as Types 3 or 4. Of the 17 Danish adverbial ClCs only two were labelled Type 1 CSs, the rest Type 3. The cross-linguistic differences are evident: in the Italian texts, the most frequent type is Type 3, i.e. the text cohesion marker that links text segments together, expresses the relation between them and indicates the transition from one to another. Considering the importance of well-formed texts in the Italian school system and society, the famous “bello stile” and “bella forma”, and the fact that in the EUROPARL-corpus we are dealing with the argumentative text
Cleft sentences. Italian-Danish in contrast
269
type, it is not surprising that so many Italian CSs are linguistic units with a specific role regarding textual organisation and cohesion. Table 13: Distribution of Cleft sentence Types 1 to 4 in the EUROPARL-corpus Italian CSs
Danish CSs
Type 1
33
30.8%
95
51.4%
Type 2
18
16.8%
75
40.5%
Type 3
52
48.6%
15
8.1%
Type 4
4
3.7%
0
0
Total
107
100%
185
100%
In the Danish texts, more than nine CSs out of ten belong to Types 1 or 2, i.e. the two types that generally focalize valency arguments (non-anaphoric and anaphoric), which typically, but not exclusively, designate first-order entities. As Table 4 showed, in a large majority of these cases, 113 to be precise, the ClC is subject³⁸. Danish Type 3 CSs account for only 8.1% of the cases, which is probably due, in part at least, to the fact that Danish is a “V2 language”, i.e. a language that has the finite verb in the second position in main declarative sentences (for references see footnote 1). The initial position, also termed the “fundamental field” (cf. e.g. Diderichsen [1946] 1971: 185, 190; Hansen and Heltoft 2011: 331 passim), may hold practically all kinds of constituents (except the finite verb) including e.g. (anaphoric and non-anaphoric) adverbial adjuncts, i.e. constituents that may express anaphoric relations and, possibly, transitions from one text segment to another³⁹, as in the following example: (87)
Kyoto-målene stiller os over for en væsentlig udfordring, og bæredygtighed bør være et gennemgående princip for udviklingen af den fremtidige energipolitik. A-kraft er ikke løsningen. Herved kan man måske nedsætte CO2-udslippet, men samtidig skaber denne energiform langt flere
38 A preponderance of subject ClCs was found also by Berretta (1994: 93) and Roggia (2009: 74) in their surveys on Italian CSs in mixed corpora. As table 13 shows, the Italian EUROPARL-corpus differs in that regard from the corpora investigated by Berretta and Roggia. 39 For a comparison between the French and Danish use of sentence initial adverbial adjuncts, which showed them to be particularly frequent in Danish, see Lundquist (2009).
270
Iørn Korzen
nye miljøproblemer. Af samme grund støtter jeg ændringsforslag 8 om at vægte forsigtighedsprincippet. (ep-99-11-16.txt:66) ‘The Kyoto targets present us with a major challenge, and sustainability should be a guiding principle for the development of future energy policy. Nuclear power is not the solution. Hereby [i.e. by using nuclear power] it may be possible to reduce CO2 emissions, but at the same time this form of energy creates far more new environmental problems. For this reason, I support Amendment 8 giving priority to the precautionary principle.’ This linguistic manoeuvre (which exists in Italian, too) gives a certain weight to the initial adverbial constituent, albeit less forcefully so than the CS construction. The situation of the subject is a different one. Since a subject placed in the fundamental field is informationally unmarked in Danish (Hansen and Heltoft 2011: 1713), a focus marking of the subject would require more “potent” linguistic means as for instance a CS structure. On the other hand, in a pro-drop language such as Italian, the mere explication of a subject will often entail a certain focus on this constituent. This may help to explain the differences in frequency of subject ClCs in Italian and Danish CSs. However, also earlier findings regarding related as well as different linguistic phenomena (as e.g. the use of implicit vs. explicit verb forms and different anaphoric structures) point to a high awareness of textual composition and unity in Italian vs. a higher attention towards individual text units in Danish, and within each text unit a particular Danish focus on specific individuals responsible for concrete activities and events (cf. Korzen 2005a/b, 2010, In press; Korzen and Gylling 2012). Thus, in addition to the different word order circumstances in the two languages, the CS distribution in the EUROPARL-corpus investigated seems to corroborate text structuring tendencies and differences encountered also in other studies.
References Bach, Svend & Jørgen Schmitt Jensen. 1990. Større Italiensk Grammatik. [Comprehensive Italian Grammar]. Copenhagen: Munksgaard. Bache, Carl & Niels Davidsen-Nielsen. 1997. Mastering English. An Advanced Grammar for non-native and Native Speakers. Berlin & New York: Mouton de Gruyter. Baroni, Marco & Silvia Bernardini. 2006. A new approach to the study of translationese: Machine-learning the difference between original and translated text. Literary and Linguistic Computing 21 (3). 259–274.
Cleft sentences. Italian-Danish in contrast
271
Bazzanella, Carla. 1994. Le facce del parlare. Un approccio pragmatico all’italiano parlato. Milano: La Nuova Italia. Benincà, Paola. 1978. Sono tre ore che ti aspetto. Rivista di grammatica generativa 3. 231–245. Berretta, Monica. 1994. Ordini marcati dei costituenti di frase in italiano. La frase scissa. Vox Romanica 53. 79–105. Berretta, Monica. 2002. Quello che voglio dire è che: le scisse da strutture topicalizzanti a connettivi testuali. In Gian Luigi Beccaria & Carla Marello (eds.), Le parole al testo. Scritti per Bice Mortara Garavelli, 15–31. Torino: Edizioni dell’Orso. Bonomi, Ilaria. 2002. L’italiano giornalistico. Dall’inizio del ‘900 ai quotidiani on line. Firenze: Franco Cesati. Bossong, Georg. 1991. Differential Object Marking in Romance and Beyond. In Dieter Wanner & Douglas A. Kibbee, New Analyses in Romance Linguistics. Selected Papers from the XVIII Linguistic Symposium on Romance Languages Urbana-Champaign, 143–170. Amsterdam & Philadelphia: John Benjamins. Byskov, J. 1910. Dansk Sproglære [Danish Grammar]. København: Det Schønbergske Forlag. D’Achille, Paolo, Domenico Proietti & Andrea Viviani. 2005. La frase scissa in italiano: aspetti e problemi. In Iørn Korzen & Paolo D’Achille (eds.), Tipologia linguistica e società. Due giornate italo-danesi di studi linguistici (Roma, 27–28 novembre 2003), 249–279. Firenze: Franco Cesati. Dardano, Maurizio. 2005 [1996]. Manualetto di linguistica italiana. 10th edn. Bologna: Zanichelli. Dardano, Maurizio & Pietro Trifone. 2001 [1997]. La nuova grammatica della lingua italiana. Bologna: Zanichelli. De Cesare, Anna-Maria. 2001. Fra teoria e pratica: sintassi, semantica e traduzioni inglesi dell’avverbio proprio. Studi italiani di linguistica teorica e applicata 30 (1). 143–169. De Cesare, Anna-Maria. 2011. L’ordine dei costituenti in italiano contemporaneo e in prospettiva contrastiva con il tedesco. Tra sintassi, pragmatica e tipologia linguistica. Tesi di abilitazione. University of Basle. De Cesare, Anna-Maria. 2012. Riflessioni sulla diffusione delle costruzioni scisse nell’italiano giornalistico odierno a partire dalla loro manifestazione nei lanci di agenzia in italiano e in inglese. Cuadernos de Filología Italiana 19. 11–39. De Cesare, Anna-Maria & Davide Garassino. In press. On the status of exhaustiveness in cleft sentences: A data-driven and cross-linguistic study of English also- / only-clefts and Italian anche- / solo-clefts. Folia linguistica. De Cesare, Anna-Maria. This vol. Cleft constructions in a contrastive perspective. Towards an operational taxonomy. Delin, Judy. 1995. Presupposition and shared knowledge in it-clefts. Language and Cognitive Processes 10 (2). 97–120. De Stefani, Elwys. 2009. Le strutture grammaticali come epifenomeni dell’interazione sociale? Riflessioni sull’uso delle costruzioni scisse nel parlato conversazionale italiano e francese. In Angela Ferrari (ed.), Atti del X Congresso SILFI. Sintassi storica e sincronica, subordinazione, coordinazione, giustapposizione (Napoli, 5–7 ottobre 2010), 1615–1631. Firenze: Franco Cesati. Diderichsen, Paul. 1971 [1946]. Elementær dansk grammatik [Elementary Danish Grammar]. 3rd edn. København: Gyldendal. Dik, Simon C. 1989. The Theory of Functional Grammar. Part I: The Structure of the Clause. Dordrecht-Providence: Foris.
272
Iørn Korzen
Dufter, Andreas. 2009. Clefting and discourse organization. Comparing Germanic and Romance. In Andreas Dufter & Daniel Jacob (eds.), Focus and background in Romance languages, 83–121. Amsterdam & Philadelphia: John Benjamins. Dyhr, Mogens. 1978. Die Satzspaltung im Dänischen und Deutschen. Eine kontrastive Analyse. Tübingen: TBL Verlag Gunter Narr. Firbas, Jan. 1966. On defining the Theme in Functional Sentence Analysis. Travaux Linguistiques de Prague (1). 267–280. Fornaciari, Raffaello. 1974 [1881]. Sintassi italiana dell’uso moderno. Firenze: Sansoni. Frison, Lorenza. 1982. Alcune differenze tra l’inglese e l’italiano nel comportamento della frase scissa. Rivista di Grammatica Generativa 7. 79–121. Frison, Lorenza. 2001. Le frasi scisse. In Lorenzo Renzi, Giampaolo Salvi & Anna Cardinaletti (eds.), Grande grammatica italiana di consultazione, vol. I, 208–239. Bologna: il Mulino. Garassino, Davide. This vol. Cleft sentences. Italian-English in contrast. Gettrup, Harald, Michael Herslund, John Pedersen & Arne Schnack. 1986. Sprog og tekst. Studiebog i sagtekstanalyse [Language and text. Textbook on the analysis of non-fiction texts]. København: Romansk Institut. Københavns Universitet. Givón, Talmy. 1976. Topic, Pronoun and Grammatical Agreement. In Charles N. Li (ed.), Subject and Topic, 149–188. New York: Academic Press. Götzsche, Hans. 1998. Some remarks on the analysis of clefting in Danish. Papers from the 16th Scandinavian Conference of Linguistics (Turku / Åbo, November 14–16, 1996), 128–140. Publications of the Department of Finnish and General Linguistics of the University of Turku. Graffi, Giorgio. 1978. On Italian cleft sentences. In Maria-Elisabeth Conte, Anna Giacalone Ramat & Paolo Ramat (eds.), Wortstellung und Bedeutung. Akten des 12. Linguistischen Kolloquiums (Pavia, 1977), 113–123. Tübingen: Max Niemeyer. Grewendorf, Günther & Cecilia Poletto. 1989. La costruzione scissa: un’analisi contrastiva. Rivista di Grammatica Generativa 14. 105–142. Guil, Pura. 1994. Es que… in italiano. In Anna Giacalone Ramat & Massimo Vedovelli (eds.), Italiano lingua seconda / lingua straniera, Atti del XXVI Congresso Internazionale di Studi della SLI (Siena, 5–7 novembre 1992), 111–126. Roma: Bulzoni. Halliday, Michael A.K. 1967. Notes on transitivity and theme in English. Part 2. Journal of Linguistics 3. 199–244. Hansen, Aage. 1933. Sætningen og dens led i moderne dansk [The sentence and its constituents in modern Danish]. København: Nyt Nordisk Forlag, Arnold Busck. Hansen, Erik. 1995. Sætningskløvning i moderne dansk [Cleft sentences in modern Danish]. In Iver Kjær, Flemming Lundgreen-Nielsen & Merete K. Jørgensen (eds.), Danske studier 1995 [Danish studies 1995], 126–146. København: C.A. Reitzel. Hansen, Erik & Lars Heltoft. 2011. Grammatik over det Danske Sprog, vols. I–III [Grammar of the Danish Language]. København: Det Danske Sprog-og Litteraturselskab. Heltoft, Lars. 1992. The Topology of Verb Second and SVO Languages. In Michael Herslund (ed.), Word Order. Two Studies on Central Issues in the Syntax of Danish and French. Copenhagen Studies in Language 15. 13–64. Herslund, Michael (ed.). 1996. Det franske sprog. Kapitel III. Valens og transitivitet [The French language. Chapter III. Valency and transitivity]. København: Handelshøjskolen i København (Copenhagen Business School). Herslund, Michael. 2003. Sur la modalité en danois et la tripartition de la phrase. Revue Belge de Philologie et d’Histoire 81 (3). 867–882.
Cleft sentences. Italian-Danish in contrast
273
Herslund, Michael. 2005. Clivage, structure thématique et anaphores. In Frédéric Lambert & Henning Nølke (eds.), La syntaxe au cœur de la grammaire. Recueil offert en hommage pour le 60e anniversaire de Claude Muller, 127–135. Rennes: Presses Universitaires de Rennes. Herslund, Michael. 2006. La topologie du français à la lumière de deux autres synstèmes. Le danois et l’ancien francais. In Kim Gerdes & Claude Muller (eds.), Lingvisticæ Investigationes XXIX(1). Fascicule special. Ordre des mots et topologie de la phrase francaise, 103–112. Amsterdam & Philadelphia: John Benjamins. Herslund, Michael. In press. Réflexions sur la topologie de la phrase française. In Hans Petter Helland & Christine Meklenborg Salvesen (eds.), Festschrift for Marianne Hobæk Haff. Oslo: Novus forlag. Hopper, Paul J. & Sandra A. Thompson. 1980. Transitivity in Grammar and Discourse. Language 56 (2). 251–299. Jespersen, Otto. 1969 [1937]. Analytic Syntax. New York: Holt, Rinehart & Winston. Koehn, Philipp. 2005. Europarl: A parallel corpus for statistical machine translation. Conference Proceedings: The Tenth Machine Translation Summit, 79–86. Phuket, Thailand. Korzen, Iørn. 1996. L’articolo italiano fra concetto ed entità, vols. I–II. Etudes Romanes 36. Copenaghen: Museum Tusculanum Press. Korzen, Iørn. 2000. Reference og andre sproglige relationer. In Gunver Skytte & Iørn Korzen, Italiensk-dansk sprogbrug i komparativt perspektiv. Reference, konnexion og diskursmarkering [Italian-Danish language usage in a comparative perspective. Reference, connexion and discourse marking], 161–619. København: Samfundslitteratur. Korzen, Iørn. 2001. Anafore e relazioni anaforiche. Un approccio pragmatico-cognitivo. Lingua nostra LXII (3–4). 107–126. Korzen, Iørn. 2005a. Struttura linguistica e schema cognitivo: tipologie a confronto. In Iørn Korzen (ed.), Lingua, cultura e intercultura: l’italiano e le altre lingue. Atti dell’VIII Convegno SILFI (Copenaghen, 22–26 Giugno 2004). Copenhagen Studies in Language 31. 123–134. Korzen, Iørn. 2005b. Lingue endocentriche e lingue esocentriche: lessico, testo e pensiero. In Iørn Korzen & Paolo D’Achille (eds.), Tipologia linguistica e società. Due giornate italo-danesi di studi linguistici (Roma, 27–28 novembre 2003). 31–54. Firenze: Franco Cesati. Korzen, Iørn. 2006. On Demonstrative Determiners in Anaphoric Noun Phrases. In Henning Nølke, Irene Baron, Hanne Korzen, Iørn Korzen & Henrik Høeg Müller (eds.), Grammatica. Festschrift in Honour of Michael Herslund, 261–277. Bern: Peter Lang. Korzen, Iørn. 2007a. Mr. Bean e la linguistica testuale comparativa. Considerazioni tipologicocomparative sulle lingue romanze e germaniche. In Manuel Barbera, Elisa Corino & Cristina Onesti (eds.), Corpora e linguistica in rete, 209–224. Perugia: Guerra Edizioni. Korzen, Iørn. 2007b. Linguistic typology, text structure and anaphors. In Iørn Korzen & Lita Lundquist (eds), Comparing Anaphors. Between Sentences, Texts and Languages. Copenhagen Studies in Language 34, 93–109. København: Samfundslitteratur Press. Korzen, Iørn. 2008. Determination in endocentric and exocentric languages. With evidence primarily from Danish and Italian. In Henrik Høeg Müller & Alex Klinge (eds.), Essays on Nominal Determination. From Morphology to Discourse Management, 79–99. Amsterdam & Philadelphia: John Benjamins.
274
Iørn Korzen
Korzen, Iørn. 2009. Struttura testuale e anafora evolutiva: tipologia romanza e tipologia germanica. In Iørn Korzen & Cristina Lavinio (eds.), Lingue, culture e testi istituzionali, 33–60. Firenze: Franco Cesati. Korzen, Iørn. 2010. Lingua, cognizione e due Costituzioni. In Jacqueline Visconti (ed.), Lingua e diritto. Livelli di analisi, 163–202. Milano: Edizioni Universitarie di Lettere Economia Diritto. Korzen, Iørn. In press. Frasi complesse e complessità frasale: il discorso politico in un’ottica tipologico-comparativa. In Carla Bruno, Simone Casini, Francesca Gallina & Raymond Siebetcheu (eds.), Plurilinguismo/Sintassi. Atti del XLVI Congresso Internazionale SLI (Siena 27–299 settembre 2012), Roma: Bulzoni. Korzen, Iørn & Morten Gylling. 2012. Text Structure in a Contrastive and Translational Perspective. On Information Density and Clause Linkage in Italian and Danish. Translation: Corpora, Computation, Cognition 2 (1). 23–46. Krifka, Manfred. 2007. Basic notions of information structure. In Caroline Fery, Gisbert Fanselow & Manfred Krifka (eds.), The Notions of Information Structure. (Interdisciplinary Studies on Information Structure 6). 13–55. Potsdam: Universitätsverlag Potsdam. Lambrecht, Knud. 2001. A framework for the analysis of cleft constructions. Linguistics 39 (3). 463–516. Lundquist, Lita. 2009. Adverbiaux initiaux en danois et en français: langue, texte, mentalité. In Iørn Korzen & Cristina Lavinio (eds.), Lingue, culture e testi istituzionali, 141–162. Firenze: Franco Cesati. Lyons, John. 1977. Semantics, vols. 1–2. Cambridge: Cambridge University Press. McEnery, Anthony M., Richard Z. Xiao & Yukio Tono. 2006. Corpus-based Language Studies: An Advanced Resource Book. Routledge Applied Linguistics Series. London: Routledge. Metzeltin, Michele. 1989. La scissione relativa in italiano e nelle altre lingue romanze. In Fabio Foresti, Elena Rizzi & Paola Benedini (eds.), L’italiano tra le lingue romanze. Atti del XX congresso internazionale di studi. Società di Linguistica Italiana, 151–169. Roma: Bulzoni. Mikkelsen, Kristian. 1975 [1911]. Dansk ordföjningslære med sproghistoriske tillæg [Danish syntax with diachronic observations]. København: Hans Reitzels Forlag. Nølke, Henning. 1983. Les adverbs paradigmatisants: Fonction et analyse. Revue Romane numéro spécial 23. København: Akademisk Forlag. Nølke, Henning. 1984. Clefting in Danish? In Finn Sørensen & Lars Heltoft (eds.), Topics in Danish Syntax. Nydanske studier & almen kommunikation, NyS [Studies in modern Danish and general communication], 14. 72–111. Panunzi, Alessandro. 2009. Strutture scisse e pseudoscisse: valori d’uso del verbo essere e articolazione dell’informazione nell’italiano parlato. In Angela Ferrari (ed.), Atti del X Congresso SILFI. Sintassi storica e sincronica, subordinazione, coordinazione, giustapposizione (Basilea, 30 giugno–3 luglio 2008), vol. II. 1121–1137. Firenze: Franco Cesati. Prince, Ellen F. 1978. A Comparison of WH-Clefts and IT-Clefts in Discourse. Language 54 (4). 883–906. Rehling, Erik. 1932. Det danske Sprog. Fremstilling for Lærere og Seminarier [The Danish language. A presentation for teachers and colleges of education]. København: J.H. Schultz. Roggia, Carlo Enrico. 2006. Costruzioni marcate tra scritto e parlato: la frase scissa. In Angela Ferrari, Parole frasi testi tra scritto e parlato. Cenobio 3. 222–230. Roggia, Carlo Enrico. 2009. Le frasi scisse in italiano. Struttura informativa e funzioni discorsive. Genève: Slatkine. Rooth, Mats. 1992. A theory of focus interpretation. Natural Language Semantics 1. 75–116.
Cleft sentences. Italian-Danish in contrast
275
Sabatini, Francesco. 1985. “L’italiano dell’uso medio”: una realtà tra le varietà linguistiche italiane. In Günter Holtus & Edgar Radtke (eds.), Gesprochenes Italienisch in Geschichte und Gegenwart, 154–184. Tübingen: Narr. Serianni, Luca. 1991 [1989]. Grammatica italiana. Italiano comune e lingua letteraria, 2nd edn. Torino: UTET. Skytte, Gunver, Iørn Korzen, Paola Polito & Erling Strudsholm (eds.). 1999. Tekststrukturering på italiensk og dansk. Resultater af en komparativ undersøgelse / Strutturazione testuale in italiano e danese. Risultati di una indagine comparativa. Copenaghen: Museum Tusculanum Press. Sornicola, Rosanna. 1991. Origine e diffusione della frase scissa nelle lingue romanze. In Dieter Kremer (ed.), Actes du XVIIIe Congrès International de Linguistique et de Philologie Romanes, vol. 3, 41–54. Tübingen: Niemeyer. Togeby, Ole. 2009. Ordkløveri? Om sætningskløvning på dansk [Hair-splitting? On sentence clefting in Danish]. In Rita Therkelsen & Eva Skafte Jensen (eds.), Dramatikken i grammatikken, 439–455. Roskilde: Institut for Kultur og Identitet, Roskilde Universitet. von Heusinger, Klaus. 1999. Intonation and information structure. Universität Konstanz. Habilitationsschrift. [Chapter 4: Semantic theories of focus] http://www.ilg.uni-stuttgart.de/ vonHeusinger/publications/abstracts/habil.html (accessed 1 November 2012). Wiwel, Hylling Georg. 1901. Synspunkter for dansk sproglære [Viewpoints on Danish grammar]. København: Det Nordiske Forlag. Zemach, Eddy M. 1970. Four Ontologies. The Journal of Philosophy LXVII (8). 231–247.
Giovanna Brianti*
Cleft sentences. A translation perspective on Italian and French 1 Introduction The aim of this paper is to provide a contrastive analysis on the use of Cleft sentences in French and Italian, based on a corpus of translated newspaper articles. I will analyse in particular the so-called explicit clefts of the type C’est moi qui l’ai dit / Sono io che l’ho detto ‘It is me who said it’ and the Italian implicit cleft (Sono stato io a dirlo). As we know (Roggia 2008, 2009), French and Italian show significant differences in the way they express focalization. In that regard, French provides an ideal ground to investigate the “compensation mechanism”, stated in the following terms by Lambrecht (2001: 488): “The occurrence of cleft constructions in a language correlates with the degree of positional freedom of prosodic accents and syntactic constituents in that language”. In French, word order and the position of prosodic accents are determined in a rigid way, which means that the frequency of Cleft sentences is predictably higher than in other languages, like Italian, offering a much greater flexibility in the ordering of constituents as well as prosodic accents. Corpus-based studies, like the introductory essay by De Cesare et al. (in this volume), surprisingly show a high incidence of Cleft sentences not only in French but also in Italian, which seems to indicate that the use of clefts cannot be explained in structural terms alone. From a textual point of view, Cleft sentences fulfil other functions, such as marking transitions, especially when they open a new paragraph, enhancing cohesion (Dufter 2009: 115) or achieving a disambiguating purpose (Carter-Thomas 2009: 148). This is particularly true for press articles that belong to an argumentative genre. The focus on translation provides a favourable standpoint in order to assess not only the frequency of Cleft sentences in French and Italian, but also to grasp better the reasons underlying the choice of a Cleft sentence among other focalizing devices. It is important, however, to be aware of the possible weight of interference that may affect a translated text in order to avoid a biased interpretation
* I would like to thank Anna-Maria De Cesare for her insightful comments and constructive suggestions throughout the elaboration of this chapter, as well as Jörn Albrecht for his careful reading of a previous version of this text. Giovanna Brianti, University of Geneva
278
Giovanna Brianti
of the data extracted from translated corpora. In other words, it is not possible to bring together the contribution of contrastive linguistics and translation studies in the realm of Cleft sentences without being aware of the methodological issues related to the compilation of corpora, on one hand, and the specific features of translated documents as opposed to original texts, on the other. This paper is organized as follows. In section 2, I will highlight the major differences between a text originally written in a given language and a translated text in terms of the “translation universals” and I will introduce the concept of “translationese” that is used to analyse the phenomenon of interference between the original and the translation. I will also define translation (or parallel) corpora and comparable corpora with their advantages and drawbacks for a contrastive analysis. In section 3, I will describe the translation corpora compiled for this study: Le Monde Diplomatique and its Italian translation published by Il Manifesto, as well as Press Europ, a multilingual electronic collection of press articles on European Union affairs, out of which I compiled two small subcorpora made of articles translated from French into Italian and from Italian into French. Section 4 will be devoted to the quantitative and qualitative analysis of data extracted from these corpora. I will focus on the distribution of explicit and implicit clefts and I will examine more carefully the alternatives to clefts in translation, which are more frequent in the Press Europ corpus as the translation is less literal than in the Monde Diplomatique-Manifesto corpus. In Section 5, I will carry out a general comparison between the data extracted from my translation corpora and the ICOCP-corpus,¹ which is a comparable corpus (see De Cesare et al. in this volume). The results show that there are some significant differences in the use of Cleft sentences in texts written in the original language and translated texts. This means that translation corpora and comparable corpora will not provide the same results and therefore are not interchangeable in the realm of contrastive linguistics and translation studies.
2 Contrastive linguistics and translation With the development of corpus-based approaches, contrastive linguistics, on the one hand, has shown a growing interest in translation (e.g. Laviosa 2002; Granger 2003; Olohan 2004). Indeed, corpora made of translated texts ideally provide large amounts of empirical data allowing the comparison of two or more languages from the lexical, syntactic, semantic or pragmatic point of view. On the
1 ICOCP-corpus: Italian Constituent Order in a Contrastive Perspective-corpus.
Cleft sentences. A translation perspective on Italian and French
279
other hand, translation studies make use of these corpora in order to highlight the specific features of translated texts. Whereas this rapprochement between contrastive linguistics and translation studies is certainly promising, we have to be aware of two possible risks of deviation. First of all, studies in contrastive linguistics do not necessarily take into account the fact that translation is often the result of a choice among several options. Moreover, as Toury’s (1995) “law of interference” shows, the translator tends to resort to an equivalent structure in the target language when it is available (see Section 4). The risk is then to draw biased conclusions from the analysis of translation corpora. As Olohan (2004: 27–28) indicates, studies in contrastive linguistics may implicitly “assume that choices in translation are determined to a large extent by the language system”, thus “fail[ing] to recognize that translators’ choices may be motivated by something other than language systemic conventions”. Secondly, translation studies tend to exacerbate the alteration in the translation process to the point of considering the language of translation as a separate linguistic subsystem. In what follows, I will first explore the issue of interference in translation before turning to methodological issues concerning translation corpora and comparable corpora.
2.1 Differences between an original text and a translated text: Translation universals and “translationese” As Garzone (2004: 35) points out, since the 1980s translation studies have undergone a “paradigm shift” by moving from comparing the original and the translated text in terms of equivalence to focusing on the translated text and its specific functions in the target linguistic and cultural setting. With the extension of corpus-based studies that allow for a more precise analysis of the translation effects on the target language it has been possible to explore the various phenomena of interference with the source language (on lexical, morpho-syntactic and pragmatic levels), which will eventually contribute to the evolution of the target language. Some recent studies (Garzone and Cardinaletti 2004; Cardinaletti and Garzone 2005) assume that the translated text is a means of spreading syntactic innovations as a result of interference with the source language. This phenomenon becomes even more relevant given the increasing amount of “covert transla-
280
Giovanna Brianti
tions” (House 1997)² that are circulated in all fields (medical, journalistic, scientific, etc.) and received as original texts given the fact that they are not specifically identified as a translation. Interference as such can be seen from two perspectives: a prescriptive one and a descriptive one. In the first case, the translator has to be aware of the negative effects of interference from the source language, also called “translationese”, that is “deviance in translated texts induced by the source language” (Johansson and Hofland 1994: 26). According to Toury (1980: 75), “theoretical considerations […] lead to hypothesising that the language used in translation tends to be interlanguage (sometimes designated ‘translationese’), or that a translation is, as it were, an ‘inter-text’ by definition”. In a more positive light, the language of translation has been defined by Frawley (1984) as a “third code”. In other words, the translation language is the result of a compromise between the structures of the source language and the target language and, in that sense, unique (Baker 1998: 482). In the same direction, Toury (1995) identifies laws of translational behaviour, among which he includes the “law of interference”.³ Corpus-based studies have been instrumental in defining the characteristics of translation as a “third code”. This descriptivist approach allows to differentiate the language of translated texts from the one employed in the documents drafted in the original language. These differences have been first formalised by Mona Baker (1993: 243–245) under the name of “universals of translation”. The universals of translation have been variably listed in terms of: Explicitation, Simplification, Disambiguation, Conventionalisation, Standardisation, etc. The translation universals have been repeatedly questioned, mainly because the categories they include are too general and not always representative of the translation process.⁴ In the same mind-set, I would question the very existence of a “third code” as a different form of language born out of translation. As we know, especially syntactic structures are rather resistant to the influence of other languages, with the exception of some limited sequences, such as multiple interrogations (Who bought what? / Chi ha comprato che cosa?).⁵ From a linguistic point 2 “A covert translation [as opposed to an overt translation] is a translation which enjoys the status of an original source text in the target culture. The translation is covert because it is not marked pragmatically as a translation text of a source text but may, conceivably, have been created in its own right” (House 1997: 69). 3 “In translation, phenomena pertaining to the make-up of the source text tend to be transferred to the target text” (Toury 1995: 275). 4 For a deeper discussion on this matter, see House (2008) and Chesterman (2010). 5 According to Benincà (1993: 288), “the cultural contact does not transfer complex and productive structures, but at most it ‘excites’ pre-existing structures and rules: otherwise, it can only transfer rigid and limited schemes” (author’s translation).
Cleft sentences. A translation perspective on Italian and French
281
of view, it is however interesting to explore the language contact phenomena in order to identify possible sources of syntactic influence due to translation.⁶ The awareness of possible cases of interference in translated texts has important methodological consequences for choosing a reliable corpus adapted to a contrastive analysis. In the following section, I will define and compare two main types of corpora: translation corpora (made of translated texts) and comparable corpora (made of texts written in the original language).
2.2 Translation corpora and comparable corpora As has been repeatedly noted (e.g. Laviosa 2002: 36–37; Granger 2003: 19; Olohan 2004: 24), we are still lacking a unified terminology for the different types of corpora used in contrastive linguistics as well as translation studies. For the purpose of this study – and adopting the most common terminology – I will distinguish between “translation” or “parallel corpora”⁷ and “comparable corpora”.⁸ “A parallel corpus consists of original, source language-texts in language A and their translated versions in language B” (Baker 1995: 230). “Parallel corpora can be used to provide information on language-pair specific translational behaviour, to posit certain equivalence relationships between lexical items or structures in source and target languages […] or to study the phenomenon of translationese […]”(Kenny 1998: 51). A comparable corpus is made of original texts in two or more languages that have been selected on the basis of their similarity and comparability, in terms of their genre, content, form, date of appearance etc. (Laviosa 2002: 36–37).⁹ 6 McLaughlin (2011: 23) makes an interesting distinction between syntactic influence and borrowing applied to the translation of press articles from English into French: “it is appropriate to refer to the impact that the presence of the original has on the syntax of the translations as influence rather than borrowing […]. An important consequence of the distinction between influence and borrowing is that any evidence of influence in the corpus will not necessarily indicate that borrowing has taken place”. 7 In the field of contrastive linguistics, the most common term is “translation corpora”, whereas the translation studies tend to prefer “parallel corpora” (Granger 2003: 20). In this study, I will use the term “translation corpora”. 8 Depending on the authors, comparable corpora can either refer to a collection of comparable original texts in two or more languages or to a collection of texts originally written in a language, alongside a collection of texts translated (from one or more languages) into that language (Baker 1995: 234). For clarity sake and for the purpose of this research, we will only retain the first definition of comparable corpora. 9 See Johansson (2007: ch. 1) for a more thorough discussion on comparable and parallel corpora.
282
Giovanna Brianti
Both types of corpora present advantages and drawbacks for contrastive analyses.¹⁰ In our case, data drawn from translation corpora may lead to very interesting observations on language specific behaviours regarding the frequency, position, form and function of Cleft sentences in press articles. It might be misleading, however, to draw definitive conclusions on the use of Cleft sentences in French and Italian from a translation corpus, given the inevitable weight of interference due to the translation process. Depending on the scope of analysis, both contrastive linguistics and translation studies will rely upon translation corpora and/ or comparable corpora. In translation studies, parallel corpora are often used to highlight the typical traces left by the translation process on the target text, the so-called “translationese”, or to illustrate the “translation universals”. Comparable corpora are preferred in broader analyses aiming at emphasizing the specific features of each of the compared languages. In contrastive linguistics, scholars rely more typically upon separate monolingual corpora which have the advantage of characterizing more precisely the spontaneous uses of a specific structure in a given language and of being more readily available in most languages. However, a growing number of studies in contrastive linguistics make use of translation corpora in order to highlight similarities and/or contrasts between two or more language systems. In the realm of Cleft sentences, we can quote the comprehensive multilingual study by Andreas Dufter (2009) based on EUROPARL, a multilingual translation corpus of European Parliament Proceedings.¹¹ The purpose of that study is to compare the “translational counterparts of English it-Clefts in German, French, Italian, Spanish and Portuguese” (Dufter 2009: 83). The same corpus has been used by Gast and Wiechmann (2012) for their study on wh-clefts in German and English. Indeed, EUROPARL offers a user-friendly platform allowing for immediate cross-linguistic comparisons. As Olohan (2004: 25) points out, however, there are two problems concerning these types of translation corpora: First of all, “it can be difficult to assign the status of ‘source text’ to one of the language versions; documents may be written in more than one language and, once translations exist, there is nothing to distinguish a source text from the other language versions”.
10 See Wecksteen (2007: 261–266) on the use of translation corpora in contrastive linguistics and translation studies. 11 The latest version of the EUROPARL parallel corpus (v7) contains documents issued in 21 languages by the European Parliament from 1996 until 2011 (http://statmt.org/europarl/ [accessed 29 July 2014]).
Cleft sentences. A translation perspective on Italian and French
283
Secondly, in order to cover the 552 language combinations offered by the 24 official languages of the European Union in a cost-effective way, the translation department resorts to a complex system of “relay or pivot languages” (mainly English, French or German).¹² In the European Parliament, however, “all parliamentary documents are published in all the official languages of the European Union and all the Members of the European Parliament have the right to speak in the official language of their choice”. With the enlargement of the European Union, the European Parliament developed the concept of ‘controlled full multilingualism’ by resorting to ‘relay languages’ in translation (Wagner, Bech & Martinez 2012).¹³ In sum, even though the most recent version of EUROPARL mentions the original language of around 80% of the speeches, we cannot exclude that some translations have been made through a ‘relay language’, especially if the source language is not one of the main European Union languages (Korzen & Gylling 2012: 29–30). If translation corpora are used mainly to provide “a reflection of the possibilities offered by the target language system” (Olohan 2004: 24), we can understand that the translation process is not taken into account. Comparable corpora, on the other hand, allow exploring the use of a given construction in two or more languages without the deforming perspective of translation. According to Sinclair (1996), “the possibilities of a comparable corpus are to compare different languages or varieties in similar circumstances of communication, but avoiding the inevitable distortion introduced by the translations of a parallel corpus”.
12 As Dr. Valter Mavric̆, from the DG TRAD (Direction de la Traduction), pointed out to me in an e-mail message on 18 October 2012, “according to our most recent statistics, the three relay languages account for 70% of the originals. An additional 10% would be covered if the original pivot was extended to three additional languages, Spanish, Italian and Polish. Thus the remaining 17 languages combined would contribute only about 10% of the originals”. The term “original language” clearly refers to the source language that serves as a basis for translation. 13 As the European Parliament’s website says about multilingualism: “the European Parliament has put in place a system of ‘relay’ languages: A text is first translated into one of the most widely used languages (English, French or German) and from there into the minor languages. Other major Community languages (Italian, Polish and Spanish) could also become relay languages in due course”. (http://www.europarl.europa.eu/aboutparliament/en/007e69770f/Multilingualism.html [accessed 29 July 2014]). “When the document contains several languages that are not covered in language units, a relay language is chosen (EN, FR, DE). The choice of a relay language depends on the capacity in these three language units, on the proportion of the relay language already contained in the source text (e.g. if a document is composed of 60% EN, 30% FR, 7% DE and 3% Latvian, English will be selected as the relay language) and on the current ability of the relay language to cover all the source languages contained in the document” (cf. the intranet page of the Demand Management Service-Planning).
284
Giovanna Brianti
All in all, the choice of a translation corpus or a comparable corpus will be defined according to the objectives of the study: if the goal is to identify specific translation features it will be preferable to compile a translation corpus; if it is more generally a contrastive analysis, two options are available, either a comparable corpus or the comparison between a translation and a comparable corpus,¹⁴ which is the option I chose in the present study. Indeed, the combination of a translation corpus and a comparable corpus allows establishing a more solid basis for the contrastive hypotheses. This position is supported and applied by a number of scholars, like Tognini-Bonelli (2001: 6), who affirms: “When working across languages it is advisable to consider the evidence of both translation corpora and comparable corpora because they have different things to offer”, or Granger (2003: 19), who suggests that “the two types of corpus should be used concurrently as each has its advantages and disadvantages”.¹⁵ The main disadvantage of a comparable corpus “lies in the difficulty of establishing comparability of texts” (Granger 2003: 19), especially when there is a large cultural difference between the languages that are compared. In the case of journalistic prose, this type of difficulty is less likely to arise given the fact that the texts belong to the same genre and timeframe. In the following sections, I will compare the data taken from two translation corpora of newspaper articles in French and Italian that I compiled manually with the data from the ICOCP-corpus (for Italian and French), which perfectly falls into the definition of a comparable corpus. This process will allow me to distinguish better the characteristic behaviour of Cleft sentences in French and Italian from the cases in which the use of Cleft sentences in the target language is clearly influenced by the source language from a quantitative and qualitative point of view.
2.3 Translation corpora of newspaper articles: Methodological issues Before comparing the use of Cleft sentences in original and translated texts, a few remarks will be necessary in order to highlight the methodological issues underlying the building of a translation corpus of journalistic texts.
14 “Translation corpora are insufficient as sources of contrastive studies, and need to be combined with comparable corpora, i.e. collections of original texts in the languages compared” (Johansson 2007: 5). 15 See also Laviosa (2002: 38–39); Olohan (2004: 37); Ondelli and Viale (2010: 576).
Cleft sentences. A translation perspective on Italian and French
285
First of all, existing parallel multilingual corpora including French and Italian are – to my knowledge – exclusively related to the European Union institutions.¹⁶ A corpus of journalistic texts has then to be compiled manually.¹⁷ Secondly, some parts of newspaper articles can be the result of a translation that has not been stated as such. For instance, most of the press releases contain translated quotations without indicating the source language; signed articles as well can contain fair portions of “covert translations”, especially in the articles composed on the basis of press releases made out of portions of texts or quotes in several languages. Thirdly, when talking about translation in the journalistic field, it is necessary to distinguish between a strictly speaking translation corpus, made of the unabridged translation of single newspaper articles, and a “free-translation corpus” (Tognini-Bonelli 2001: 7), composed of articles in the source language that have been freely translated and adapted in the target language.¹⁸ For obvious reasons, it is not possible to implement a contrastive linguistic analysis on the basis of a “free-translation corpus”.
2.3.1 French and Italian newspaper corpora Concerning translated press articles in French and Italian, we could distinguish three types of corpora: a. unabridged translation of a single source, such as the Italian translation of Le Monde Diplomatique, which offers a “faithful” translation of a large selection of articles taken from every monthly issue of the French journal; b. “free-translation” of a selection of articles such as Le Courrier International¹⁹ for French and Internazionale for Italian. In those weekly newspapers, the translations are rarely integral and most often provide a summary of the
16 Cf. EUROPARL and the European Union Lingua project, started in 1993 at the University of Nancy, which includes original and translated texts in 10 languages (English, French, German, Italian, Spanish, Portuguese, Danish, Finnish, Swedish and Greek) (Laviosa 2002: 103–105). 17 See for example Ondelli and Viale’s (2010) study based on a parallel corpus of translated newspaper articles from English into Italian. 18 As Bielsa and Bassnett (2009: 57) point out, the task of translation in news agencies is intimately connected with selecting, editing, completing or reducing a press article. 19 Le Courrier International contains articles from various sources in several languages and is translated into French by a team of professional translators (Lavault-Olléon and Sauron 2009).
286
c.
Giovanna Brianti
main information contained in the original version of a press article. Moreover they are very often adapted to the style of the target audience;²⁰ mixed translation strategy: in some cases the translation is unabridged and in others the translator will resort to the “free-translation” style. This approach is used by Press Europ. This multilingual website offers a number of unabridged translations of original articles, as well as a summary of relevant news coming from different sources.
In addition to these sources that exclusively provide translated press articles, it is possible to find a number of translations in the main national daily papers (Corriere della Sera, Le Monde etc.). In the next section, I will present the two corpora that I compiled manually from two sources: Le Monde Diplomatique and its Italian version published by Il Manifesto, and a set of articles drawn from Press Europ that are divided into two subcorpora: articles translated from French into Italian and articles translated from Italian into French.
3 Presentation of the corpora compiled for the study of Cleft sentences in French and Italian 3.1 Le Monde Diplomatique-Il Manifesto Le Monde Diplomatique,²¹ a monthly journal specialized in political analysis and commentary, has a worldwide audience thanks to its 47 international editions in 28 languages (including 39 printed and 8 electronic versions). The Italian translation of Le Monde Diplomatique,²² published since 1994 as a monthly supplement of the left-wing daily newspaper Il Manifesto,²³ offers a
20 “It is obviously when translating press articles that the translator becomes most like a journalist. As the literary translator necessarily has in himself a latent writer, a general press or specialized translator takes pleasure in adopting the quality of a journalistic style to the public for whom he writes. It is up to him to recreate a captivating title and headline, to explain an acronym or an original cultural fact, in sum to adapt his text to his reader, as the journalist did it when he wrote the original text, in a properly target-oriented process” (Lavault-Olléon and Sauron 2009) (author’s translation). 21 http://www.monde-diplomatique.fr/. 22 Le Monde Diplomatique-Il Manifesto: http://www.monde-diplomatique.it/. 23 For a thorough presentation of both journals, see Pederzoli (2006).
Cleft sentences. A translation perspective on Italian and French
287
reliable and consistent basis for the constitution of a French-Italian translation corpus. Indeed, most of the articles of Le Monde Diplomatique are fully translated into Italian and available online without charge 45 days after the original French publication, which makes them easily accessible. Moreover, the translation is as faithful as possible to the source text.²⁴ A personal communication with the staff of Il Manifesto allowed me to clarify that the translations are not necessarily made by professional translators, which could explain the word for word translation style. The revision of the articles (as well as the creation of titles, subtitles and abstracts) is then made by the editorial staff.²⁵ As this monthly supplement of Il Manifesto is the only Italian journal mostly composed of translations from French, it is often used for contrastive linguistic analyses.²⁶ The corpus that I assembled manually includes six issues of the journal from December 2011 until May 2012. As we will see in the analysis of translated examples from Le Monde Diplomatique, the fidelity to the original often results in a word for word translation of the whole text in general and of the Cleft sentences in particular, which will require some caution in the interpretation of the quantitative results as they will not necessarily correspond to the frequency of clefts in a corpus of press articles originally written in Italian.
3.2 Press Europ Born out of a more recent European initiative, Press Europ, which was launched by the European Commission and a consortium led by the Courrier International in May 2009, is the first multilingual website compiling press articles on European Union affairs.²⁷ The free Press Europ website publishes articles from the main European daily newspapers translated into 10 languages. We can find two types of articles: the unabridged original version of articles selected from a number of national sources, fully translated into 10 languages and signed by the
24 As stated by Pederzoli (2006), “On the basis of the agreement with the French editor, the Italian edition is bound to offer a translation that is as faithful as possible to the source text. The name of the translator is indicated for every article, which is a sign of a certain transparency and seriousness in terms of fidelity to the original” (author’s translation). 25 E-mail message from Geraldina Colotti, Il Manifesto, 13.10.2012. 26 For example, the study by Vecchiato (2005) includes a short section on Cleft sentences based on a corpus of articles drawn from Le Monde Diplomatique and Il Manifesto (see 4 for a more detailed presentation of that study). 27 http://europa.eu/rapid/pressReleasesAction.do?reference=MEMO/09/248&format=HTML&a ged=0&language=EN&guiLanguage=en (accessed 16 July 2012). Unfortunately, the Press Europ site has stopped publishing new articles on-line on December 20, 2013 due to lack of funding.
288
Giovanna Brianti
translator; a compilation of several sources around a theme or news item made by the editorial staff of Press Europ and translated into 10 languages. I excluded from my corpus the articles compiled from a collection of original texts as they correspond to the “free-translation” style, which is rather an adaptation than a translation. I included in my corpus articles originally written in French and published in French newspapers, mainly Le Monde,²⁸ together with their Italian translation. The original version of the article is available on the site, which allows for an easy inter-lingual comparison. I also included a more limited number of Italian articles translated into French, which will allow me to check how clefts are translated from Italian into French.²⁹ As advertised on the website, the articles are translated by professional translators and then revised by journalists. I carefully selected the original sources as being written in French or Italian, in order to avoid the risk of examining translations of source texts that are already the result of translation into one of the European Union “relay language”, as is the practice in the European institutions. The bidirectional Press Europ corpus that I have compiled manually spans over January to August 2012 and is composed of two subcorpora: articles originally written in French and translated into Italian; articles originally written in Italian and translated into French. The size of this corpus is much smaller than the Monde Diplomatique-Manifesto corpus inasmuch as the number of French original articles is very limited compared to Le Monde Diplomatique. In the direction Italian-French, the number of Italian original articles is even scarcer. Despite these quantitative restrictions, it will nevertheless be possible to observe some relevant differences between our two corpora in translation strategies concerning Cleft sentences. Table 1 and Table 2 present both corpora: Table 1: Monde Diplomatique-Manifesto corpus: Overview Abbreviation Dates Monde Diplomatique
MD
Manifesto (Italian translation) MAN
Total words
December 2011-May 2012 (6 issues) 217,635 Idem
213,999
28 Other sources are Libération, Le Figaro, La Tribune, Les Echos, Télérama, France Inter, as well as Le Temps (Geneva) and Les Coulisses de Bruxelles. 29 The Italian sources are La Stampa, La Repubblica, Il Corriere della Sera, Il Sole-24 Ore, Il fatto quotidiano and Linkiesta.
Cleft sentences. A translation perspective on Italian and French
289
Table 2: Press Europ corpus: Overview
Press Europ (French original)
Abbreviation
Dates
PE-FR
January– 32,283 August 2012
Press Europ (Italian original)
idem
Press Europ PE-IT (Italian translation)
Total words
30,656
Abbreviation
Total words
PE-IT
19,313
Press Europ PE-FR (French translation)
19,198
4 Cleft sentences in translation In this section, I will turn to the analysis of data. I will focus my comparative analysis on two types of clefts, namely explicit clefts and implicit clefts. Explicit clefts are found in French and Italian and correspond to the general definition of Cleft sentences (C’est Marie qui a téléphoné / È Maria che ha telefonato ‘It is Mary who called’): the matrix clause is headed by a copula and the verb of the subordinate clause is conjugated.³⁰ The implicit cleft, which is restricted to Italian, is formed with an implicit subordinate clause introduced by the preposition a followed by an infinitive and occurs exclusively with subjects (È stata Maria a telefonare ‘It was Mary who called’;³¹ Frison 1988: 201–206; Berretta 1994: 85). I have excluded from this study the following subtypes of Cleft sentences as they are not present in my corpora: pseudo-conditional Cleft sentences and Cleft sentences with verbs different than être / essere ‘to be’ (see De Cesare et al. in this volume). I included in the quantitative data a small number of Cleft sentences with gerundive subordinate clauses and found one occurrence of inferential cleft that I will quote in 4.2.3. Within the rich contrastive bibliography on Cleft sentences (D’Achille, Proietti and Viviani 2005), there are a few recent studies devoted to the comparison between French and Italian (Schöpp 2005; Roggia 2008; De Stefani 2009), but,
30 “A cleft construction (CC) is a complex sentence structure consisting of a matrix clause headed by a copula and a relative or relative-like clause whose relativized argument is coindexed with the predicative argument of the copula. Taken together, the matrix and the relative express a logically simple proposition, which can also be expressed in the form of a single clause without a change in truth conditions” (Lambrecht 2001: 467). 31 My statistics on implicit clefts include a very small number of the so-called implicit Reverse clefts of the type (A dirlo sono stato io, lit. ‘To say it was I’) (1 occurrence in the MD-MAN-corpus and 3 occurrences in the PE-IT/PE-FR corpus).
290
Giovanna Brianti
to my knowledge, only one briefly addressing the issue of translation between these two languages based on a corpus of translated texts.³² In the volume entitled L’italiano delle traduzioni, edited by Anna Cardinaletti and Giuliana Garzone (2005), the essay by Sara Vecchiato analyzes, among other phenomena of syntactic interference between French and Italian, the translation of Cleft sentences from French into Italian. Her comparison of a few translated articles from Le Monde Diplomatique³³ and a recent mystery novel allows her to highlight the fact that the translators of Il Manifesto typically reproduce the French Cleft sentence (or replace it sometimes with an implicit cleft), whereas the translator of the novel tends to avoid the use of Cleft sentences in Italian. This study interestingly reveals that the ratio of clefts in translation can vary according to the translator. The comparison between my two translation corpora will also allow me to stress some interesting quantitative differences in both translations.
4.1 Monde Diplomatique-Manifesto corpus: Quantitative and qualitative results In this section I will first examine the distribution of explicit and implicit Cleft sentences (4.1.1); I will then analyse the distribution of Cleft sentences according to the grammatical category of the cleft constituent (4.1.2) and I will finally consider the cases where a French cleft is not translated with a cleft in Italian (4.1.3).
4.1.1 Distribution of explicit and implicit Cleft sentences Table 3 shows the total amount of clefts in French and Italian.
32 Other studies based on translation corpora involving French and/or Italian are Ondelli and Viale (2010) English-Italian; Carter-Thomas (2009) French-English; Muller (2003) FrenchGerman; as well as Dufter (2009), which compares English and its translational counterparts in German, French, Italian, Spanish and Portuguese. 33 Vecchiato’s study is limited to 24 articles of Le Monde Diplomatique (year 2003) and their Italian translation.
Cleft sentences. A translation perspective on Italian and French
291
Table 3: Monde Diplomatique-Manifesto (MD-MAN-corpus): Quantitative results (total number of examples and % of total clefts) Occurrences
Explicit clefts
Implicit clefts³⁴
Monde Diplomatique 104
104
–
48
Manifesto
70 (67.3%)
17 17 (16.35%) (16.35%)
41
104
French clefts translated with other constructions
Per 100,000 words
In this table we notice that the French Cleft sentences are mainly translated with clefts in Italian (83.7% of total number of examples), which shows that the translators mainly chose the closest available structure in Italian. A more striking difference is given by the small percentage of implicit clefts in the Italian translation of Le Monde Diplomatique (16.35% of total number of examples). The implicit Cleft sentence is possible only when the focalized element is the subject and is preferred to the explicit cleft in order to avoid the duplication of the conjugated verb in the sentence (Frison 1988: 202). According to Berretta (1994) and Roggia (2009), it is the most frequent form of Cleft sentences in written Italian,³⁵ whereas oral Italian tends to prefer the explicit cleft.³⁶ In their study on journalistic language, Bonomi et al. (2002: 328) observe that implicit clefts are more common than explicit clefts, because that type of construction is more adapted to a formal style³⁷. As the data I extracted from the MD-MAN-corpus contradict the results of a number of studies on the use of clefts in written Italian in general and journalistic language in particular (see De Cesare et al. in this volume), in this case we see a clear effect of interference between the source language, which does not have an implicit cleft, on the target language. It is interesting to notice that in most cases the implicit cleft is chosen in the translation when in the French text the copula
34 This column includes one example of implicit Reverse cleft (c’est un sentiment de gâchis qui prédomine / a predominare è un sentimento di confusione ‘it is a sentiment of waste that prevails’ [MD-MAN 12.11]). 35 Gil (2004: 379–380) explains the higher frequency of implicit clefts by their higher degree of grammaticalization. 36 Berretta (1994: 92) found 84.3% of implicit clefts in her corpus made of newspaper articles and oral conversations, whereas Roggia’s data (2009: 71) on written and oral Italian confirm that when the subject is focalized, written Italian prefers the implicit cleft (57%), whereas spoken Italian chooses the explicit cleft (79%). 37 See also Garassino’s study on Italian and English Cleft sentences based on the ICOCP-corpus (this volume).
292
Giovanna Brianti
is plural and the verb of the subordinate clause is in a past tense (c’étaient ‘it was’). It seems that the translator tends to prefer the implicit cleft in those cases where an agreement is needed between the copula and the verb of the subordinate clause, namely in the cases where the standard present form c’est … que is not to be found, as in examples (1) and (2):³⁸ (1)
a.
“Carolis n’était pas conscient du fait que, certes, le CSA l’avait nommé, mais que c’étaient les différentes tutelles [notamment le ministère des finances et le ministère de la culture] qui comptaient,” juge aujourd’hui M. Philippe Baudillon, ex-patron de France 2, ancien diplomate proche de M. de Villepin. (MD 04.12)³⁹
b. “Carolis non era cosciente del fatto che, certo, il Csa lo aveva nominato, ma che a contare erano le varie tutele [in particolare il ministero delle finanze e il ministero della cultura],” analizza oggi Philippe Baudillon, ex direttore generale di France 2 ed ex diplomatico vicino a de Villepin. (MAN 04.12) ‘“Carolis was not conscious of the fact, of course, that the CSA had named him, but that it was the different regulating bodies [notably the Ministry of Finance and the Ministry of Culture] that counted,” affirms Philippe Baudillon today, ex director of France 2, former diplomat close to de Villepin.’ c’étaient les différentes tutelles ‘it was the different regulating bodies a contare ‘to count (2)
a.
qui comptaient that counted’
erano le varie tutele were the different regulating bodies’
Ce sont les oligarques qui ont servi d’entremetteurs, négociant la composition d’un gouvernement “proeuropéen” autour des bonnes tables du Club des entrepreneurs de Belgrade. (MD 03.12)
38 “It is worth noting that in the explicit form, essere ‘to be’ remains preferably in the present tense even when the other verb is in the past or future […], whereas in the implicit form essere has to match the tense of the other verb” (Berretta 1994: 92) (author’s translation). 39 The source of the examples is abbreviated in the following way: MD (Monde Diplomatique, French edition; MAN (Monde Diplomatique-Manifesto, Italian edition), MDE (Monde Diplomatique, English edition). The examples will be followed by full translations in English and, in some cases, also by glosses.
Cleft sentences. A translation perspective on Italian and French
293
b. Sono stati gli oligarchi a fare da mediatori contrattando la composizione di un governo “pro europeo” nei buoni ristoranti del Club degli imprenditori di Belgrado. (MAN 03.12) ‘When no clear majority emerged in the spring 2008 Serbian elections, the oligarchs acted as go-betweens in negotiating the composition of a pro-European government over dinners in a Belgrade business club.’ (MDE 03.12) Ce sont les oligarques ‘It is the oligarchs Sono stati ‘It was
qui ont servi d’entremetteurs who served as go-betweens’
gli oligarchi a fare da mediatori the oligarchs to do as go-betweens’
4.1.2 Distribution of grammatical categories in Cleft sentences Table 4 presents the distribution of cleft constituents according to their grammatical category. The second column shows that the cleft constituents are mainly subjects and adverbials in French. In the Italian translation, the category of cleft subjects appears mainly in explicit clefts (3rd column), but also in implicit clefts (4th column). As in French, adverbials are well represented in Cleft sentences. The 5th column shows the relatively few cases in which a French cleft is translated with another syntactical construction in Italian.⁴⁰ This table makes evident the predictable fact that clefts tend to put a constituent presenting a given information in the focal position: a subject (very often a pronoun) or an adverbial. According to Berretta (2002:19), “the Cleft sentence has a linear structure that is more inclined to insert given focal elements represented by light constituents like pronouns or adverbials”.⁴¹ In example (3) we have a cleft subject (ce type de discours / questo genere di discorsi ‘this type of rhetoric’), in example (4) a cleft pronoun with the function of a subject (elle ‘it’ / questa comunità ‘this community’), and in example (5) a locative adverbial (là / lì ‘there’):
40 The predominance of adverbials is also found in Korzen’s essay (in this volume) on Italian and Danish Cleft sentences based on the EUROPARL corpus. 41 Author’s translation.
294
Giovanna Brianti
Table 4: Cleft sentences (total number of examples and % of total clefts in MD-MAN-corpus) FR-MD
IT-MAN explicit clefts
IT-MAN implicit clefts
IT-MAN French clefts translated with other constructions
Subjects
51 (49%)
26 (25%)
17 (16.3%)
8 (7.7%)
Objects
4 (3.8%)
3 (2.9%)
–
1 (0.95%)
Adverbials
46 (44.2%)
39 (37.5%)
–
7 (6.7%)
Gerundives
3 (3%)
2 (1.9%)
–
1 (0.95%)
TOTAL GRAND TOTAL: 104
104
70 (67.3%)
17 (16.3%)
17 (16.3%)
(3)
a.
C’est ce type de discours qui a permis à la droite d’échapper au “scénario des temps difficiles”. (MD 01.12)
b. È questo genere di discorsi che ha permesso alla destra di sfuggire dallo “scenario dei tempi difficili”. (MAN 01.12) ‘It was a popular line, and the reason the right fastened on to it is because it allowed them to flip the script of the hard times scenario.’ (MDE 01.12) C’est ce type de discours ‘It is this type of rhetoric d’échapper to escape (4)
a.
qui a permis à la droite that allowed the political right’
au scénario des temps difficiles from the crisis scenario’
“Les élites politiques de tous les pays des Balkans tirent leur légitimité de la ‘communauté internationale’. C’est elle qui attribue les brevets de compatibilité démocratique, de modernité, d’efficacité,” explique M. Zaimi. (MD 03.12)
b. È questa comunità che concede attestati di compatibilità democratica, di modernità, di efficienza – spiega Zaimi. (MAN 03.12) “The international community confers legitimacy on the political elite in all the Balkan countries. It gives a stamp of bureaucratic compatibility, modernity and efficiency,” said Zaimi. (MDE 03.12) C’est elle ‘It is it
qui attribue les brevets de compatibilité démocratique… that attributes certificates of democratic compatibility…’
Cleft sentences. A translation perspective on Italian and French
(5)
a.
295
“Mon père est arrivé de Chine continentale à 18 ans [à la fin des années 1940], ma mère vient de Penghu [archipel des Pescadores, au large de la côte ouest de Taïwan]. C’est là que je suis né. Et maintenant, je suis à Kaohsiung. Je vis à Taïwan”. (MD 02.12)
b. “Mio padre è arrivato dalla Cina continentale a 18 anni [alla fine degli anni quaranta], mia madre viene da Penghu [le isole Pescadores, al largo della costa ovest di Taiwan]. È lì che sono nato. E oggi sono a Kaohsiung. Vivo a Taiwan”. (MAN 02.12) “My father came from mainland China when he was 18 [at the end of the 1940s], my mother is from the Penghu Islands [off the west coast of Taiwan], which is where I was born. And now I’m in Kaohsiung. I live in Taiwan.” (MDE 02.12) C’est là ‘It is there
que je suis né that I was born’
The high incidence of adverbials in the French version of Le Monde Diplomatique can be explained by the frequency of phrases of the type C’est ainsi que ‘It is that way that’, always translated with è così che, which has become a grammaticalized expression, as in example (6):⁴² (6)
a.
C’est ainsi que le 15-M s’est réinventé et a pu poursuivre ses actions en décidant d’arrêter le campement et l’occupation permanente de la Puerta del Sol. (MD 05.12)
b. È così che il 15-M si è reinventato e ha potuto proseguire la sua azione decidendo di smantellare l’accampamento e di porre fine all’occupazione permanente di Puerta del Sol. (MAN 05.12) ‘After 15-M decided to end the permanent occupation of the square on 12 June 2011 it reinvented itself, and expanded to reach a broader, more working-class public by putting down local roots.’ (MDE 05.12) C’est ainsi ‘It is that way
que le 15-M s’est réinventé that the 15-M reinvented itself’
42 In the MD-MAN-corpus we found 4 examples of that kind, which amounts to 9% of cleft adverbials.
296
Giovanna Brianti
The category of objects is relatively infrequent in the cleft position. It can be expressed by a noun phrase as in example (7) (cet invisible / questo invisibile ‘that invisible’) or by a pronoun as in example (8) (elle / lei ‘it’): (7)
a.
La carte ordinaire (le dedans) n’offre aux yeux du lecteur qu’un tout petit morceau du monde, le reste est invisible (le dehors). Et c’est cet invisible que L’Atlas du Monde diplomatique 2012 essaie de rendre visible. (MD 03.12)
b. La mappa ordinaria (il dentro) offre al lettore solo un piccolo pezzo del mondo, il resto è invisibile (il fuori). È questo invisibile che l’Atlas di Le Monde Diplomatique prova a rendere visibile. (MAN 03.12) ‘A normal map (the inside) only shows the reader a tiny part of the world; the rest (the outside) is invisible. And it is this invisible part that the 2012 Atlas tries to make us see.’ (MDE 03.12) Et c’est cet invisible ‘And it is that invisible part
que L’Atlas du Monde diplomatique 2012 that the Atlas of the Monde Diplomatique
essaie de rendre visible seeks to make visible’ (8)
a.
“La voici”. M. Arcenio Osorio pointe du doigt l’immense montagne qui surplombe le village de Santiago Lachiguiri, dans l’Etat mexicain d’Oaxaca, isthme de Tehuantepec. “Elle donne à boire à toutes les villes de la région et pour nous, les Zapotèques, elle est sacrée. C’est elle que nous voulions protéger de manière officielle”, ajoute le secrétaire de l’assemblée communautaire, une structure traditionnelle élue qui organise la population du village. (MD 12.11)
b. “Eccola”. Arcenio Osorio indica l’immensa montagna che sovrasta il villaggio di Santiago Lachiguiri, nello stato messicano di Oaxaca, istmo di Tehuantepec. “Dà da bere a tutte le città della regione e per noi zapotechi è sacra. È lei che avremmo voluto proteggere in modo ufficiale”, aggiunge il segretario dell’assemblea comunitaria, una struttura tradizionale eletta che organizza la popolazione del villaggio. (MAN 12.11) “That’s the one,” said Arcenio Osorio, pointing at the huge mountain that towers over the village of Santiago Lachiguiri, in Oaxaca state, part of southwestern Mexico’s Isthmus of Tehuantepec. “It provides water to all the towns in the area, and to us, the Zapotec people, it’s
Cleft sentences. A translation perspective on Italian and French
297
sacred. That’s the mountain we wanted official protection for.” Osorio is secretary of the community assembly, a traditional elected body that represents the people of the village. (MDE 01.12) C’est elle que nous voulions protéger de manière officielle ‘It is it [the mountain] that we wanted to protect officially’ Another reason for clefting is more typically textual, where the cleft constituent (mainly subject or adverbial) often occurs at the beginning of a new paragraph and marks a transition between what was said before and the new information that follows. According to Dufter (2009: 115), “clefting is but one of several means for fulfilling a variety of functions, which include […] discursive goals such as enhancing cohesion, marking transitions, or differentiating between propositional content that is under discussion and information that is not”. In our corpus we found 30 occurrences of Cleft sentences in paragraph opening position (29% of total examples, marked with #), among which 14 cleft subjects and 16 cleft adverbials. In most cases, the cleft constituent summarizes what was said in the previous paragraph and introduces the new information that follows. In example (9a) the prepositional phrase dans ce contexte ‘in this context’ summarizes the political situation in Syria stated in the first part of the article, whereas the clause les révoltes arabes ‘Arab uprisings’ introduces a new concept: (9)
a.
#C’est dans ce contexte qu’ont éclaté les révoltes arabes. Elles revendiquent la liberté, la dignité (karama), la démocratie et la justice sociale. (MD 04.12)
b. #E’ in questo contesto che sono scoppiate le rivolte arabe. Esse rivendicano la libertà, la dignità (karama), la democrazia e la giustizia sociale. (MAN 04.12) ‘It was in the context of this instability that the Arab uprisings broke out, with demands for freedom, dignity (karama), democracy and social justice.’ (MDE 04.12) C’est dans ce contexte qu’ont éclaté les révoltes arabes ‘It is in this context that broke out Arab uprisings’ The connecting role of the Cleft sentence is often reinforced by the presence of pragmatic or textual connectives. Berretta (1994: 100) calls this type of clefts “textual clefts” and underlines the fact that they are often associated with textual
298
Giovanna Brianti
connectives like ma ‘but’ or e ‘and’.⁴³ By “textual connectives” we refer to a rather wide category of pragmatic connectives that are used to signal the transition between the various parts of the text. In some cases, these connectives may also indicate the semantic relationship between the propositional content of the events referred to in the text (Andorno 2003: 182). Our corpus is particularly rich of these connectives associated with cleft subjects (8 occurrences out of 14) or cleft adverbials (2 occurrences out of 16). The most frequent type of connective is mais ‘but’ that occurs 5 times with cleft subjects (out of 8 examples) and once with an adverbial cleft (out of 2 examples). In example (10a), the value of Mais at the beginning of a new paragraph is not adversative but it reinforces the emphasis put on Jean-Claude Michéa that Mr Le Pen quotes after a number of other philosophers and politicians who are against globalization. The same interpretation can be given to mais in example (11a) as the journalists are introduced as a new category of people who manifest against Putin in Russia: (10)
a.
#Mais c’est encore le philosophe Jean-Claude Michéa qui semble l’avoir le plus impressionnée, à la suite, précise-t-elle, “de conversations, de débats passionnés qui m’ont opposée à certains de mes amis sur des sujets aussi importants que la laïcité, la république, le libreéchange ou la fin de l’euro.” (MD 04.12)
b. #Ma è soprattutto il filosofo Jean-Claude Michéa che sembra averla particolarmente impressionata, a seguito, precisa, “di conversazioni e dibattiti appassionati che mi hanno contrapposta ad alcuni amici su temi importanti come la laicità, la Repubblica, il libero scambio o la fine dell’euro.” (MAN 04.12) However, it is mainly the philosopher Jean-Claude Michéa who seemed to have impressed her most, because, she specifies, “of conversations and heated debates that pitted me against some of my friends on subjects as important as secularism, the Republic, free market, or the end of the Euro.”
43 “Finally, the type that I would call merely textual is the one in which the cleft constituent recalls and/or summarises the previous portion of text […], and the Cleft sentence begins or prepares a further development of the text. In this way the Cleft sentence simply becomes a structure that marks the transition between two parts of a text, and represents a sort of signal enhancing discourse articulation […]. Examples of this kind typically appear at the beginning of new textual units […] and are frequently underlined by the textual connectives e and ma” (author’s translation).
Cleft sentences. A translation perspective on Italian and French
299
Mais c’est encore le philosophe Jean-Claude Michéa ‘However, it is mainly the philosopher Jean-Claude Michéa qui semble l’avoir le plus impressionnée who seemed to have impressed her most’ (11)
a.
#Mais ce sont surtout les journalistes mobilisés qui trouvent dans les manifestations la confirmation d’une réalité qu’ils ont préfabriquée, ignorant la diversité des conditions sociales des individus prenant part aux cortèges. (MD 05.12)
b. #Ma sono soprattutto i giornalisti mobilitati, a trovare nelle manifestazioni la conferma di una realtà che loro stessi hanno prefabbricato, ignorando la diversità di condizione sociale di coloro che prendono parte ai cortei. (MAN 05.12) ‘And the media has had the reality that it constructed confirmed by the protests (provided it ignored the actual diversity in any crowd of demonstrators).’ (MDE 05.12) Mais ce sont surtout les journalistes mobilisés manifestations ‘However, it is mainly the journalists involved demonstrations
qui trouvent dans les who found in the
la confirmation d’une réalité qu’ils ont préfabriquée the confirmation of a reality that they have prefabricated’ Ma sono soprattutto i giornalisti mobilitati, manifestazioni… ‘However, are mainly the journalists involved demonstrations…’
a trovare nelle to find in the
In the example (12), mais ‘but’ cannot be interpreted as adversative since the previous part of the article states other reasons for the success of berlusconisme. The introduction of Mais at the beginning of a new paragraph, followed by the keyword berlusconisme in a cleft position, gives emphasis to the fact that the success of that trend is cultural or anthropological. It is interesting to note that the Italian translator wrongly interprets this connective as an adversative by translating it with the adversative conjunction però ‘however’ that he/she postpones after the adverbial (È sul piano culturale però ‘it is on the cultural level however’). In that case Ma ‘But’, placed at the beginning of the sentence would have translated more effectively the emphatic meaning of Mais:
300
(12)
Giovanna Brianti
a.
#Mais c’est sur le plan culturel, voire anthropologique, que le berlusconisme a remporté ses plus éclatantes victoires, en imposant les codes culturels de la “néo-télévision”. (MD 12.11)
b. #È sul piano culturale però, perfino antropologico, che il berlusconismo ha ottenuto le sue più eclatanti vittorie, imponendo i codici culturali della “neo-televisione”. (MAN 12.11) ‘But it is on the cultural or anthropological level that Berlusconism won its most spectacular victories, by imposing the cultural norms of ‘neo-television.’’
4.1.3 Alternatives to clefts in translation In the MD-MAN-corpus the number of cases in which a Cleft sentence is not translated with a cleft in Italian is rather low (17 examples out of 104, about 16%). It would be imprudent to draw definitive conclusions from such a limited number of occurrences, therefore the following observations will merely suggest possible avenues for further exploration. In this section, I will first explore the ways in which the Italian translation renders the information structure without a cleft constituent at the syntactic and discourse level. I will then examine the cases in which there is no interpretative equivalence between the original sentence and its translation. From the quantitative point of view, the translator mainly resorts to an alternative structure when the cleft constituent is an adverbial (11 cases out of 17). In example (13b), the prepositional phrase Nella rete ‘on the Web’ remains in the opening position of the sentence without being inserted into a cleft. In this case, the disambiguating effect of the French cleft is not expressed in the Italian translation:⁴⁴ (13)
a.
Les commentaires célébrant la “révolution Facebook” en Tunisie ne sont d’aucun secours pour comprendre comment les blogueurs tunisiens sont passés de l’enthousiasme au désenchantement. Mani-
44 “The main discourse motivation behind a writer’s decision to employ a c’est-cleft is one of disambiguation. Use of the c’est-cleft enables the writer to confer a certain textual prominence on the clefted item, by imbuing it not only with thematicity but also with focal highlighting. The reader is given more precise processing instructions than in the case of a non-cleft equivalent with regard to the desired interpretation and ambiguity is reduced to a minimum” (CarterThomas 2009: 148).
Cleft sentences. A translation perspective on Italian and French
301
chaeus résume ainsi leur état d’esprit actuel: “Ils sont en deuil”. C’est sur la Toile qu’ils ont compris que le régime leur était insupportable et qu’ils pouvaient le critiquer publiquement. (MD 02.12) b. I commenti che celebrano la “rivoluzione Facebook” in Tunisia non sono di nessun aiuto per comprendere come i blogger tunisini siano passati dall’entusiasmo al disincanto. Manichaeus riassume così il loro attuale stato d’animo: “Sono in lutto”. Nella rete avevano compreso che il regime era loro insopportabile e che potevano criticarlo pubblicamente. (MAN 02.12) ‘The media celebration of Tunisia’s “Facebook revolution” does not help us understand why the enthusiasm of the bloggers has turned to disillusion. One blogger, Manichaeus, wrote: “They are in mourning.” The internet had made them aware they hated the regime and could publicly criticise it.’ (MDE 02.12) C’est sur la Toile insupportable… ‘It is on the web unbearable…’
qu’ils ont compris que le régime leur était that they understood that the regime was
Nella rete avevano compreso che il regime era loro insopportabile… ‘On the web they understood that the regime was unbearable…’ In the example (14b), the non-hypothetical se ‘if’ used with a causal meaning allows to keep the causal interpretation of the French embedded clause: (14)
a.
A en croire un récent sondage, 46% des électeurs républicains du Mississippi estiment que les mariages inter-raciaux devraient être interdits. “C’est pour une bonne raison que Dieu nous a créés de couleurs différentes, et il convient d’honorer Sa volonté en n’allant pas épouser quelqu’un d’une autre couleur que celle qu’Il a choisie pour nous”, argumente une Mississippienne dans un courriel envoyé à l’institut qui a réalisé ce sondage. (MD 04.12)
b. A voler credere ad un recente sondaggio, il 46% degli elettori repubblicani del Mississippi ritiene che i matrimoni misti dovrebbero essere proibiti. “Se Dio ci ha creati di colori diversi è per un buon motivo, ed è opportuno onorare la sua volontà evitando di andare a sposare qualcuno di un colore differente rispetto a quello che Egli ha scelto
302
Giovanna Brianti
per noi”, spiega una donna del Mississippi in una e-mail inviata all’istituto che ha realizzato il sondaggio. (MAN 04.12) ‘A recent survey revealed that 46% of Mississippi’s Republican voters believe inter-racial marriage should be banned. “I believe God made us a different colour for a reason and should be honoured by not marrying outside of the race that God picked for me,” wrote a woman in an email to the survey institute.’ (MDE 06.12) C’est pour une bonne raison que Dieu nous a créés de couleurs différentes ‘It is for a good reason that God created us different colours’ Se Dio ci ha creati di colori diversi ‘If God created us different colours
è per un buon motivo it is for a good reason’
The passive mode in the French original text seems to trigger declefting, either with adverbials or subjects (5 occurrences out of 17). In example (15), the passive mode is maintained in the Italian translation (il messaggio è destinato al mondo intero ‘the message is addressed to the whole world’), whereas the prepositional phrase al mondo intero ‘to the whole world’ is not inserted into a Cleft sentence: (15)
a.
C’est effectivement au monde entier – la fameuse “communauté internationale” – qu’est destiné le message. Au Nicaragua même, il tombe à plat. (MD 05.12)
b. Infatti il messaggio è destinato al mondo intero – la famosa “comunità internazionale”. Perché proprio in Nicaragua, ha registrato un pesante fallimento. (MAN 05.12) ‘His message was for foreign consumption; it fell flat in Nicaragua.’ (MDE 06.12) C’est effectivement au monde entier ‘It is indeed to the whole world Infatti il messaggio ‘Indeed the message
qu’est destiné le message that the message is addressed’
è destinato al mondo intero is addressed to the whole world’
In example (16a), the cleft subject is translated with a passive form with an agent (da duemila persone ‘of 2,000 people’):
Cleft sentences. A translation perspective on Italian and French
(16)
a.
303
Ce sont finalement quelque deux mille personnes parmi lesquelles des militants historiques des luttes minoritaires, comme M. Charles Barron, ancien membre des Black Panthers, aujourd’hui conseiller municipal de New York qui marchent dans ce quartier. (MD 05.12)
b. Alla fine, il corteo che sfila per il quartiere sarà composto da circa duemila persone – tra le quali militanti storici di gruppi minoritari, come Charles Barron, ex membro dei Black Panthers, oggi consigliere municipale di New York. (MAN 05.12) ‘In all there were 2,000 marchers, including the minority rights activist Charles Barron, a former Black Panther and now a NYC council member.’ (MDE 05.12) Ce sont finalement quelque deux mille personnes ‘It is finally around 2,000 people
qui marchent who march
dans ce quartier in this neighbourhood’ Alla fine, il corteo che sfila per il quartiere sarà composto da circa duemila persone ‘Finally, the procession that marches through the neighbourhood will be made of around 2,000 people’ On the level of discourse, it is worth noting that the translator tends to break a series of clefts by resorting to a different structure in order to avoid repetition. In example (17b) the translator did not insert così ‘thus’ in a Cleft sentence because of the proximity of another Cleft sentence: (17)
a.
Un tel esprit de consensus n’a jamais été la marque de M. Sarkozy. Quand il s’est agi d’imprimer son style, au cours de son mandat, il ne s’est guère encombré de bienséance: c’est ainsi en petit comité et sur les conseils de M. Alain Minc qu’il décida, en janvier 2008, la suppression totale de la publicité sur les chaînes de France Télévisions. En novembre 2010, le Parlement s’opposa toutefois à cette décision, en permettant au groupe public, à travers deux amendements législatifs, de conserver de la publicité avant 20 heures. C’est également sans débat que le chef de l’Etat s’arrogea la prérogative de nommer directement les futurs présidents de l’audiovisuel public, alors que les “sages” du CSA, instance réputée indépendante, en étaient chargés précédemment. (MD 04.12)
304
Giovanna Brianti
b. Un tale riguardo per il consenso non è mai stato nelle corde di Sarkozy. Quando, nel corso del mandato presidenziale, si è trattato di dar prova del suo stile, non si è mai posto problemi di buona educazione: così, nel gennaio 2008, in un comitato ristretto – e seguendo i consigli di Alain Minc – ha deciso la soppressione totale della pubblicità sui canali di France Télévisions […]. È sempre senza dibattito che il capo dello stato si è arrogato la prerogativa di nominare direttamente i futuri presidenti dell’audiovisivo pubblico, togliendo questo compito ai “saggi” del Csa, un’istanza ritenuta indipendente. (MAN 04.12) ‘Such a consensual spirit was never the strength of Sarkozy. When it came to making his mark, when in office, he was never gifted with good manners: it is thus in a small group and on the advice of Alain Minc that he decided, in January 2008, to suppress publicity on the channels of ‘France Télévisions.’ […] It is also without debate that Sarkozy took the liberty to directly name the future presidents of the public audio-visual media, while taking this responsibility away from the ‘wise men’ of CSA, an agency renowned for its independence.’ As it often happens in translation, in some cases it is not possible to establish any interpretative equivalence between the original and the Italian version. The absence of cleft in Italian cannot be explained in contrastive terms, but may result from an inaccurate reading of the information structure of the original. In our corpus, this happens in some cases, where the disambiguating effect of the Cleft sentence is not expressed in the Italian translation. In the example (18a), taken from a local news article reporting oral conversation among village people, the sentence structure is doubly marked: by the left dislocation of the object (le village ‘the village’) and the cleft with a subject (ce sont les enfants qui le font vivre ‘it is the children who make it live’). That sentence is replaced by a standard passive in Italian (Il paese è mantenuto in vita dai bambini ‘The village is kept alive by the children’), which can be interpreted as a case of standardisation as the passive mode is less marked than the combination of the left dislocation and clefting: (18)
a.
Où se rencontrer? Les bistrots ont fermé. Les services publics? Un car scolaire, une mairie (deux permanences par semaine), une école, et c’est tout. Le village, ce sont les enfants qui le font vivre: leurs pas
Cleft sentences. A translation perspective on Italian and French
305
et leurs rires résonnent chaque jour dans les rues, à 8h29, à 11h31, à 13h29 et à 16h31. (MD 04.12) b. Dove incontrarsi? I bistrot hanno chiuso. I servizi pubblici? Uno scuolabus, un municipio (con due giorni di apertura settimanale), una scuola, ed è tutto. Il paese è mantenuto in vita dai bambini: i loro passi, le loro risate risuonano ogni giorno nelle strade alle 8 h 29, alle 11 h 31, alle 13 h 29 e alle 16 h 31. (MAN 04.12) ‘Where to meet? The cafes are closed. And public services? A schoolbus, a townhall (open twice a week), a school, and that’s all. The village, it is the children who give it life: their steps and their laughter resonate every day in the streets at 8:29, at 11:31, at 13:29 and at 16:31.’ Le village, ce sont les enfants ‘The village, it is the children
qui le font vivre who make it live’
Il paese è mantenuto in vita dai bambini ‘The village is kept alive by the children’ In (19b), the Italian reader is not guided in interpreting the contrast between the positive vote and the lack of enthusiasm of the Croatian population, since not only is there no emphasis on the prepositional phrase senza grande entusiasmo ‘without great enthusiasm’, but also the contrastive connective en réalité ‘in fact’ has been replaced by the consecutive così ‘thus’: (19)
a.
La Croatie deviendra, le 1er juillet 2013, le vingt-huitième Etat membre de l’Union européenne. Les électeurs ont approuvé le 22 janvier l’adhésion de leur pays. Si le “oui” a remporté près de 67% des suffrages, la très faible participation (43% des inscrits) limite la portée de ce vote. En réalité, c’est sans enthousiasme que les Croates s’apprêtent à rejoindre l’Union. Laquelle a perdu beaucoup de son attrait dans tous les pays de la région. (MD 03.12)
b. Il 1° luglio 2013, la Croazia diventerà il ventottesimo Stato membro dell’Ue. Gli elettori hanno approvato il 22 gennaio l’adesione del loro paese. Sebbene l’esito abbia premiato il “sì” con il 67%, la bassissima partecipazione (il 43% degli iscritti) limita la portata del voto. I croati si apprestano così a entrare senza grande entusiasmo in un’Unione che, in tutti i paesi della regione, ha perso gran parte del suo fascino. (MAN 03.12)
306
Giovanna Brianti
‘Croatia will become the 28th member of the European Union next year after its 67% approval vote in a referendum this January. However, the turnout was only 43%. European membership has lost its appeal for the Balkans.’ (MDE 03.12) En réalité, c’est sans enthousiasme rejoindre l’Union ‘In fact, it is without enthusiasm the European Union’ I croati si apprestano così un’Unione… ‘Croatians are about European Union…’
que les Croates s’apprêtent à that Croatians are about to join
a entrare senza grande entusiasmo in to join without great enthusiasm the
In (20b), the absence of the cleft in the Italian version creates some imbalance in the correlative interpretation of the sentence as the information structure of the original text is not maintained: (20)
a.
Tout autant qu’une mainmise d’un clan sur le FSLN, c’est cet affrontement idéologique qui a provoqué la scission. (MD 05.12)
b. Al pari dell’assoggettamento del Fsln da parte di un clan, questo scontro ideologico ha provocato la scissione. (MAN 05.12) ‘It was as much this ideological clash as the power grab by Ortega’s faction that led to the split.’ (MDE 06.12) c’est cet affrontement idéologique qui a provoqué la scission. ‘it is this ideological clash that provoked the rift’ questo scontro ideologico ‘this ideological clash
ha provocato la scissione provoked the rift’
4.2 Press Europ bidirectional corpus: Quantitative and qualitative results As shown in Table 2, the PE-corpus is much smaller than the MD-MAN-corpus, inasmuch as it amounts to about 15% only of its total amount of words. It is nevertheless interesting to compare two translation corpora in order to highlight possible differences in translation choices. Moreover, the PE-corpus is bidirectional, which gives the opportunity to check how Cleft sentences occurring in Italian
Cleft sentences. A translation perspective on Italian and French
307
newspapers are translated into French. In this section, I will first examine the distribution of explicit and implicit clefts (4.2.1); I will then briefly comment the distribution of clefts according to grammatical category (4.2.2) and I will finally illustrate the cases where a Cleft sentence is not translated with a cleft, in both directions (4.2.3). In the conclusion (4.2.4), I will compare the data extracted from the MD-MAN-corpus with those of the PE-corpus, which will emphasize some important differences due to the translation process. Table 5: PE-corpus FR-IT: Quantitative results (total number of examples and % of total clefts) Occurrences
Explicit clefts
Implicit Clefts
Press Europ FR
18
18
–
Press Europ IT
18
10 (55.5%)
3 (16.7%)
French clefts translated with other constructions
Per 100,000 Words 56
5 (27.8%)
42
Table 6: PE-corpus IT-FR: Quantitative results (total number of examples and % of total clefts) Occurrences
Explicit clefts
Implicit clefts
Italian no clefts
Press Europ IT
22
5 (22.7%)
5⁴⁵ (22.7%)
12⁴⁶ (54.6%)
Press Europ FR
22
16 (72.8%)
–
Italian clefts translated with other constructions
Per 100,000 words 52
6 (27.2%)
83
The figures given in Table 5 and Table 6 show once again that the use of clefts is more frequent in French than in Italian. This observation becomes even more relevant in the translations from Italian into French where in 12 cases the French translator inserts a cleft where the Italian original does not resort to such a construction. On the other hand, in 6 cases an Italian Cleft sentence is not translated with an equivalent construction in French.
45 This column includes 3 cases in which the Italian original resorts to implicit Reverse Cleft sentences. These constructions are not translated with a cleft in French in 2 cases, whereas in one case it is translated with an inferential cleft (see [30b]). 46 This column shows the cases in which there is no cleft in Italian (12), whereas the French translation includes an explicit cleft.
308
Giovanna Brianti
In examples (21b) and (22b), the prepositional phrases in questo spirito ‘in this spirit’ and A Giugliano ‘In Giugliano’ are inserted into a Cleft sentence in the French translation. (21)
a.
“Nessuno può sindacare su quel che facciamo”, parlava anche in questo spirito. (La Stampa, 05.01.12)
b. “Personne ne peut se mêler de ce que nous faisons”, a-t-il dit – c’est aussi dans cet esprit-là qu’il s’exprimait. (PE 05.01.12)⁴⁷ “No one can criticise what we do,” he said – was uttered in this spirit. (PE 05.01.12) parlava ‘he spoke
anche in questo spirito also in this spirit’
c’est aussi dans cet esprit-là ‘it is also in this spirit (22)
a.
qu’il s’exprimait that he spoke’
A Giugliano, vicino a Napoli, viene stampata oltre la metà del denaro falso prodotto in Europa. (La Repubblica, 23.03.12)
b. C’est depuis les alentours de Giugliano, un fief de la mafia napolitaine, que proviennent près de la moitié des faux billets en euros en circulation. (PE 23.03.12)
‘The region around the city of Giugliano, a strong-hold of the Neapolitan mafia, provides nearly half of the counterfeit euros in circulation.’ (PE 23.03.12) A Giugliano, vicino a Napoli, viene stampata oltre la metà del denaro falso ‘In Giugliano close to Naples, money’
is printed over half of the counterfeit
C’est depuis les alentours de Giugliano … la moitié des faux billets… ‘It is from the surroundings of Giugliano, … counterfeit euros…’
que proviennent près de that come half of the
47 The translations (into Italian, French or English) found on the Press Europ website are labelled PE.
Cleft sentences. A translation perspective on Italian and French
309
4.2.1 Distribution of explicit and implicit Cleft sentences As Table 5 and Table 6 show, the proportion of implicit clefts is higher in the original articles (5 occurrences in Italian and 3 in the French translation), which corresponds better to the standard use in written Italian. In the French-Italian translation, the percentage of implicit clefts (16.7%) is still relatively high as compared with explicit clefts with subjects (22.2%), which is due to the fact that the translation is less literal than in the MD-MAN-corpus, where we find proportionally more explicit clefts (25%) in the Italian translation than implicit clefts (16.3%). In the examples (23) and (24), the subject (le capitalisme sauvage ‘unbridled capitalism’/ la classe media ‘middle class’) is part of an explicit cleft in French and an implicit cleft in Italian: (23)
a.
Cette fois, c’est le capitalisme sauvage qui en fait les frais, à en juger par celles [les blagues] qui circulent actuellement à Bucarest. (Le Monde 24.08.12)
b. Questa volta, a giudicare da quelle [le barzellette] che circolano attualmente a Bucarest, è stato il capitalismo selvaggio a farne le spese. (PE 24.08.12) ‘To judge from the gags that are doing the rounds in Bucharest, this time around, runaway capitalism is the main target.’ (PE 24.08.12) c’est le capitalisme sauvage ‘it is unbridled capitalism è stato il capitalismo selvaggio ‘it was unbridled capitalism (24)
a.
qui en fait les frais that pays the price’ a farne le spese to pay the price’
È la classe media a mancare, sono i borghesi che abitavano i piani tra quello nobile e le mansarde. (Corriere della Sera, 12.04.12)
b. C’est la classe moyenne qui fait défaut, les bourgeois qui habitaient entre l’étage noble et les mansardes. (PE 12.04.12) ‘It’s the middle class that is leaving, the bourgeois who lived between the first floor and the attic.’ (PE 12.04.12) È la classe media ‘It is the middle class C’est la classe moyenne ‘It is the middle class
a mancare to be leaving’ qui fait défaut that is leaving’
310
Giovanna Brianti
4.2.2 Distribution of grammatical categories in Cleft sentences The distribution of Cleft sentences according to grammatical categories in Table 7 and Table 8 shows the prevalence of subjects both in French and Italian original articles, followed by adverbials. Table 7: PE-corpus FR-IT: Quantitative results (absolute frequency and % of total number of occurrences) Press Europ FR-IT
FR explicit clefts
IT explicit clefts
IT implicit clefts
French clefts not translated with clefts
Subjects
11 (61%)
4 (22.2%)
3 (16.7%)
4 (22.2%)
Objects
1 (5.6%)
1 (5.6%)
Adverbials
5 (27.8%)
5 (27.8%)
1 (5.6%)
Gerundives
1 (5.6%)
0
0
TOTAL GRAND TOTAL: 18
18
10 (55.6%)
3 (16.7%)
5 (27.8%)
Table 8: PE-corpus IT-FR: Quantitative results (absolute frequency and % of total number of occurrences)⁴⁸ Press Europ IT-FR
IT explicit clefts
IT implicit clefts
Subjects
2 (9.1%)
Objects
IT No clefts⁴⁸
FR explicit clefts
Italian clefts not translated with clefts
5 (22.7%) 5 (22.7%)
5 (22.7%)
3 (13.6%)
1 (4.5%)
–
1 (4.5%)
2 (9.1%)
1 (4.5%)
Adverbials
2 (9.1%)
–
6 (27.3%)
8 (36.4%)
–
Gerundives
–
–
–
1 (4.5%)
–
TOTAL GRAND TOTAL: 22
5 (22.7%) 5 (22.7%) 12 (54.5%)
16 (72.7%) 6 (27.3%) (4+2missing translations)
48 This column takes into account the cases in which the French translation inserts clefts that do not occur in the original Italian texts.
Cleft sentences. A translation perspective on Italian and French
311
In example (25), the cleft subject is a pronoun (eux / loro ‘them’): (25)
a.
“En 1985, le ministre social-démocrate des finances et son équipe avaient été surnommés ‘la droite financière’ du parti. Palme était Premier ministre, mais il a laissé faire. Ce sont eux qui ont déréglementé les marchés et libéralisé les banques”. (Le Monde, 20.02.12)
b. “Nel 1985 il ministro delle finanze socialdemocratico e la sua équipe erano stati soprannominati “la destra finanziaria” del partito. Palme era primo ministro, ma ha lasciato fare, sono loro che hanno deregolamentato i mercati e liberalizzato le banche”. (PE 20.02.12) “In 1985, the social-democratic finance minister and his team were referred to as “the financial right” of the party. Palme was prime minister, but he allowed them free rein. They were the ones who deregulated the markets and liberalised the banks.” (PE 20.02.12) Ce sont eux qui ont déréglementé les marchés et libéralisé les banques ‘It is them who deregulated the markets and liberalized the banks’ In example (26), we find a prepositional phrase (contre ce choix brutal / contro questa scelta brutale ‘against this brutal choice’) in cleft position: (26)
a.
C’est contre ce choix brutal que se bat Giovanni Savino, 33 ans, éducateur spécialisé (Le Monde, 30.03.12)
b. Ed è contro questa scelta brutale che si batte Giovanni Savino, 33 anni, educatore specializzato. (PE 30.03.12) ‘Specialist educator Giovanni Savino, age 33, has devoted his life to preventing young people from opting for this most brutal choice.’ (PE 30.03.12) C’est contre ce choix brutal ‘It is against this brutal choice
que se bat Giovanni Savino that fights Giovanni Savino’
4.2.3 Alternatives to clefts in translation Since the translation is less literal than in the MD-MAN-corpus, we find more cases where the cleft is not translated in both directions. It is interesting to note that in 3 cases out of 5, the Italian translation has a standard word order, as in (27b):
312
(27)
Giovanna Brianti
a.
“C’est une situation personnelle de solitude absolue qui m’a poussé à écrire ce texte qui a pris la forme d’une parabole: je parle d’un pays qui meurt parce qu’il n’accepte pas sa propre fin, et qu’il n’accepte pas l’autre.” (Le Monde, 17.06.12)
b. “Mi trovavo in una situazione personale di solitudine assoluta, che mi ha spinto a scrivere questo testo che ha preso la forma di una parabola: parlo di un paese che muore perché non accetta la propria fine e non accetta l’altro.” (PE 17.06.12) ‘It was my personal situation of absolute solitude that prompted the writing of the text which took the form of a parable about a country that dies because it refuses to accept its own transience, and is hostile to other identities.’ (PE 17.06.12) C’est une situation personnelle de solitude absolue qui m’a poussé à écrire ce texte ‘It is a personal situation of absolute solitude that prompted me to write this text’ Mi trovavo in una situazione personale di solitudine assoluta, ‘I found myself in a personal situation of absolute solitude che mi ha spinto a scrivere questo testo that prompted me to write this text’ In 2 cases, however, the French cleft is translated with a passive as in (28b): (28)
a.
Certains les gardaient en souvenir [les pésétas] mais avec la crise, la nostalgie n’a plus sa place, explique M. Pino. C’est Mme Ameijeira Rivas, l’opticienne, qui a eu l’idée de tout ceci, se souvenant que la banque d’Espagne avait calculé qu’il restait plus de 1,7 milliard d’euros de pesetas en circulation. (Le Monde, 06.01.12)
b. L’idea è venuta ad Ameijeira Rivas, l’ottica, che si è ricordata che la Banca di Spagna aveva calcolato che rimanevano in circolazione più di 1,7 miliardi di euro di pesetas. (PE 06.01.12) ‘Sometimes they were kept as souvenirs, but, as Mr Pino explains, people cannot afford to be nostalgic in the current crisis. It was Miss Ameijeira Rivas, the optician, who came up with the idea for the scheme, when she remembered that the Spanish central bank had
Cleft sentences. A translation perspective on Italian and French
313
calculated that more than 1.7 billion euros worth of pesetas remain in circulation.’ (PE 06.01.12) C’est Mme Ameijeira Rivas, l’opticienne, ‘It is Ameijeira Rivas, the optician,
qui a eu l’idée de tout ceci who had the whole idea’
L’idea è venuta ad Ameijeira Rivas, l’ottica ‘The idea came to Ameijeira Rivas, the optician’ In the Italian-French direction, in 6 cases the French translation does not use a cleft to translate an Italian cleft. This is particularly true with the Italian implicit Reverse clefts (29a): (29)
a.
A guidare la classifica sono gli inglesi, con concessioni record di 1.600.000 ettari, seguiti da italiani, tedeschi, francesi e nord-americani. (La Repubblica, 15.03.12)
b. Les Britanniques sont en tête, avec un record de 1,6 millions d’hectares de terres cultivées, suivis par les Italiens, les Allemands, les Français et les Américains. (PE 15.03.12) ‘The British are in the lead with a record 1.6 million hectares of cultivated land, followed by the Italians, the Germans, the French and the Americans.’ (PE 15.03.12) A guidare la classifica sono gli inglesi ‘To be in the lead are the British’ Les Britanniques sont en tête ‘The British are in the lead’ In one case the French translation resorts to an inferential cleft of the type c’est… parce que ‘it is…because’ (cf. Atayan and Wienen in this volume) with a causal interpretation (30b): (30)
a.
Ad accelerare la proliferazione della jatropha è il fatto che il biodiesel rappresenterà in futuro il 71% delle importazioni agroenergetiche UE. (La Repubblica, 15.03.12)
b. Si la culture du jatropha est si prolifique, c’est entre autres parce que le biodiesel représentera à l’avenir 71% des importations en agrocarburants de l’UE. (PE 15.03.12)
314
Giovanna Brianti
‘If jatropha plantations are so widespread, it is, among other reasons, because biodiesel is expected to represent 71% of biofuel imports to the EU.’ (PE 15.03.12) Ad accelerare la proliferazione della jatropha è il fatto che il biodiesel… ‘To accelerate the proliferation of jatropha is the fact that biodiesel…’ Si la culture du jatropha est si prolifique, c’est … parce que le biodiesel représentera…71% des importations en agrocarburants ‘If jatropha plantations are so widespread, it is… because biodiesel is expected to represent…71% of biofuel imports’
4.2.4 Comparison between the MD-MAN-corpus and the PE-corpus By comparing our two translation corpora, we can draw some tentative conclusions: – the Manifesto translation tends to be more literal, therefore the proportion of explicit clefts over implicit clefts is higher (67.3%) than in the PE-corpus (55.5%); – the Press Europ translations show more flexibility in the choice of equivalent structures to clefts. The frequency of cases where the French clefts are not translated with clefts is of 27.8% in the Italian translation from French and of 27.2% in the French translation from Italian; – the quantitative discrepancy between the two corpora is indicative of different translation strategies. In the case of the explicit clefts extracted from the MD-MAN-corpus, I would talk about a case of syntactical influence on the Italian translations (cf. 2.1), whereas the PE-corpus, by tending to avoid literal translation, shows a wider variety of options. Naturally, such an observation remains a hypothesis given the small size of the corpora.
5 Translation corpora vs. comparable corpora: Concluding remarks In this essay I have shown that, whereas the dialogue between contrastive linguistics and translation studies is certainly welcome, it is necessary to be aware of the weight of syntactic interference that inevitably characterizes the translated text
Cleft sentences. A translation perspective on Italian and French
315
as opposed to a text originally written in one language. More concretely, the data extracted from translation corpora will have to be measured against the figures drawn from comparable corpora in order to identify more objectively similarities and differences in the use of clefts in two languages. In conclusion, I will now compare the two translation corpora that I have assembled for this study with the data extracted from the ICOCP-corpus (which is a comparable corpus according to the definition given in 2.2) for Italian and French. My observations will concern the general frequency of Cleft sentences in French and Italian, the ratio of explicit and implicit clefts, as well as the distribution of cleft constituents according to their grammatical category. As Table 9 shows, the overall frequency of clefts in translated press articles, counting both translation corpora (41/100,000 words), is lower than the results of the ICOCP-corpus for French (75/100,000 words) and Italian (79/100,000 words). As expected, the French version (either original or translated) shows a slightly higher rate of clefts. This is mostly visible in the French translation from Italian in the PE-corpus, which means that French – as predicted by the compensation mechanism – will resort more often than Italian to an explicit Cleft sentence. Table 9: Parallel and comparable corpora Absolute frequency Relative frequency (per 100,000 words) Translation corpora
FR original
122
49
IT translation
100
41
IT original
10
52
FR translation
16
83
ICOCP corpus FR original
300
75
IT original
413
79
As shown in Table 10, the ratio of implicit clefts in Italian translation is very low compared with the ICOCP-corpus concerning Italian, namely about half (39% vs. 79% in the ICOCP-corpus). In the MD-MAN-corpus, the proportion of implicit clefts as compared to explicit clefts is of 39.5%. Since French does not have an implicit cleft, we see the effect of translation in the Italian target texts that give a significant preference to explicit clefts when the cleft constituent is a subject (60.5% in translation corpora vs. 21% in ICOCP).
316
Giovanna Brianti
Table 10: Explicit and implicit clefts in Italian (total occurrences, % of total number of examples of each category and relative frequency per 100,000 words) Explicit clefts with subjects
Implicit clefts with subjects
Total
Relative frequency (per 100,000 words)
IT transl.MAN
26 (60.5%)
17 (39.5%)
43 (100%)
19.5
IT transl. PE
4 (57%)
3 (43%)
7 (100%)
22.8
ICOCP IT
31 (21%)
116 (79%)
147 (100%)
38
In three cases only, the Italian translation uses an implicit Reverse cleft sentence (see [1b], [31b] and [32b]). This contrasts with the relatively high frequency of such constructions in ICOCP (38/100,000 words): (31)
a.
Après quarante-trois jours d’une mobilisation sociale sans précédent, durant laquelle les services publics et les entreprises ont tourné au ralenti, c’est un sentiment de gâchis qui prédomine. (MD 12.11)
b. Dopo quarantatré giorni di una mobilitazione sociale senza precedenti, durante la quale i servizi pubblici e le imprese hanno girato a rilento, a predominare è un sentimento di confusione. (MAN 12.11) ‘After 43 days of unprecedented social demonstrations, during which public services and businesses have been slowed down, it is a sentiment of waste that prevails.’ c’est un sentiment de gâchis ‘it is a sentiment of waste
qui prédomine that prevails’
a predominare è un sentimento di confusione ‘to prevail is a sentiment of waste’ (32)
a.
Comme les Blancs sont deux fois plus nombreux que les Noirs, c’est le Parti républicain qui gagne, sauf à certaines élections locales, comme à Jackson. (MD 04.12)
b. Poi, siccome i bianchi sono due volte più numerosi dei neri, a vincere è il Partito repubblicano, fatta eccezione per alcune elezioni locali, come a Jackson. (MAN 04.12) ‘Since the whites outnumber the blacks two to one, the Republicans win, except in certain local elections, like Jackson.’ (MDE 06.12)
Cleft sentences. A translation perspective on Italian and French
317
c’est le Parti républicain qui gagne ‘it is the Republican party that wins’ a vincere ‘to win
è il Partito repubblicano is the Republican Party’
Regarding the grammatical category of the cleft constituent, the results are globally similar to those of ICOCP-corpus in terms of the preference given to subjects, followed by adverbials and objects. As we can see in Table 11, in the ICOCP-corpus the subject is the most common cleft category (68% in IT and 56% in FR), followed by adverbials (26% in IT and 29% in FR) and objects (6% in IT and 15% in FR). Table 11: Cleft sentences (% of total clefts in every corpus) FR-MD IT-MAN (original) (transl.)
FR-PE IT-PE IT-PE FR-PE ICOCP FR ICOCP (original) (transl.) (original) (transl.) IT
Subjects
49
41.3
61.1
38.9
31.8
22.7
56
68
Objects
3.8
2.9
5.6
5.6
4.5
9.1
15
6
Adverbials
44.2
37.5
27.8
27.8
9.1
36.4
29
26
Gerundives
3
1.9
5.6
–
–
4.5
0.4
0
We can identify, however, some specific features related to the translation process. As the Italian translation of Le Monde Diplomatique is very literal, the number of clefts remains almost identical in both versions. The Italian translation of Press Europ, which is less literal, tends to produce more clefts with subjects (38.9%) rather than with adverbials (27.8%), whereas the French translation shows a higher rate of clefts with adverbials (36.4%) like the original French of Le Monde Diplomatique (44.2%). Translation corpora allow taking into account the cases in which a cleft is not translated with an identical construction. As we have seen in 4.2.3, the French translation of the Italian PE-corpus tends to insert a Cleft sentence where there is none in the Italian original but not always. This tendency can be partly explained with the compensation mechanism. It is important, however, to explore the textual reasons that motivate the use of clefts in both languages according to the context (marking transition or reducing ambiguity in the interpretation). Thanks to corpus-based approaches, translation has become a focus of interest for contrastive linguistics. As Johansson (2007) confirms:
318
Giovanna Brianti
As translation shows what elements may be associated across languages, it is fruitful to base a contrastive study on a comparison of original texts and their translations. At the same time, translations have special characteristics of their own and may deviate from original texts in the target language. Given an appropriate corpus model it is, however, possible to control for translation-specific features. (Johansson 2007: 3–4)
The contrastive study on clefts that I carried out in this article on the basis of self-assembled translation corpora intends to bring more awareness of the advantages and limits in dealing with translated material. In the case of Cleft sentences, I was able to highlight some differences in translation strategies concerning, for example, the choice of an implicit cleft in Italian as an alternative to an explicit cleft. As I have tried to show, the combined analysis of translation corpora and comparable corpora is necessary in order to distinguish between the possibilities offered by each language’s systemic conventions and the stylistic options chosen by the translator.
References Andorno, Cecilia. 2003. Linguistica testuale. Un’introduzione. Roma: Carocci. Atayan, Vahram & Ursula Wienen. This vol. Inferential Cleft constructions in translation. French c’est que in political texts. Baker, Mona. 1993. Corpus Linguistics and Translation Studies: Implications and Applications. In Mona Baker, Gill Francis & Elena Tognini-Bonelli (eds.), Text and Technology: In Honour of John Sinclair, 233–50. Amsterdam & Philadelphia: John Benjamins. Baker, Mona. 1995. Corpora in Translation Studies. An Overview and Some Suggestions for Future Research. Target 7 (2). 233–243. Baker, Mona. 1998. Réexplorer la langue de la traduction: une approche par corpus. Meta 43 (4). 480–485. Benincà, Paola. 1993. Sintassi. In Alberto A. Sobrero (ed.), Introduzione all’italiano contemporaneo. Le strutture, 247–290. Bari: Laterza. Berretta, Monica. 1994. Ordini marcati dei costituenti di frase in italiano. La frase scissa. Vox Romanica 53. 79–105. Berretta, Monica. 2002. Quello che voglio dire è che: le scisse da strutture topicalizzanti a connettivi testuali. In Gian Luigi Beccaria & Carla Marello (eds.), La parola al testo. Scritti per Bice Mortara Garavelli, 15–31. Torino: Edizioni dell’Orso. Bielsa, Esperança & Susan Bassnett. 2009. Translation in Global News. London & New York: Routledge. Bonomi, Ilaria, Elena Catalfamo, Laura Nacci & Francesca Travisi. 2002. La lingua dei quotidiani on line. In Ilaria Bonomi, L’italiano giornalistico. Dall’inizio del ’900 ai quotidiani on line, 326–329. Firenze: Franco Cesati.
Cleft sentences. A translation perspective on Italian and French
319
Cardinaletti, Anna & Giuliana Garzone (eds.). 2005. L’italiano delle traduzioni. Milano: Franco Angeli. Carter-Thomas, Shirley. 2009. The French c’est-cleft: Functional and formal motivations. In David Banks, Simon Eason & Janet Omrod (eds.), La linguistique systémique fonctionnelle et la langue française, 127–156. Paris: L’Harmattan. Chesterman, Andrew. 2010. Why study translation universals? HELDA – The Digital Repository of University of Helsinki. https://helda.helsinki.fi/bitstream/handle/10138/24319/ATH vol 1 art 4 Chesterman.pdf (accessed 22 October 2012). D’Achille, Paolo, Domenico Proietti & Andrea Viviani. 2005. La frase scissa in italiano: aspetti e problemi. In Iørn Korzen & Paolo D’Achille (eds.), Tipologia linguistica e società. Due giornate italo-danesi di studi linguistici (Roma, 27–28 novembre 2003), 249–279. Firenze: Franco Cesati. De Cesare, Anna-Maria, Davide Garassino, Rocío Agar Marco & Laura Baranzini. This vol. Form and frequency of Italian Cleft constructions in a corpus of electronic news. A contrastive perspective with French, Spanish, German and English. De Stefani, Elwys. 2009. Le strutture grammaticali come epifenomeni dell’interazione sociale? Riflessioni sull’uso delle costruzioni scisse nel parlato conversazionale italiano e francese. In Angela Ferrari (ed.), Atti del X Congresso SILFI. Sintassi storica e sincronica dell’italiano: subordinazione, coordinazione, giustapposizione (Basilea, 30 giugno–3 luglio 2008), vol. 3, 1615–1631. Firenze: Franco Cesati. Dufter, Andreas. 2009. Clefting and discourse organization: Comparing Germanic and Romance. In Andreas Dufter & Daniel Jacob (eds.), Focus and Background in Romance Languages, 83–121. Amsterdam & Philadelphia: John Benjamins. Frawley, William. 1984. Prolegomenon to a Theory of Translation. In William Frawley (ed.), Translation: Literary, Linguistic, and Philosophical Perspectives, 387–392. London & Toronto: Associated University Press. Frison, Lorenza. 1988. Le frasi scisse. In Lorenzo Renzi (ed.), Grande grammatica italiana di consultazione, vol. 1, 194–225. Bologna: Il Mulino. Garassino, Davide. This vol. Cleft sentences. Italian-English in contrast. Garzone, Giuliana. 2004. Traduzione e interferenza: il punto di vista della traduttologia. In Giuliana Garzone & Anna Cardinaletti (eds.), Lingua, mediazione linguistica e interferenza, 105–127. Milano: Franco Angeli. Gast, Volker & Daniel Wiechmann. 2012. W(h)-Clefts im Deutschen und Englischen. Eine quantitative Untersuchung auf Grundlage des Europarl-Korpus. In Lutz Gunkel & Gisela Zifonun (eds.), Jahrbuch des IDS 2011, 333–362. Berlin & New York: Mouton de Gruyter. Gil, Alberto. 2004. Textstrukturelle Dimensionen der Satzspaltung im Italienischen. In Alberto Gil, Dietmar Osthus & Claudia Polzin-Hausmann (eds.), Romanische Sprachwissenschaft. Zeugnisse für Vielfalt und Profil eines Faches. Festschrift für Christian Schmitt zum 60. Geburtstag, vol. 3, 369–385. Bern: Peter Lang. Granger, Sylviane. 2003. The corpus approach: A common way for Contrastive Linguistics and Translation Studies? In Sylviane Granger, Jacques Lerot & Stephanie Petch-Tyson (eds.), Corpus-based Approaches to Contrastive Linguistics and Translation Studies, 17–29. Amsterdam & New York: Rodopi. House, Juliane. 1997. Translation Quality Assessment. A Model Revisited. Tübingen: Narr House, Juliane. 2008. Beyond Intervention: Universals in Translation? Translationswissenschaft und Fachkommunikation 1 (1). 6–19. http://www.transkom.eu/bd01nr01/
320
Giovanna Brianti
transkom_01_01_02_House_Beyond_Intervention.20080707.pdf (accessed 22 October 2012). Johansson, Stig & Knut Hofland. 1994. Towards an English-Norwegian parallel corpus. In Udo Fries, Gunnel Tottie & Peter Schneider (eds.), Creating and Using English Language Corpora, 25–37. Amsterdam & New York: Rodopi. Johansson, Stig. 2007. Seeing through Multilingual Corpora. On the use of corpora in contrastive studies. Amsterdam & Philadelphia: John Benjamins. Kenny, Dorothy. 1998. Corpora in Translation Studies. In Mona Baker (ed.), Routledge Encyclopedia of Translation Studies, 50–53. London & New York: Routledge. Korzen, Iørn & Morten Gylling. 2012. Text Structure in a Contrastive and Translational Perspective. On Information Density and Clause Linkage in Italian and Danish. Translation: Computation, Corpora, Cognition 2 (1). 23–46. http://www.t-c3.org/index.php/t-c3/ article/view/12 (accessed 29 July 2014) Korzen, Iørn. This vol. Cleft sentences. Italian-Danish in contrast. Lambrecht, Knud. 2001. A Framework for the Analysis of Cleft Constructions. Linguistics 39 (3). 463–516. Lavault-Olléon, Elisabeth & Véronique Sauron. 2009. Journaliste et traducteur: deux métiers, deux réalités. ILCEA 11. 1–11. http://ilcea.revues.org/index210.html (accessed 20 November 2012). Laviosa, Sara. 2002. Corpus-based Translation Studies. Theory, Findings, Applications. Amsterdam & New York: Rodopi. McLaughlin, Mairi. 2011. Syntactic Borrowing in Contemporary French. A Linguistic Analysis of News Translation. Oxford: Legenda. Muller, Claude. 2003. Traduire les clivées du français en allemand. In Michael Herslund (ed.), Aspects linguistiques de la traduction, 149–167. Bordeaux: Presses Universitaires de Bordeaux. Olohan, Maeve. 2004. Introducing Corpora in Translation Studies. London & New York: Routledge. Ondelli, Stefano & Matteo Viale. 2010. Evidenze quantitative sull’italiano tradotto in un corpus giornalistico. In Sergio Bolasco, Isabella Chiari & Luca Giuliano (eds.), JADT 2010: 10th International Conference on Statistical Analysis of Textual Data. (9–11 June 2010, Sapienza University Rome), 573–584. Milano: Edizioni Universitarie di Lettere Economia Diritto. Pederzoli, Roberta. 2006. Le Monde Diplomatique in versione francese e italiana: la traduzione dei titoli anaforici. Intralinea 8. http://www.intralinea.org/archive/article/1637 (accessed 1 November 2012). Roggia, Carlo Enrico. 2008. Frasi scisse in italiano e in francese orale: evidenze dal C-ORAL-ROM. Cuadernos de filología italiana 15. 9–29. Roggia, Carlo Enrico. 2009. Le frasi scisse in italiano. Struttura informativa e funzioni discorsive. Genève: Slatkine. Schöpp, Frank. 2005. Fokus-Konstruktionen im Italienischen – mit Vergleichen zum Französischen. Beiträge zur Fremdsprachenvermittlung 43. 85–105. Sinclair, John. 1996. Preliminary Recommendations on Corpus Typology. In EAGLES http://www. ilc.cnr.it/EAGLES96/corpustyp/node1.html (accessed 22 October 2012). Tognini-Bonelli, Elena. 2001. Corpus Linguistics at Work. Amsterdam & Philadelphia: John Benjamins. Toury, Gideon. 1980. In search of a Theory of Translation. Tel Aviv: The Porter Institute for Poetics and Semiotics.
Cleft sentences. A translation perspective on Italian and French
321
Toury, Gideon. 1995. Translation Studies and beyond. Amsterdam & Philadelphia: John Benjamins. Vecchiato, Sara. 2005. Interferenza e strategie stilistiche nella traduzione dal francese all’italiano. In Anna Cardinaletti & Giuliana Garzone (eds.), L’italiano delle traduzioni, 153–192. Milano: Franco Angeli. Wagner, Emma, Svend Bech & Jesús M. Martínez. 2012. Translating for the European Institutions, 2nd edn. Manchester & Kinderhook (NY): St. Jerome Publishing Wecksteen, Corinne. 2007. Le corpus en traductologie: un moyen d’observation pour une approche réaliste de la traduction. Application à quelques phénomènes connotatifs. In Michel Ballard & Carmen Pineira-Tresmontant (eds.), Les corpus en linguistique et en traductologie. 261–282. Arras: Artois presses université.
Part II. Romance and Germanic Cleft Constructions in Contrast
Elisabeth Stark*
Frequency, form and function of Cleft constructions in the Swiss SMS corpus 1 Introduction Following Jespersen (1937) and subsequent work, Lambrecht (2001) provides the following definition of a Cleft construction (see also Ferrari et al. 2008: 224‒225 for Italian): A cleft construction […] is a complex sentence structure consisting of a matrix clause headed by a copula and a relative or relative-like clause whose relativized argument is coindexed with the predicative argument of the copula. Taken together, the matrix and the relative express a logically simple proposition, which can also be expressed in the form of a single clause without a change in truth conditions. (Lambrecht 2001: 466)
We will strictly adhere to this definition and consider Cleft constructions to be only those sentences which can be reduced without any substantial syntactic and semantic change to simple sentences with the same vericonditional properties as the corresponding Cleft sentence; see examples (1) to (4): (1)
a.
It
is
the
wife
that
decides.
b. C’
est
l’
épouse
qui
décide.
c.
la
sposa
che
decide.
ist
die
Frau,
die
entscheidet.
rel
3s.decide
È
d. Es
expl 3s.be art.def wife
* I would like to thank Anna-Maria De Cesare for having invited me to her exciting project and workshop, the Swiss National Science Foundation for substantial financial support of the project „SMS communication in Switzerland“, CRSII1_136230, and the anonymous reviewers for having helped considerably improving the paper. All remaining errors are, of course, mine. Elisabeth Stark, University of Zurich
326
(2)
Elisabeth Stark
a.
What
b. Ce que c.
I
want
je
veux,
Quello che
d. Was rel
is c’
est
music. de la
musique.
voglio
è
musica.
ich
will,
ist
Musik.
1s
1s.want
expl 3s.be
art.part
music
(Dufter 2009: 83) (3)
a.
Music
b. Musik music (4)
is
what
I
want.
ist,
was
ich
will.
3s.be
rel
1s
1s.want
(Dufter 2009: 89–90)
That’s what is difficult to understand.
The examples in (1) give English, French, Italian and German instances of a socalled it-cleft, which is constituted by an expletive subject (Engl. it, Fr. ce, G. es, also pro in pro-drop languages like Standard Italian or Spanish), a tensed form of a copula, the clefted constituent, a wh-element and the predicate attributed to the element expressed in the clefted constituent. The examples in (2) feature the same languages and are instances of a so-called wh-cleft (or pseudo-cleft; we will use the term wh-cleft in what follows), beginning with a wh-element, the predicate, a tensed form of the copula verb and the clefted constituent. Example (3) shows examples of reverse wh-clefts, non-existent in French and Italian according to Lambrecht (2001) or Van den Steen (2005), in which the clefted constituent is fronted before the copula verb, the wh-element and the predicate. So-called that-clefts, shown in (4), are similarly non-existent in French and Italian and begin with a demonstrative or presentative element, referring anaphorically back to the referent of the clefted constituent, before a tensed form of the copula, a wh-element and the predicate. For each of these examples, there exist equivalent sentences without the expletive and/or wh-elements and the copula: The wife decides and its equivalents in the other languages for (1), I want music etc. for (2) and (3) and That is difficult to understand for (4). The overall function of clefts is normally associated with textual coherence due to their specific information structure related properties (see below and e.g. Büring 2006: 145; for French wh-clefts Krötsch and Sabban 1990; Blanche-Benveniste 2010: 185), and they are no longer considered simple stylistic variants of the corresponding monoclausal propositions (cf. e.g. Engebretsen 2012: 125).
Frequency, form and function of Cleft constructions in the Swiss SMS corpus
327
In what follows, we will try to give a substantial answer to two research questions: 1) Which types of Cleft constructions do we find, how often, and in which language in our SMS data? 2) Do we see any language-specific conventionalization tendencies (“possible fossilizations of highly frequent subtypes of clefts”, Dufter 2009: 115)? As far as we know, this study is the first empirical research on the use of Cleft constructions in text messages (SMS). We will take into account clefts in nondialectal German, Swiss German, French and Italian text messages of the newly established Swiss reference corpus of text messages (cf. www.sms4science.ch). Little is known about the use of Cleft constructions in the non-standard varieties of Switzerland, especially Swiss German dialects, and we hope to enlarge our knowledge on this specific type of syntactic structure in a special form of communication, texting (cf. Dürscheid 2002), and of the German dialectal varieties used in Swiss text messages. The paper is structured as follows: after this short introduction, we will provide an overview of the most important findings in the literature concerning the function of Cleft constructions and their language specific distribution in different types of communication forms and/or registers (Section 2). From this, we will formulate three guiding hypotheses for our analysis. Section 3 will briefly introduce our database, the newly established Swiss reference corpus of text messages (cf. www.sms4science.ch), before we present our results in Section 4. Section 5 seeks to interpret these findings against the background of preceding studies on clefts in German and Romance languages, and a short conclusion in Section 6 will reiterate the most important insights we have gained from our research.
2 Distribution and functions of Cleft constructions: State of the art and hypotheses Although no research has been conducted into the distribution of Cleft constructions in new forms of communication, especially in computer-mediated communication (CMC),¹ the specialized functional literature gives some insights into their functional and variationist profile, which might help develop some guiding hypotheses about their probable distribution in our data. 1 A search for “Cleft constructions” and “computer-mediated communication” in the data bases LLBA and MLA returned zero results.
328
Elisabeth Stark
Prince (1978) and Collins (1991) state that it-clefts occur more often in (formal) writing. They seem to be typically used in written discourse, whereby the more formal conception of the text (type) in question appears more important than the medium (i.e., the phonic or graphical realization; cf. Koch and Oesterreicher [1990] 2011 for this fundamental distinction). This is in line with Roggia’s (2008) results on the overall distribution of it-clefts in French and Italian, based on the phonic corpus C-ORAL-ROM (cf. Roggia 2008: 19). As for the influence of the medium on French Cleft sentence types, contrary to Prince’s statement, Dufter (2008: 42) finds that “[…] c’est-clefts [= the French equivalent to it-clefts, ES] occur about 2.5 times more often in speech than in writing […]”. Dufter (2008) based his research on the corpus C-ORAL-ROM for phonic (formal and informal situation types; cf. Cresti and Moneglia 2005) and on FRANTEXT for formal written French. Bearing these divergent statements in mind, we will formulate our Hypothesis I as follows: it-clefts and their equivalents are expected to appear quite rarely in text messages, the latter ones being typical instances of informal communication (cf. the discussion in Stark 2011). According to Collins (1991), wh-clefts (and reverse wh-clefts) are typically found in spoken language, especially in dialogues between participants who are (close) friends. This means that the rather informal conception of a message favors the occurrence of wh-clefts. Here, we are confronted additionally with a language-specific bias (which holds generally for the use of Cleft sentences, see below): they seem to be quite rare in French (cf. Van den Steen 2005: 277) and German (cf. the results of Gast and Wiechmann 2012 for the use of wh-clefts in English vs. German in the EUROPARL corpus). We thus have to formulate Hypothesis II in two parts: a. wh-clefts should be particularly frequent in text messages; b. wh-clefts are in general rather infrequent in the French, German and probably also Italian subcorpora. Concerning language-specific preferences for certain Cleft construction (sub-) types, Dufter (2009), in his empirical investigation of translations of English itclefts in the EUROPARL-corpus into French, Portuguese, Italian, Spanish and German, notices a decreasing likelihood of translating English it-clefts with Cleft sentences, in the above-mentioned order. While the equivalents of it-clefts are used frequently in French and Portuguese, they are used slightly less frequently in Italian and Spanish, and German uses different constructions, especially Verb (V)-Subject (S) word order, to substitute them. (For another empirical study corroborating this view on language specific frequencies and functions of clefts, cf. Engebretsen 2012). In order to explain these empirical findings, Dufter (2009) reminds us of an important functional subcategorization of it-clefts, arguing against a monofunctional approach (cf. particularly Lambrecht 1994, 2001), which would see focus marking as the only function of it-clefts. Prince (1978)
Frequency, form and function of Cleft constructions in the Swiss SMS corpus
329
distinguishes essentially three sub-types of it-clefts (see also e.g. Gundel 2008): first, a Type A, called stressed-focus cleft, where either the clefted constituent constitutes the only narrow focus of the sentence (example [5]) or which may have also multiple foci (example [6]): (5)
It was Peter who stole the money (not John).
(6)
It is the money that we have to save and the economy that we have to rebound.
Second, it-clefts may be of Type B, i.e., informative-presupposition cleft, where the focus is restricted to the embedded clause. This type is especially frequent with clefted adverbials. One subtype, frequently found with manner or cause adverbials, is also called the cohesive cleft (strongly preferred in French and Portuguese, but almost non-existent in German and not so frequent in Italian or Spanish; cf. also Van den Steen 2005; Roggia 2008): (7)
It is for this reason that […] the Commission has decided to make 2004 the European Year of Education through Sport. (Dufter 2009: 103)
Finally, Type C it-clefts can be called all focus / all new and are often found at the beginning of a discourse act or unit. They are almost non-existent in German. Socalled presentational clefts or have-clefts with this function, such as in example (8), are particularly common in French (cf. also Lambrecht 2001, ch. 4). They will thus be taken into account as a structural variant of Type C it-clefts, even if this denomination might be slightly misleading (we are following Lambrecht 2001 here in using the English terms for non-English structures which are, however, functionally equivalent to them; cf. Lambrecht 2001: 468): (8)
A: Qu’est-ce qui t’ arrive? wh 2s.obj 3s.happen ‘What’s wrong with you?’ B: J’ ai ma 1s 1s.have 1s.poss ‘My mother is ill’
(B is crying)
mère qui est malade. mother rel 3s.be sick
Dufter (2009: 108–109) explains that it-clefts are chosen only rarely in German to translate the English occurrences in the EUROPARL-corpus because German has the ability to permit in situ constituent scope marking, topicalization (V-S-
330
Elisabeth Stark
inversion) and scrambling, which is heavily restricted or non-existent in French (Dufter 2009: 109–110; for similar tendencies in Spanish vs. French see Van den Steen 2005; for comparable findings on French vs. Italian see Roggia 2008). Modal particles are also used in German to signal presupposed content (Dufter 2009: 110–112). These alternative structures make possible the overall avoidance of Type B and C in German, which indicates different conventionalization tendencies for different functions of Cleft sentences in different languages. These observations lead to our Hypothesis III: We expect more clefts in French than in Italian (see also Roggia 2008 for this assumption) and still slightly more in Italian than in the German / Swiss German text messages, especially where Type B and Type C-it-clefts are concerned. Let us now summarize our three hypotheses: Hypothesis I: it-clefts and their equivalents are expected to appear quite rarely in text messages, as they are typical instances of informal communication. Hypothesis II: a. wh-clefts should be particularly frequent in text messages; b. wh-clefts are in general rather infrequent in the French, German and probably also Italian subcorpora. Hypothesis III: We expect more clefts in French than in Italian and more in Italian than in the German / Swiss German text messages, especially where Type B and Type C-itclefts are concerned.
3 Database: The Swiss SMS corpus Our study is based on the newly established corpus SMS4science.ch. In the fall / winter of 2009 and in the spring of 2011, 25,947 authentic text messages were collected with the support of Swisscom, the most important telecommunication provider in Switzerland, as part of a larger international project (cf. www.sms4science. org). The corpus comprises about 500,000 words. In total, 2,784 participants sent their original text messages to a free phone number. Of course, we cannot verify the “authenticity” of the data (i.e., we cannot exclude that someone wrote a text message precisely for the study, thus not or less reflecting his or her normal usage), but we have to accept this general methodological problem for any kind of empiri-
Frequency, form and function of Cleft constructions in the Swiss SMS corpus
331
cal study where speakers or writers know that their language products will be analyzed. The messages have subsequently been anonymized and are currently set up in a corpus that constitutes an online resource for scientific needs (cf. Dürscheid and Stark 2011 for further details on the corpus). Out of the 2,784 participants, 1,316 filled in an anonymous online questionnaire containing sociodemographic information about their sex, age, regional origin, mother tongues, education, reading and writing habits etc. (cf. www.sms4science.ch for overall information about the average sex, age etc. of the participants). Furthermore, 10,718 text messages are written in Swiss German dialects, 7,224 in non-dialectally marked German, 4,627 in French, 26 in a French dialect or Francoprovençal, 1,481 in (also dialectally marked) Italian, 1,085 in Romansh and 539 in English (the remaining 247 text messages contain Slavic languages, Scandinavian languages etc.). For this study, we searched manually for Cleft constructions of all types in the Swiss German, non-dialectal German, French, Italian and English text messages.² There were no Cleft constructions found in the English text messages, which leaves us with the following four languages and varieties, respectively, constituting the base of our findings: Table 1: Number of text messages / words per language / variety with Cleft constructions
SMS
Words (approximately)
Swiss German
10,718
218,000
Non-dialectal German
7,224
140,000
French
4,627
94,000
Italian
1,481
30,500
The following examples³ illustrate the kind of linguistic variation we find in the text messages from Switzerland: (9)
oh schnuggiputz hallo, i miss you! Du hesches bald gschafft :-) vernünftig bisch mitem zug gange, super! Cu soon :-*. (Swiss German, English) ‘Oh Darling, hello, I miss you! You have almost made it :-) It was reasonable that you took the train, super! See you soon :-*’
2 The manual research by reading each text message was effectuated by Alexandre Torea, a master’s student at the University of Zurich, whom we thank for this valuable empirical work. 3 The English translations of the text messages remain sometimes voluntarily non-idiomatic and very literal in order to make clear the original structure of the messages.
332
Elisabeth Stark
(10)
Na, hats bei euch auch geschneit? Ich kann heut abend so gegen 20.30 bei euch sein. Schick mir doch ne kurze Antwort, wenn dich meine SMS erreicht hat. CU Jörg (non-dialectal German, English) ‘So, did it snow at your place as well? I can be at your place at around 20.30. Please send me a short answer when you have got my text message. See you, Jörg’
(11)
Mé fiche de Geo son ché toi?t la se soir pr ke je viene les chercheR? (French) ‘Are my notes of geography at your place? Are you at home this evening so that I can come and get them?’
(12)
Hey scusami ho visto il tuo sms solo oggi. Bè è come l’hai detto te, la lingua nn è x niente facile. Pure io sono convinta che l’abbia dato a qualke persona (Italian, English) ‘Hi, sorry, I have only seen your text message today. Well, as you said, the language is not easy at all. And I am convinced as well that s/he has given it to some other person’
(13)
Joey angel, am still @mymother’s place where we had dinner w my parents. Miss u, xxx sss (English)
As illustrated in these examples, text messages deviate from the written standard, especially with regard to register-specific spelling strategies such as the absence of apostrophes or the integration of apocopes, imitating the spoken language (cf. German hats instead of hat’s for hat es, ‘has it’, or heut instead of heute, ‘today’, in [10]); “phonetic spelling” (cf. e.g. Anis 2007), particularly frequent in French, where the phoneme-grapheme-relation is very indirect and where silent consonant letters are very frequent and accordingly omitted in text messages (e.g. mé for mes, ‘mine’, son for sont, ‘are’, etc., in [11]); letter-and number-homophones (cf. Thurlow 2003), like t for homophonic t’es, informal French for tu es, ‘you are’, in [11]); consonantal skeletons (like pr for pour, ‘for’, in [11]); logographs like x for per, the mathematical operation, but also the preposition per, ‘for’, in Italian (cf. example [12]), and all kinds of abbreviations like w for with in (13). This enumeration is far from being exhaustive, and we will not go into the details of these specific spelling strategies (cf. for more information, see Dürscheid and Stark 2013), as they do not directly affect the number, distribution or formal set-up of the Cleft sentences we have found in our corpus. The empirical results of this study are presented in the next section.
Frequency, form and function of Cleft constructions in the Swiss SMS corpus
333
4 Results 4.1 Overall results The following table shows, in absolute numbers, all Cleft constructions and Cleft construction types found in the Italian, non-dialectal German, Swiss German and French text messages of our corpus: Table 2: Overview of Cleft construction types in the corpus
it-clefts
wh-clefts
Reverse wh-clefts
Italian
5 (4+1 have)
–
–
Non-dialectal German
1
2
–
Swiss German
9
1
–
French
55 (38+17 have)
3
–
As displayed in Table 2, the absolute numbers are very low, with one exception: French, where we find 55 it-clefts, among which 17 are formed with avoir (Hello tu pourrai m excuser auprès de Chloé pcq g un genou qui me fait vraiment mal [hello 2s could 1s.obj apologise to Chloé because 1s.have art.indef knee rel 1s.obj hurt much] ‘Hello, could you excuse my absence to Chloé because I have a knee that is really aching…’). A typical example with a clearly contrastive focus on the clefted constituent is the following one from our SMS corpus: (14)
Ce n’est pas ce jeudi que je m’absente pour voir les anciennes photos de Rafz, mais jeudi prochain et Goran sera là.Bonne soirée!Patricia (French, it-cleft) ‘It is not this Thursday that I will be off to watch the old pictures of Rafz, but next Thursday, and Goran will be there. Good night! Patricia’
(15)
Liebe Florina, was ich dir anbieten kann sind Karten für die letzten beiden Vorstellungen am 5./6. Dezember, ginge das? PS […] (German, wh-cleft) ‘Dear Florina, what I can offer you are tickets for the last two presentations on December 5th or 6th, would that be fine? PS […]’
This German example is a clear case of a wh-cleft, a type that is rare in the corpus and even more so in the Romance subcorpora (see above, Table 2). It is important to note that clefts are generally very infrequent in our corpus, and that they only appear in French in slightly over 1% of the respective text mes-
334
Elisabeth Stark
sages (see Table 3). This is perfectly in line, for example, with the findings of Roggia (2008:15), who compares French and Italian clefts in the corpus C-ORALROM: Table 3: Percentages of text messages with clefts
it-clefts
wh-clefts
Reverse wh-clefts
Italian
5/1,481 = 0.34%
–
–
Non-dialectal German
1/7,224 = 0.014%
2/7,224 = 0.028%
–
Swiss German
9/10,718 = 0.08%
1/10,718 = 0.01%
–
French
55/4,627 = 1.19%
¾,627 = 0.07%
–
4.2 Functions of it-clefts If we have a closer look at the examples, it is sometimes difficult to decide which type of it-clefts (as those constitute the majority of our occurrences) we are confronted with: Type A, stressed focus cleft; Type B, informative-presupposition cleft; or Type C, all focus / all new (see above, Section 2): (16)
No nn sn io ke ti pacco sempre, 6 tu ke arrivi sempre al momento sbagliato J (Type A, multiple-focus?) ‘It is not me who snubs you always, it is you who is always arriving at a bad moment’
(17)
Sisch im fall gar nöd s telefon, wo kaputt isch, sondern d sim charte! Ha jetz jensti grät usprobiert und alli hend di glich fehlermäldig age! (Type A) ‘It happens to be not the phone which is broke but the SIM card! I have tried many different phones, and all indicated the same error!’
(18)
Coucou!c’est pas aujourd’hui qu’on va voir ce film à l’abc?c’est à quelle heure? (Type A) ‘Hi ! Isn’t it today that we will go and watch that movie at abc? At which time?’
We consider these three examples as belonging to Type A, i.e., the clefted constituent is in narrow focus. Examples (16) and (17) are cases where this is particu-
Frequency, form and function of Cleft constructions in the Swiss SMS corpus
335
larly evident, as the contrast is made explicit in the second coordinated sentence. Example (18) is more problematic as the contrast is not easy to reconstruct, thus it could contain an emphatic focus just to highlight the fact that the appointment in question will indeed take place that same day. Example (19) belongs to Type B, i.e., a type of it-cleft in which the focus is on the embedded sentence rather than the clefted constituent, which takes up something that has been said before by means of an anaphoric elements (here: ça, ‘that’). (19)
C’est pour ca que je ne prends JAMAIS les trains le dimanche soir… Sinon tu vas? Et ton Damian? […] (Type B) ‘It is for that reason that I never take the train on Sunday evenings… Will you go there? And your Damian?’
(20)
Oups, moby isch ja no. Grad temporär vergässä, tztztz- ja, bi glaub chli zimli müed ;-) wohnig nöd okay, nun- das isch nöd die gsi wo uf dich wartet, ganz eifach :-) gmüetliche abig u live bis i ca. 48std. Grueß vo frau fatigue (Type C) ‘Oops, moby will be. I forgot that temporarily, tztztz – well, I think I am quite tired ;-) Apartment was not okay, well – that was not the one which waited for you, simply J cosy evening, and live in about 48 hours. Greetings from Mrs Tired’
(21)
Non,je suis entrain d’arriver,c’est ma maman qui m’amène!j’te dis quand j’arrive!à toute,becs (Type C) ‘No, I am about to arrive, it is my mum who gives me a ride! I will tell you when I am there, see you in a minute, bye’
Examples (20) and (21) are examples of Type C, i.e., the information in the Cleft sentence is all new (das isch nöd die gsi wo uf dich wartet, Swiss German, example [20]; c’est ma maman qui m’amène, French, example [21]). In example (20), die (‘it’) is in a rather sloppy anaphoric relation to wohnig (‘apartment’), not referring back to the visited apartment, but to the one and only apartment that is potentially waiting for the author of the text message. Additionally, the negation of the fact that the visited one was the one “that waited” for the author of the text message renders the proposition in the whole sentence new. In (21), this is even more evident, as the mother of the author of the text message is not aforementioned, nor is the fact that s/he is being given a ride.
336
Elisabeth Stark
(22)
ciao bel…un’ora fa sono uscita dopo 3 ore di esame di chimica…ahahah… divertente..-.-” .. vabbé..dovrei averlo passato.cmq..sono mega distrutta psicologicamente :-/ ho la testa che scoppia, e domani ho fisica…=( […] (Type C, presentational have-cleft) ‘Hi beautiful one…one hour ago, I left after 3 hours of chemistry exams… ahahaah…amusing..-.-=…well..I should have passed it.. However, I am completely destroyed psychologically ;-/ I have the head that is exploding, and tomorrow, it will be physics…=( […]’
(23)
Yo,bien débuté la semaine?ya samuel qui propose de prendre l’auto pour ce soir.Tu aimerais venir avec? (Type C, presentational have-cleft) ‘Yo, did you have a good start to the week? There is Samuel who proposes to take the car for tonight. Would you like to join us?’
As for their information structure, examples (22) and (23) are quite similar to the preceding examples (Type C, all new); however, they show a morphosyntactic peculiarity: instead of an expletive subject and a form of the copula verb to be, we find either the first person or an expletive and a form of to have (ho la testa che scoppia [1s.have art.def head rel 3s.burst] ‘I have the head that is exploding’, Italian, example [22]; ya Samuel qui propose…[expl.loc.have Samuel rel 3s.propose] ‘There is Samuel who proposes…’, French, example [23]) as the predicate of the clefted part of the sentence. This is quite frequent in Romance languages (presentational cleft, see above, Section 2). If we ask ourselves how frequently these subtypes of it-clefts occur in our corpus and in the respective subcorpora, we find the following results: Table 4: Overview of it-cleft functions in the corpus
Type A
Type B
Type C
Italian
2 (40%)
–
3 (60%)
Non-dialectal German
1 (100%)
–
–
Swiss German
5 (55.6%)
–
4 (44.4%)
French
25 (45.46%)
4 (7.27%)
9+17 have-clefts (47.27%)
Thus, only French exhibits the whole range of possible it-cleft functions in our data, and it shows even more all new and presentational it-clefts (Type C) than the other two types, such as Italian (which has, however, very few attested examples). Types B and C have not been observed in our non-dialectal German text
Frequency, form and function of Cleft constructions in the Swiss SMS corpus
337
messages. Up to this point, our results confirm the overall tendencies observed and partially explained by Dufter (2009) for the EUROPARL-corpus (see above, Section 2). What is striking, given the fact that Swiss German is a diatopic variety of German, is that Type C is quite frequent in Swiss German text messages, a fact that – to our knowledge – has never been stated before.⁴ In what follows, we will look more closely at French Type C occurrences in our corpus. Example (24) demonstrates the classification difficulties we occasionally encountered (and which might have slightly altered the quantitative results presented in Table 4), as text messages are often embedded into longer dialogues which we ignore. Thus, we do not know whether Marion in example (24), below, has been mentioned before or whether it is part of a list of alternatives, and whether Jose being picked up by somebody is already given knowledge. If this was the case, we would face a Type C-occurrence here, but if Marion is one of several potential people to pick up Jose, the example could also be a Type A-occurrence of a French it-cleft. (24)
Hello. C est marion qui uient chercher jose ce soir. (Type A or C?) ‘Hello. It is Marion who will come and pick up Jose this evening.’
Having a closer look at rather unambiguous occurrences of Type C in the French subcorpus, we observe certain lexicalization tendencies, which have also been mentioned by Dufter (2009). Thus, 5 out of 9 Type C-c’est-clefts show the manner adverbial avec plaisir/avec joie (‘with pleasure / joy’) as the clefted constituent: (25)
Ah c gentil! :) C avec plaisir ke je serai venu ms je pourrai pa etre la…Ms fais signe pr le prochain coup! Bonne journée, Gui ‘Ah, that is nice ! J It is with pleasure that I would have come, but I will not be able to be there…But please inform me next time! Have a nice day, Gui’
(26)
oki..ben j pense qui va trouver Ça bizar aussi..lol euh oui j suis tjs partante:-) c’est avec plaisir de repasser une soirée avec toi:) hi bisous chérie Jtm⁵ ‘OK, well, I think that somebody will find that strange, too…lol euh yes, I am still taking part :-) It is with pleasure that I will spend another evening with you ;-9 Hi, kisses, darling, I love you’
4 Elvira Glaser, personal communication. 5 Here, the Cleft sentence is structurally incomplete (it would have been c’est avec plaisir que je repasse une soirée avec toi). The author of the text message most probably mixed up two differ-
338
Elisabeth Stark
(27)
Quand tu aura un moment c’est avec grand plaisir que je T’invite :) ca serait trop cool! Gros bisous ‘Once you find a moment, it is with great pleasure that I will invite you J that would be too cool! Big kisses’
(28)
On peut ariver dans laprem,et c avec joie ke je te donne un kou de main! ‘We can arrive in the afternoon, and it is with pleasure / joy that I will help you a bit’
(29)
Coucou Lexi,c est avec joie,que je viendrai demain apres-midi voir votre princesse,vers 16h00,youpiii! Je n aurai pas le temps de faire un cake,mai ca n est que partie remise,promis;-) […] ‘Hi there Lexi, it is with pleasure / joy that I will come tomorrow afternoon and see your princess, around 16h00, yippee! I will not have the time to make a cake, but that is only suspended, I promise;-) […]’
The formulaic character of these clefts is easy to detect. Example (30) is lexically similar, but with a different information structure: (30)
Dommage pour moi c est pas l envie qui manque d y manger une fois … mais je ne suis pas la je l ai dit à nicky hier au boulot je vous souhaite une bonne soirée (Type A or C?) ‘That’s a pity for me, it is not the will that is missing to eat there once… but I will not be there I told Nicky yesterday at work I wish you a nice evening’
Here, the clefted constituent (the subject of the unmarked corresponding sentence l’envie d’y manger ne manque pas) also expresses an emotional state of pleasure / joy / will, but the negation in the clefted part permits an alternative interpretation of this c’est-cleft as Type A (‘it is not the will, but the occasion…’). The following two examples are even closer to lexicalization of the whole Cleft construction than examples (25) to (29) above:
ent syntactic constructions (the clefted one and c’est un plaisir de repasser une soirée avec toi). Informationally speaking, the resulting expression is nevertheless clearly Type C of our it-clefts and will accordingly be considered as such.
Frequency, form and function of Cleft constructions in the Swiss SMS corpus
339
(31)
C toi ki voi.ben passe me prendr ver NNNN,je devrai etr pret.tte (Type A or C?) ‘It’s up to you, well, come and pick me up around NNNN, I should be ready by then’
(32)
Bon… C’est toi qui vois.. J’espere que ca ira mais mm a 2h du mat tu sais ke tu px venir hin ;) Prends soin de toi, bisous JE T’AIME ‘Well…It’s up to you…I hope that it will be possible, even at 2 o’clock in the morning you know that you can come there ;) Take care, kisses I love you’
Even if we could still imagine a literal reading of (31) in the sense that the addressee of this text message (toi, ‘you’) is the one to decide, in contrast to the sender or somebody else, this interpretation is very improbable in (32). In our opinion, C’est toi qui vois, which occurs twice in our corpus, is lexicalized, just as it’s up to you (to decide), an original Cleft structure also found in English.
5 Discussion As we have seen in Section 4, we find relatively few occurrences of Cleft constructions in our data. This is certainly partially explained by the fact that text messages are graphical, but rarely conceptually written; they are also different from dialogical and argumentative speech in that direct interaction and online observation of the interlocutor’s performance are not possible, among other aspects. They are short, as their original name tells us (Short message service, SMS), and only quite rarely deploy an internal textual structure where clefts are needed as coherence-creating devices. French prevails in that it exhibits 58 Cleft constructions altogether (see Tables 2 and 3), a number far outranking the findings for Swiss German (10), Italian (5) and non-dialectal German (3). This is a confirmation of the often stated preference French has for clefts (see e.g. Dufter 2008, where he discovers about three times higher percentages of clefts in his French corpora compared to the study of Collins 1991 on English), but the extreme differences in the use of Cleft sentences certainly require further comparative studies on clefts in computer-mediated communication. As for the structural sub-types, we have found only very few wh-clefts in our subcorpora (and relatively more in non-dialectal German than in the other subcorpora, as expected); it-clefts constitute by far the most important type.
340
Elisabeth Stark
We have also found a surprising partial confirmation of the general observation (cf. Miller 2006) according to which the frequency of (it-) Cleft sentences diminishes in the languages of Europe from West to East. We have found the most clefts in the French subcorpus, followed by the Italian, followed by the Swiss German and subsequently the non-dialectal German data, and these results correspond grosso modo to Dufter’s (2009) study using the EUROPARL-corpus. If one considers any map of Switzerland, one will state that French is the language spoken in the Western part (Romandie), close to the Swiss German area. Germany and Austria, both regions where Standard German is also spoken as a mother tongue, are situated to the (North-)East of Switzerland. Perhaps the fact that the position of Swiss German results in intensive contact with French provides an explanation for the different figures we have found in our Swiss German subcorpus compared to the non-dialectal German one (see Tables 2 and 3, above). In fact, there are relatively more Cleft sentences in Swiss German than in nondialectal German, and we have found even Type C it-clefts (example [20]), which could not easily be translated literally into non-dialectal or Standard German. Furthermore, the expected presence of presentational have-clefts in Italian and French is confirmed by our data (Type C; see Table 4 and examples [22] and [23]). Type B is attested only in French. Among the Type C occurrences in French, a considerable number of them show clear lexicalization tendencies (c’est avec plaisir / joie que, c’est toi qui vois; cf. examples [25] to [29] and [31] and [32]). Even if these results are generally in line with, for example, Dufter’s findings (2009), we must admit that we encountered several classification problems due to missing context (cf. examples [24], [30] and [31], at least). If we compare our results to our initial three hypotheses (see above, end of Section 2), the following conclusions can be drawn: Hypothesis I is confirmed, i.e., it-clefts and their Romance and Germanic structural equivalents do not appear very frequently in text messages, at least not in our corpus. Apparently, the informal nature is not as important when it comes to the appearance of certain linguistic features or stylistic variants as the medium in which informal communication is conducted – results for negation marking in French text messages and chats show similar tendencies, in that the expected features of informal phonic French do not appear to the same extent in these informal, but graphic communication forms (cf. Van Compernolle 2008; Stark 2012). Hypothesis II is not confirmed, but language-specific restrictions formulated therein are confirmed. Thus, we find some wh-clefts in the German subcorpus, but to a much lesser extent than expected, and the German-Romance contrast is borne out (see Tables 2 and 3). Furthermore, 2 out of 3 non-dialectal German cleft
Frequency, form and function of Cleft constructions in the Swiss SMS corpus
341
constructions are wh-clefts, 1 out of 10 Swiss German, and only 3 out of 58 French clefts are wh-clefts, with no attestation of a wh-cleft in Italian at all. Hypothesis III is confirmed. We do find significantly more clefts in French than in Italian and more in Italian than in German and Swiss German text messages, especially as far as Type B and C-it-clefts are concerned, which is in line with Dufter’s (2009) findings. What is particularly striking here is the strong prevalence of French clefts in our data, which represent more than 11 times the Italian and German occurrences (Fr. 58, It. 5, G. 3 occurrences), and more than 10 times the Swiss German occurrences. While the French preference for clefts over other focalizing and coherence marking devices is a well-known fact, the extreme quantitative imbalance in the usage of clefts in our data is still striking. We leave the search for a possible explanation to future research, but would like to conclude this discussion section with a brief general reflection on the variationist profile of text messages. Robust quantitative results for certain linguistic features of (French) text messages that allow further-reaching conclusions about their variationist profile are not available in great magnitude at this point in time, and systematic empirical sociolinguistic and stylistic research is still almost absent for the new electronic communication forms in French. Yet, some pilot studies such as Van Compernolle (2008), Stark (2011, 2012) and Stark and Riedel (2013) indicate that the medial aspect of text messages’ graphical nature plays a greater role than expected. Consequently, despite the fact that we do find typical features of informal French in text messages (negation particle drop, absence of past participle agreement, subject-verb disagreement etc.), we find it to a lesser extent than in informal corpora of phonic French. Our findings on clefts and their distribution point in a similar direction: the varieties used in text messages are most certainly different from speech, but not as informal as often assumed (see the fact that more it-clefts are found than expected, in particular more Type C it-clefts than Type A it-clefts). Of course, there is also the factor of the reduced textual characteristics of text messages, which renders a high frequency of clefts improbable, as these are often considered to be coherence creating devices. Very often, there is a reduced necessity or possibility for exhaustive / contrastive foci against a presupposed background, and there is a greater tendency towards all new-information All in all, we consider text messages in particular and the new electronic communication forms in general as a fruitful empirical domain to learn more about language usage and medium dependent and/or register variation.
342
Elisabeth Stark
6 Conclusion Our investigation into the types and distribution of clefts in the newly established Swiss reference corpus of text messages, more specifically into its French, Italian, Swiss German and non-dialectal German subcorpora, has shown that our data confirms the French preference for clefts in contrast to the other three languages and varieties, but that clefts are in general quite infrequent in text messages, perhaps due to their graphic and not precisely textual nature. It has revealed that the all new subtype C of it-clefts (following Prince 1978) prevails in French and Italian, and that Swiss German text messages show surprisingly more clefts than the non-dialectal German text messages, a fact not noticed until now and which needs further investigation and explanation. Finally, this study has revealed that text messages are not just “transcribed informal speech”, and that their graphic nature and specific communicative profile leads to a different distribution of certain variety markers or indicators when compared with informal phonic corpora, especially for French. Further research is required in order to understand more about language use and probably also language change happening in computer-mediated communication.
References http://www.frantext.fr http://www.sms4science.ch Anis, Jacques. 2007. Neography: Unconventional spelling in French SMS text messages. In Brenda Danet & Susan Herring (eds.), The Multilingual Internet: Language, Culture and Communication Online, 87‒115. New York: Oxford University Press. Blanche-Benveniste, Claire. 2010. Les pseudo-clivées et l’effet deux points. In Marie-José Béguelin, Mathieu Avanzi & Gilles Corminbœuf (eds.), La parataxe. Structures, marquages et exploitations discursives, vol. 2, 185‒217. Bern: Peter Lang. Büring, Daniel. 2006. Intonation und Informationsstruktur. In Hardarik Blühdorn, Eva Breindl & Ulrich Hermann Wassner (eds.), Text – Verstehen: Grammatik und darüber hinaus, 144‒163. Berlin & New York: Mouton de Gruyter. Collins, Peter. 1991. Cleft and Pseudo-cleft Constructions in English. London: Routledge. Cresti, Emanuela & Massimo Moneglia. 2005. C-ORAL-ROM. Integrated Reference Corpora for Spoken Romance Languages. Amsterdam & Philadelphia: John Benjamins. Dürscheid, Christa. 2002. E-Mail und SMS – ein Vergleich. In Arne Ziegler & Christa Dürscheid (eds.), Kommunikationsform e-mail, 93‒114. Tübingen: Stauffenburg. Dürscheid, Christa & Elisabeth Stark. 2011. SMS4science: An international corpus-based texting project and the specific challenges for multilingual Switzerland. In Crispin Thurlow & Kristine Mroczek (eds.), Digital Discourse: Language in the New Media, 299‒320. Oxford: Oxford University Press.
Frequency, form and function of Cleft constructions in the Swiss SMS corpus
343
Dürscheid, Christa & Elisabeth Stark. 2013. Anything goes? SMS, phonographisches Schreiben und Morphemkonstanz. In Martin Neef & Carmen Scherer (eds.), Die Schnittstelle von Morphologie und geschriebener Sprache. 189–209. Berlin & New York: Mouton de Gruyter. Dufter, Andreas. 2008. On explaining the rise of c’est-clefts in French. In Ulrich Detges & Richard Waltereit (eds.), The Paradox of Grammatical Change: Perspectives from Romance, 31–56. Amsterdam & Philadelphia: John Benjamins. Dufter, Andreas. 2009. Clefting and discourse organization: Comparing Germanic and Romance. In Andreas Dufter & Daniel Jacob (eds.), Focus and Background in Romance Languages, 83–121. Amsterdam & Philadelphia: John Benjamins. Engebretsen, Steinar. 2012. Norwegische Satzspaltungen und ihre analogen bzw. nichtanalogen deutschen Entsprechungen: eine korpusbasierte Untersuchung. Deutsche Sprache 40 (2). 124‒150. Ferrari, Angela, Luca Cignetti, Anna-Maria De Cesare, Letizia Lala, Magda Mandelli, Claudia Ricci & Enrico Roggia. 2008. L’interfaccia lingua – testo. Natura e funzioni dell’articolazione informativa dell’enunciato. Alessandria: Edizioni dellʼOrso. Gast, Volker & Daniel Wiechmann. 2012. W(h)-Clefts im Deutschen und Englischen. Eine quantitative Untersuchung auf Grundlage des Europarl-Korpus. In Lutz Gunkel & Gisela Zifonun (eds.), Jahrbuch des IDS 2011, 333‒362. Berlin & New York: Mouton de Gruyter. Gundel, Jeannette K. 2008. Contrastive perspectives on cleft sentences. In María de los Angeles, J. Lachlan Mackenzie & Elsa M. Álvarez Gonzalez (eds.), Languages and Cultures in Contrast and Comparison, 869‒887. Amsterdam & Philadelphia: John Benjamins. Jespersen, Otto. 1937. Analytic Syntax. London: Allan & Unwin. Koch, Peter & Wulf Oesterreicher. 2011 [1990]. Gesprochene Sprache in der Romania. Französisch – Italienisch – Spanisch. Berlin & New York: Mouton de Gruyter. Krötsch, Monique & Annette Sabban. 1990. “Bleu, je veux” – Remarques sur la focalisation en français. Zeitschrift für romanische Philologie 96 (1). 80‒98. Lambrecht, Knud. 1994. Information Structure and Sentence Form: Topic, Focus, and the Mental Representations of Discourse Referents. Cambridge: Cambridge University Press. Lambrecht, Knud. 2001. A framework for the analysis of cleft constructions. Linguistics 39 (3). 463–516. Miller, Jim. 2006. Focus in the languages of Europe. In Giuliano Bernini & Marcia L. Schwartz (eds.), Pragmatic Organization of Discourse in the Languages of Europe, 121‒214. Berlin & New York: Mouton de Gruyter. Prince, Ellen F. 1978. A comparison of wh-clefts and it-clefts in discourse. Language 54 (4). 883–906. Roggia, Carlo Enrico. 2008. Frasi scisse in italiano e in francese orale: Evidenze dal C-ORAL-ROM. Cuadernos de Filología Italiana 15. 9‒29. Stark, Elisabeth. 2011. La morphosyntaxe dans les SMS suisses francophones: Le marquage de l’accord sujet – verbe conjugué. Linguistik online 48. Stark, Elisabeth. 2012. Negation marking in French text messages. Linguisticæ Investigationes 35 (2). 341‒366. Stark, Elisabeth & Isabelle Riedel. 2013. L’accord du participe passé dans les SMS francophones du corpus SMS suisse. Romanistisches Jahrbuch 63 (1). 116–138. Dürscheid, Christa & Elisabeth Stark. 2013. Anything goes? SMS, phonographisches Schreiben und Morphemkonstanz. In Martin Neef & Carmen Scherer (eds.), Die Schnittstelle von Morphologie und geschriebener Sprache. 189–209. Berlin & New York: Mouton de Gruyter.
344
Elisabeth Stark
Thurlow, Crispin. 2003. Generation Txt? The sociolinguistics of young people’s text-messaging. Discourse Analysis Online 1 (1). 1–27. Van Compernolle, Rémi A. 2008. Morphosyntactic and phonological constraints on negative particle variation in French-language chat discourse. Language Variation and Change 20 (2). 317‒339. Van den Steen, Katleen. 2005. Cleft constructions in French and Spanish. In Nicole Delbecque, Johan van der Auwera & Dirk Geeraerts (eds.), Perspectives on Variation: Sociolinguistic, Historical, Comparative, 275–290. Berlin & New York: Mouton de Gruyter.
Vahram Atayan and Ursula Wienen*
Inferential cleft constructions in translation French c’est que in political texts
1 Introduction Inferential cleft constructions have been the object of a number of linguistic and contrastive studies in recent years.¹ Various functions of these constructions have been discussed in view of both written and oral communication. The present study examines the French inferential c’est que-construction and its functions from a translation-oriented standpoint. Our aim is to verify whether the results of our previous study (Wienen and Atayan 2013), in which we analyzed translations of Ferdinand de Saussure’s Cours de Linguistique générale in German, English, Spanish, Italian, Russian and Armenian, are also applicable to other text types. In this context, we must first consider the fact that the French expression c’est que has direct equivalents in English, the Romance languages and many other languages,² but not in German. There are, of course, similar constructions (Es ist nicht so, dass, ‘It is not so that’, cf. [1f]), which are sometimes considered as equivalents: (1)
(a) Fr.
c’est que je n’ai pas le temps
(b) E.
it’s that I don’t have time
(c) Sp. es que no tengo tiempo (d) It.
è che non ho tempo
(e) G. * es ist, dass ich keine Zeit habe (f) G.
Es ist nicht so, daß wir bei Null angefangen hätten. (Delahunty 2001: 518)
* We would like to thank the editor of the volume and the reviewer Tinka Reichmann for their most valuable remarks and suggestions. 1 E.g., Bearth (1999), Declerck (1992), Delahunty (1995, 2001), Delahunty and Gatzkiewicz (2000), España Villasante (1996), Fernández Leborans (1992), Guil (1994), Pusch (sine anno, 2003). 2 Cf. e.g., Delahunty (1995: 342, 2001: 518). Vahram Atayan, University of Heidelberg Ursula Wienen, Cologne University of Applied Sciences
346
Vahram Atayan and Ursula Wienen
However, even in those cases where such (indirect) equivalents exist, they are not necessarily used in translations since the frequency of inferentials seems to differ considerably in the languages we studied.³ The preliminary results of our previous study concerning the Cours de linguistique générale, the famous academic text based on university lectures held by Saussure (passim), suggest that the translator’s decision to choose a certain option is affected by the languages involved in the translation process. The main options available to the translator include the following: 1) literal translation, i.e., using an Inferential cleft construction in the target language, 2) the use of generic nouns in anaphoric constructions, and 3) constructions with explicit markers of causality (connectives or meta-causal markers). The present study is based on a corpus of articles from Le Monde diplomatique (LMD) and speech transcripts from the European Parliament Proceedings Parallel Corpus (EUROPARL). It includes German, English, Italian, and Spanish as target languages. We will first briefly discuss the forms and functions of the French c’est que-construction as presented in the relevant literature. In a second step, we aim to present the results of our analysis with regard to the use of the construction in the original texts of our corpus. Finally, the options selected in the translated texts will be examined to identify general preferences and specific solutions used by the translators. As we are dealing with a limited number of texts, the main purpose of this article is to present the explorative study we conducted and our resulting hypotheses about Inferential cleft constructions in general, before focusing specifically on their use in political texts and their translation. These results, however, are tentative and require confirmation by further studies.
2 Forms and functions of inferentials⁴ Constructions like the French c’est que and their equivalents in other languages are termed “inferentials” on account of their role in guiding the hearer’s or reader’s interpretation of a text (cf. e.g., Delahunty 1995: 342). Normally, an inferential consists of a copula in the third person singular indicative and a complementizer (e.g., Spanish es que), with an expletive preceding it in the non-pro-drop languages (e.g., French c’est que). In English, the equivalent inferential may be used either with or without a complementizer (it’s that or simply it’s) (cf. Delahunty
3 Cf. Wienen and Atayan (2013). 4 Cf. Wienen and Atayan (2013).
Inferential cleft constructions in translation
347
1995: 343). Although the copula typically appears in the present tense, other tenses are also possible (cf. Delahunty 2001: 523, it was that). Frequently, a negative and a positive form are combined, e.g., It is not that […] it is just that.⁵ In addition, although these formulations are largely grammaticalized (cf. e.g., Berretta 2002: 25, who speaks of a “formula praticamente fissa” for è che), modalizers or argumentative connectors may be inserted (e.g., English it is just that, French c’est peut-être que, c’est surtout que) (cf. Wienen 2006: 133). Inferentials are frequently used to begin an answer to a why-question. They may also be combined with a (preceding) conditional clause (e.g., if […] it’s that; French si […] c’est que) (cf. Declerck 1992: 211).⁶ In line with the arguments brought forward by Declerck (1992),⁷ we exclude forms such as It may be that […] from our study, although they are sometimes considered to be inferentials (cf. e.g., Delahunty 1995: 343). Functional categories like “possibility and reflection” (cf. Delahunty 1995: 347) are thus excluded as well. For a brief discussion on whether inferentials should be considered as clefts, see Wienen and Atayan (2013). Inferentials appear to serve more or less similar purposes in the languages considered here and in other studies (cf. Delahunty 1995: 361–362), although the difficulty of categorizing them is often emphasized (cf. e.g., Delahunty 1995: 354–355; Delahunty and Gatzkiewicz 2000: 307). The following seem to be the most important functions: 1) inferentials are indicators of thematic and logical relevance. They guide the hearer’s or reader’s interpretations and inferences and fulfill discourse repair functions.⁸ 2) Most authors take the view that inferentials indicate a reason or an explanation for a fact, the statement of a fact⁹ or a conse-
5 For negative constructions cf. Delahunty (1995: 356–359); for tandem constructions, e.g., Delahunty (1995: 345). 6 For different constructions in English cf. e.g., Delahunty (1995). Declerck (1992: 206) also deals with inferentials in questions (is it that […]?). 7 Declerck (1992: 207–208) considers this construction to be an “instance of extraposition”: “[…] the it is that-sentence used in (5) [It may be that you have received report of her death from other sources. V.A./U.W.] is not an inferential it is that-sentence of the type we are discussing in this article. Rather, it is an instance of extraposition: It may be that you have received report… is the extraposed version of That you have received report… may be [true]; this appears, amongst other things, from the fact that the next sentence begins with It may also be true that…”. 8 Cf. Delahunty (1995: 341, 2001: 520); cf. also Delahunty and Gatzkiewicz (2000: 301): “Inferentials function as metalinguistic devices which characterize the relevance of the proposition represented by their clause to the processing of an utterance”. They are described by Bearth (1999: 270) as serving a “compensating or corrective function”. 9 Cf. e.g., Bally (sine anno: 182); Declerck (1992: 205, 221); Bearth (1999: 250, 272); Delahunty (1995: 349, 352–353); Berretta (2002: 25), who speaks of a “connettivo causale (meglio: debolmente causale)” for Italian è che).
348
Vahram Atayan and Ursula Wienen
quence, result, or conclusion.¹⁰ 3) Inferential constructions may help to convey focus and contrast functions and thus facilitate the selection of options from a subset of relevant explanations, and the exclusion of irrelevant ones (cf. Bearth 1999: 252, 262).¹¹ 4) Inferentials may also have a modalizing function: Pusch (sine anno) in his research on spoken language argues for a more general description of inferentials and considers them as mainly serving to indicate a higher degree of subjectivity, either restricting the speaker’s responsibility for what he or she is saying (cf. Pusch sine anno: 8) or expressing a higher degree of politeness (cf. Pusch sine anno: 9, 11).
3 Functions of inferential constructions in political texts 3.1 The corpus For the present study, we evaluated 35 French examples extracted from LMD and their translations (between 20 and 25 translations per target language), as well as 25 examples from the transcripts of the speeches of the EUROPARL corpus (21 to 25 translations per target language). The choice of two types of political texts – political commentaries and parliamentary debates – is due to the characteristics of the construction, which is the focus of the present paper. Although c’est que-constructions are manifestations of complex textual and pragmatic relations that are considered typical of an elaborate use of language, they appear to be more frequent in oral communication (cf. Pusch 2003: 296). The two text types chosen seem to be particularly suitable for a combined analysis of these potentially contradictory features: a political commentary in a newspaper, that is a written medium, is normally also considered as a conceptually written text. In contrast to informative genres belonging to mass media communication, however, it is characterized by a higher degree of emotional involvement and a higher occurrence of elements of simulated dialogicity. According to Koch and Oesterreicher (2008: 201), these characteristics tend to belong to the language of proximity. Additionally, newspaper commentaries tend to contain a high per10 Cf. Delahunty (1995: 353). Cf. also Delahunty (2001: 530): “That is, the inferential proposition may be either P or Q, while contextual assumptions play the complementary role. When the inferential proposition plays the antecedent role it is interpreted as a cause, reason, and the like; when it plays the consequent role, it is interpreted as a consequence, result, and similar roles”. 11 Cf. also Delahunty (1995: 354) and Declerck (1992: 213–215).
Inferential cleft constructions in translation
349
centage of rhetorical means and lexical elements expressing an opinion (cf. Klein 2000: 753). C’est que-constructions are therefore, at least potentially, part of the argumentative inventory of this text type. A speech in parliament, however, is delivered orally but prepared beforehand (“written to be read”, according to Halliday’s terminology [passim]). This type of discourse also contains linguistic and rhetorical means which are part of the language of proximity, especially with regard to the emotionalization of communication. According to Klein (2000: 749), speakers in parliament are keen to display a “passionate stance” (“leidenschaftliches Engagement”). The two text types therefore seem to be similar with regard to their position between the language of proximity and the language of distance. However, due to the frequent reference to the situational activity in parliament, there is a crucial difference between the two text types regarding expressions of personal involvement. Personal involvement is conveyed primarily by the use of the first person and the less frequent use of the second person (singular and plural).
3.2 Inferential c’est que-constructions: Contexts of use In our previous study on textual integration and the functions of c’est que-constructions in the French original of the Cours de linguistique générale (Wienen and Atayan 2013), we dealt with, among others, the following points:¹² 1. text-structuring use of c’est que-constructions in the simulated dialogue as an answer to pourquoi-questions, 2. selection of preceding punctuation (full stop, comma, and semicolon occur at nearly the same frequency), 3. possible integration in a conditional marker… c’est que-construction (in approximately one quarter of all instances). It is clear from these results and the corpus used in the present study that there are no instances of c’est que-constructions in phatic contexts.¹³ As far as punctuation is concerned, the construction is always preceded by a comma. These findings may be attributed to the frequency of the conditional marker… c’est queconstruction (si… c’est que) in both text types. This construction favors the use of a comma, making the phatic use impossible. Example (2) shows how the c’est
12 It should be noted that since it is based on university lectures, the Cours is situated between speech and writing as well. 13 The high frequency of such structures in the Cours de linguistique générale is probably due to the explicative nature of the text.
350
Vahram Atayan and Ursula Wienen
que-sequence in an argumentative context functions as an explanation of the preceding utterance. The explanation is marked by a colon: (2)
Le droit de chercher la vérité et de la faire connaître fait partie des droits fondamentaux du citoyen dans une démocratie. L’analyse des motivations comme celle des conséquences ne doit venir que dans un deuxième temps, après qu’on s’est approché, tant que faire se peut, de la vérité. Les individus comme les groupes ont besoin de connaître leur passé: c’est que leur identité en dépend, même si elle ne s’y épuise pas. Frappé par la maladie d’Alzheimer, l’individu sans mémoire perd son identité, cesse d’être lui-même. (LMD)¹⁴ ‘The right to look for truth and to make it available to everyone belongs to the basic civil rights of any democracy. Motivations and consequences play a secondary role and should only be considered once as much of the truth as possible has been discovered. Both individuals and communities need to know about their past because their identity depends on it, although it does not stop there. Someone suffering from Alzheimer’s loses their memory and thus also their identity, they stop being themselves.’ (Translation: V.A./U.W.)
In example (3), a tandem construction (cf. Delahunty 1995: 345) generates a reformulation (Ce n’est pas que – c’est que/It’s not that – it’s that), in which one element of the reciprocal relation is excluded and the other element is validated: (3)
Dans l’ordre traditionnel, il y a toujours la possibilité de rendre, à Dieu, à la nature, ou à quelque instance que ce soit, sous la forme du sacrifice. C’est ce qui assure l’équilibre symbolique des êtres et des choses. Aujourd’hui, nous n’avons plus personne à qui rendre, à qui restituer la dette symbolique – et c’est cela la malédiction de notre culture. Ce n’est pas que le don y soit impossible, c’est que le contre-don y soit impossible, puisque toutes les voies sacrificielles ont été neutralisées et désamorcées (il ne reste plus qu’une parodie de sacrifice, visible dans toutes les formes actuelles de la victimalité). (LMD)
14 Since the cited examples from our corpora (LMD and EUROPARL) can easily be found online with the help of a search engine, no particular websites are indicated.
Inferential cleft constructions in translation
351
‘Within the traditional order it was always possible to repay God, or nature, or another higher authority, by sacrifice. This safeguarded the symbolic equilibrium between human beings and everything else. Today there is no one left to compensate, to whom we might repay our symbolic debt. This is the curse of our culture: although giving is not impossible, giving back is impossible, because sacrifice has had its importance and power taken away, and what remains is a caricature of sacrifice (like contemporary ideas of victimisation).’ (Translation: LMD) It is striking that in inferential structures introduced by a conditional marker, the circumstances expressed in the si-clause are systematically presupposed. Conditional interpretations do not seem plausible in any of the examples. This correlates with the first of the proposed functions of c’est que-constructions, i.e., the indication of general thematic and logical relevance. The clause introduced by si specifies the actual frame of reference in the preceding context and allows for the appropriate inference. At the same time, the validity of the c’est que-clause is not restricted by the conditional context. This possibly fosters the modalizing function, i.e., the reduction of the speaker’s responsibility (which, however, is particularly due to the realization of the actual circumstances in the subordinate clause and the desemanticized main clause¹⁵). This presupposing effect may be strengthened by the integration of the structure as a whole into a subordinate context, as in examples (4)–(6), where subordination is triggered by savoir and fait and the truth-implicating montre: (4)
Cette partie de l’élite n’est pas convaincue par la position officielle selon laquelle il vaut mieux capitaliser sur les concessions librement consenties que de mener des combats perdus d’avance. Elles supportent mal leur perte d’influence sur la politique étrangère, par rapport à ce qui se passait sous le président Boris Eltsine. Dans le même temps, elles savent que si M. Poutine a pu réagir si vite au 11 septembre, c’est qu’il a peu consulté. Le président russe entend avant tout mettre à profit ses relations nouvelles avec les Occidentaux pour progresser dans la voie tracée depuis deux ans. (LMD)
15 In this sense, Inferential cleft constructions resemble presentative structures, in particular (pseudo)relatives (e.g., “Voilà Pierre qui arrive”, cf. Lambrecht 2000). They may trigger a reduction of the responsibility of the sender, especially in argumentative contexts, and therefore also cause an effect of indisputability (cf. Ferrari 2002: 179; Atayan 2006: 473, 478).
352
Vahram Atayan and Ursula Wienen
‘They [Some of Russia’s political elite; V.A./U.W.] remain unconvinced by the official line that it is better to capitalise on concessions freely granted than to embark on a hopeless struggle. They find their loss of influence on foreign policy hard to bear, in contrast with the way things were under Boris Yeltsin. But they are aware that the only reason that Putin was able to react so promptly to 11 September was because he did not consult much. Putin is determined to make the most of his new relationship with the West to advance the policy of the past two years.’ (Translation: LMD) (5)
La manière dont nos gouvernements ont privilégié des positions nationales et l’image, vraie ou illusoire, d’un intérêt national par rapport à l’intérêt européen nous montre que si des crises peuvent être fécondes, c’est que nous avons besoin d’une relance, effectivement, d’une politique européenne étrangère et de sécurité et qu’il faudra la chercher dans un nouveau fédéralisme à inventer pour le siècle qui s’ouvre. (EUROPARL) ‘The way our governments prioritised national positions and the real or illusory idea of national interest over the European interest shows us that, if crises can bear fruit, then what we need is, indeed, to launch a new European foreign and security policy based on a new federalism for the new century.’ (Translation: EUROPARL)
(6)
Surtout, comment ne pas relever l’hypocrisie du rapport qui déplore que “les conditions de vie des réfugiés se font de plus en plus malaisées”, et qui a le cynisme d’en rendre responsables, notamment, “la forte croissance démographique” sans dire un mot de la guerre menée par l’État d’Israël contre le peuple palestinien et sans dénoncer le fait que, si des milliers de Palestiniens sont contraints de vivre dans les camps de réfugiés, c’est que l’État d’Israël, avec le soutien des grandes puissances, y compris d’Europe, s’oppose à ce qu’ils puissent revenir sur les terres dont ils ont été chassés? (EUROPARL) ‘Above all, however, how can we fail to mention the hypocrisy of the report which criticises the fact that ‘refugees’ living conditions are becoming increasingly difficult’ and which has the gall to blame, in particular, the ‘high population growth’ but which does not even mention the war conducted by the State of Israel against the Palestinian nation? The report does not condemn the fact that, although thousands of Palestinians are forced to live in refugee camps, this is because the State of Israel, with the support of the most powerful nations, Europe included, is opposed
Inferential cleft constructions in translation
353
to them being able to return to the land from which they were driven?’ (Translation: EUROPARL)
3.3 Textual functions of c’est que-constructions As mentioned in Section 2, previous studies have revealed cause, reason or explanation, and consequence, result and conclusion as possible textual functions of c’est que-constructions. For our analysis it is important to distinguish two possible relations between the c’est que-construction and the preceding clause: on the one hand, the purely causal implicative relation without a persuasive scope, on the other hand the argumentative relation verbalized with a persuasive intention (cf. Eggs 2000: 397).
3.3.1 Argumentative use in a narrow sense With regard to the argumentative use of inferentials, we have to distinguish between epistemic and deontic argumentation (cf. Eggs 2000: 398). As far as epistemic argumentation is concerned, its conclusive function does not normally allow for si-constructions as the corresponding circumstances are presupposed. Therefore, they cannot appear as a potentially contestable claim which would be necessary for their status of a conclusion. Example (7) shows an instance of epistemic argumentation with the c’est que-sequence in argument function, without the presupposing si-context (our corpus contained only two examples of this type): (7)
L’existence des camps en URSS était connue, mais on la dissimulait sous prétexte qu’il ne fallait pas désespérer Billancourt. Du coup, lorsque la vérité s’est enfin imposée un jour, la classe ouvrière a été encore plus désespérée. Le droit de chercher la vérité et de la faire connaître fait partie des droits fondamentaux du citoyen dans une démocratie. L’analyse des motivations comme celle des conséquences ne doit venir que dans un deuxième temps, après qu’on s’est approché, tant que faire se peut, de la vérité. Les individus comme les groupes ont besoin de connaître leur passé: c’est que leur identité en dépend, même si elle ne s’y épuise pas. Frappé par la maladie d’Alzheimer, l’individu sans mémoire perd son identité, cesse d’être lui-même. (LMD)
354
Vahram Atayan and Ursula Wienen
‘People knew that labour camps existed in the USSR, but they were hushed up under the excuse that it would demoralise the working classes. However, when the truth finally came out, it demoralised the working classes even more. The right to look for truth and to make it available to everyone belongs to the basic civil rights of any democracy. Motivations and consequences play a secondary role and should only be considered once as much of the truth as possible has been discovered. Both individuals and communities need to know about their past because their identity depends on it, although it does not stop there. Someone suffering from Alzheimer’s loses their memory and thus also their identity, they stop being themselves.’ (Translation: V.A./U.W.) Vice versa, the si-clause may serve as an argument for the veracity or plausibility of the circumstances described in the c’est que-clause: (8)
La question des libertés individuelles n’était pas d’actualité pour cet ancien ministre de Nasser jusqu’au jour où il fut jeté en prison sur ordre du président Sadate. “Si un ministre est condamné à dix ans de prison pour avoir exprimé son désaccord avec le président de la République et présenté sa démission, c’est qu’il y a une crise profonde au sein du régime”, se justifie M. Fayek […]. (LMD) ‘The question of individual liberties was a non-issue for Fayek until he was imprisoned under Sadat. “When a minister is sentenced to 10 years in jail after voicing his disagreement with the president and tendering his resignation, then the regime is in crisis right down to its very core”, he says.’ (Translation: LMD)
The instances of deontic argumentation in our corpus can be divided into two subcategories. The first subcategory consists of invitations to act. In these cases, the importance of the action is expressed in the si-clause and explained in the c’est que-clause. Claiming and arguing for the importance of an action corresponds to a deontic type of argumentation, as in examples (9) and (10): (9)
S’il est aujourd’hui important, sinon nécessaire, qu’un certain nombre de chercheurs indépendants s’associent au mouvement social, c’est que nous sommes confrontés à une politique de mondialisation. (LMD)
Inferential cleft constructions in translation
355
‘It is important, if not necessary today that a certain number of independent researchers take part in strikes since we are facing a policy of globalisation.’ (Translation: V.A./U.W.) (10)
S’il est important, Monsieur le Président, deux mois et demi après le naufrage de l’Erika, de revenir sur la marée noire, c’est que l’ampleur de la catastrophe économique, sociale, écologique, qui frappe notre façade Atlantique est en réalité beaucoup plus considérable que les premières évaluations ne l’avaient laissé entendre. (EUROPARL) ‘Mr President, the reason why it is important to return to the subject of the oil slick, two and a half months after the wreck of the Erika, is because the extent of the economic, social and ecological disaster affecting our Atlantic coastline is, in fact, far greater than the first assessments gave us to understand.’ (Translation: EUROPARL)
While example (9) shows a generic invitation to act, the invitation in example (10) is embedded in a specific interactive situation. This is also highlighted by the direct address to the President of the European Parliament. The second category of deontic argumentation comprises justifications of the speaker’s own actions in the past or present. Examples of this subcategory can be found very often in the EUROPARL corpus. The fact that c’est que-clauses often appear in the first person singular or plural in this part of the corpus seems to confirm our classification. In particular, speakers justify their actions in the context of parliamentary procedure, their linguistic actions (such as the expression of a political position) and their voting behavior. Examples (11) and (12) illustrate the first case: in example (11), the speaker justifies the omission of an action. The fact that the circumstances expressed in the si-clause are presented as being outside the sender’s control is an important advantage in this context: (11)
Je crois qu’il s’agira là d’une des questions-clés. Et si je ne peux pas vous donner, dès aujourd’hui, de réponse, c’est que je souhaitais précisément, dans cette note d’orientation, inviter les États membres à prendre position, car il n’y aura pas d’espace européen de la recherche qui ne tienne qu’à la volonté de la Commission. (EUROPARL) ‘I believe that this will be one of the key issues. If I am unable to give you an immediate reply on this today, it is because it was precisely my intention, in this policy paper, to invite the Member States to adopt a position, for there will never be a European Research Area if it stems solely from the will of the Commission.’ (Translation: EUROPARL)
356
Vahram Atayan and Ursula Wienen
In example (12), however, a decision which has already been made is subsequently justified: (12)
Monsieur le Président, Monsieur le Commissaire, chers amis, si nous avons souhaité reporter ce rapport, c’est que nous tenons à la cohérence des travaux de notre Parlement pour peser sur les orientations des politiques européennes. (EUROPARL) ‘Mr President, Commissioner, ladies and gentlemen, the reason we wished to defer this report was that we wanted to ensure consistency in the work of the European Parliament as it influenced the course of European policy.’ (Translation: EUROPARL)
A clear argumentative function can be identified in instances containing speech act verbs, which may be either generic ([13] and [14]) or specific to a debate in parliament (15): (13)
Nous avons donc demandé un financement important pour la vulgarisation des travaux de la Convention, en prévoyant la possibilité pour ceux qui seront contre cette Convention ou ses résultats de faire entendre leur voix - ce qui n’est pas mon cas a priori. Si je parle ainsi, c’est que je pense que ce programme est important. (EUROPARL) ‘We have therefore asked for substantial funding to popularise the work of the Convention, while giving those who are against this Convention or its results the opportunity to make themselves heard – which theoretically speaking is not my case. I am speaking out like this because I believe this programme to be important.’ (Translation: EUROPARL)
(14)
Monsieur le Commissaire, j’entends bien votre réponse, mais d’un autre côté, si j’ai posé cette question, c’est que nous sommes inquiets, justement, étant donné que dans le cadre du rapport sur le troisième programme-cadre de la recherche, figure un projet appelé “International Thermonuclear Experimental Reactor”. (EUROPARL) ‘Commissioner, thank you for your answer. However, I asked this question because we are rightly concerned that the report on the third framework programme for research contains a project entitled ‘International Thermonuclear Experimental Reactor’.’ (Translation: EUROPARL)
Inferential cleft constructions in translation
(15)
357
Cela étant dit, il est clair que l’avenir du système de transport européen passe impérativement par la revitalisation rapide du secteur ferroviaire. Si je plaide en ce sens, c’est que je suis persuadé que le rail a un rôle central à jouer dans la solution des problèmes de mobilité auxquels l’Union européenne est confrontée. (EUROPARL) ‘Having said that, it is clear that the future of the European transport system necessarily involves revitalising the rail sector and in short order. I am speaking in favour of this because I believe that rail has a key role to play in solving the mobility problems facing the European Union.’ (Translation: EUROPARL)
Speakers often justify their voting behavior by presenting their ideas, judgments, opinions and feelings. In these contexts, the use of the first person and lexical items expressing emotion is frequent: (16)
Si j’ai accepté, pour finir, de me joindre à la résolution qui sera soumise au vote de nos collègues, c’est que j’ai pu mesurer l’émotion réelle d’une partie de l’opinion publique mondiale; ces inquiétudes, jointes aux propos de notre collègue au sujet de la réaction du gouvernement japonais, et d’autres organismes internationaux qui invitent les populations à prendre des précautions, à faire des réserves en prévision de cette fin du monde annoncée, me confirment dans l’idée que les précautions préconisées par notre Parlement sont, dans leur sagesse, susceptibles de calmer les peurs et peuvent donc être utiles. (EUROPARL) ‘If, in the end, I have agreed to associate myself to the resolution to be put to the vote in this House, it is because I have been able to gauge the real feeling of some part of public opinion world-wide. These concerns, combined with the statements of our fellow Member on the subject of the reaction of the Japanese Government, and of other international bodies, inviting people to take precautions and to lay in stocks in expectation of the anticipated end of the world, have convinced me that the precautions advocated by this Parliament are, in their wisdom, likely to soothe fears and may therefore prove useful.’ (Translation: EUROPARL)
(17)
Monsieur le Président, si je n’ai pas voté pour le rapport concernant le régime de l’impôt AIEM applicable aux îles Canaries, c’est que je ne veux pas suivre le commissaire Bolkestein et notre rapporteur et me plier à toute la “Spanish connection” mise en place pour commettre un péché
358
Vahram Atayan and Ursula Wienen
mortel contre les principes du marché unique et de la libre circulation des marchandises. (EUROPARL) ‘Mr President, although I did not vote against the report on the arrangements concerning the AIEM tax applicable in the Canary Islands, that was because I did not want simply to follow the example of Commissioner Bolkestein and our rapporteur by accepting the entire ‘Spanish connection’ which has been set up in such a way as to totally violate the principles of the single market and the free movement of goods.’ (Translation: EUROPARL) (18)
Si nous nous sommes abstenus, c’est que les tentatives de dernière minute de dénaturer le rapport en voulant affaiblir la directive “Habitat” et en demander la modification ne nous semblent pas aller dans le sens du respect des directives que nous votons. (EUROPARL) ‘The reason why we abstained, then, is that we do not feel that the last-minute attempts to change the nature of the report, seeking to weaken the Habitats Directive and to urge that it be amended, are in line with respecting the directives on which we are voting.’ (Translation: EUROPARL)
(19)
Si nous sommes opposés au rapport Brok, ce n’est pas comme un signe d’ostracisme à l’encontre de nos partenaires d’Europe centrale ou de l’Est, ou de l’Europe méridionale, bien au contraire, c’est que ce rapport pose des conditions qui sont proprement inadmissibles, s’agissant de pays qui sont européens depuis des siècles et auxquels, mis à part peut-être une condition de respect du pluralisme politique, nous n’avons pas à imposer quelque exigence supplémentaire que ce soit. (EUROPARL) ‘Our opposition to the Brok report is not an indication of ostracism of our partners in Eastern and Central Europe, or southern Europe, quite the contrary, it means that this report sets conditions which are absolutely intolerable, with regard to countries which have been European for centuries and of which we should not be making any other demand whatsoever except, perhaps, for the one condition that political pluralism be respected.’ (Translation: EUROPARL)
Inferential cleft constructions in translation
359
3.3.2 Explicative use Although causal connections are frequent in our corpus, an argumentative connection is not always discernible in the causal contexts. Therefore, a purely explicative interpretation seems to be plausible in some instances. In these cases, the facts and circumstances presupposed in the si-clause are explained by the c’est que-clause: (20)
N’oublions pas non plus le rôle du public, à propos duquel Marilyn Monroe déclare elle-même: “Si je suis une star, c’est que le public a fait de moi une star, pas les studios ni personne, mais le public.” (LMD) ‘Remember that Marilyn Monroe said: “If I am a star, the people made me a star. Not a studio, not a single person, but the people”.’ (Translation: LMD)
(21)
Simpliste, mais “politiquement correcte”, cette analyse fournissait aux décideurs ainsi qu’aux idéologues de Washington et d’ailleurs une couverture utile. Si l’islam passait pour la principale menace, c’est que l’Iran, l’Irak et l’Arabie saoudite produisaient la majorité du pétrole mondial. (LMD) ‘This simple-minded, but politically convenient, analysis provided an extremely useful cover for policy-makers and ideologues in Washington and elsewhere. Islam was seen as the biggest threat because much of the world’s oil comes from Iran, Iraq and Saudi Arabia.’ (Translation: LMD)
(22)
Pour M. Mounir Fakhri Abdennour, un riche homme d’affaires et un des trois membres du Parlement copte, élu député du Wafd aux élections législatives de l’automne 2000, “les coptes doivent s’impliquer dans la vie politique et sociale. Et se faire les avocats du nationalisme égyptien, avant d’être ceux des problèmes coptes”. S’ils sont absents de la vie politique, c’est que le parti au pouvoir les écarte des candidatures. Et puis, les coptes sont réticents à s’engager. (LMD) ‘Munir Fakhri Abdennour, a wealthy businessman and one of three Copts elected to parliament last year, thinks that “Copts need to get involved in the political and social life of the country. And to advocate Egyptian nationalism before Coptic issues.” If Copts are absent from political life, it is not just because the ruling National Democratic Party passes them
360
Vahram Atayan and Ursula Wienen
over as candidates. Copts are reluctant to go into politics.’ (Translation: LMD) (23)
En réalité, la droite et l’extrême droite forment un front du refus qui recourt à la provocation, à l’intimidation et au combat parce qu’ils craignent la paix. Et s’ils s’inquiètent, c’est que, désormais, de plus en plus d’Israéliens peuvent comprendre que, depuis trois ans, ils ont été abusés. (LMD) ‘The right and far-right bluntly refuse to discuss the matter and resort to provocation, intimidation and confrontation because they fear peace. They are worried because more and more Israelis realise that the government has been misleading them for more than three years.’ (Translation: LMD)
(24)
Mais tout expliquer par l’argent ne fait pas sens. Si le néofondamentalisme progresse, c’est qu’il répond à une demande sur le “marché religieux”. (LMD) ‘But it is nonsense to claim that Saudi money explains everything. Fundamentalism is spreading because it satisfies a particular demand in the religious market.’ (Translation: LMD)
However, in many of the instances where a causal explicative interpretation seems suitable, a secondary argumentative intention appears to be present. The explanation given in the c’est que-clause is realized by lexical items with affective and connotative meanings which typically convey a critical attitude towards a topic of the discussion which remains thematically salient as the text continues. Hence, we are dealing with a case of indirect argumentation involving the use of a pseudo-explanation. The following examples illustrate this: (25)
Si ce potentiel n’est pas pris en compte à son maximum, c’est que certaines énergies sont délaissées par les décideurs, souvent dénués de toute culture environnementale, et que d’autres sont favorisées par le biais d’alliances entre groupes. (LMD) ‘If this potential is not always fully used, it is because the decision-makers, who are often out of touch with environmental concerns, simply do not consider some energy sources, while they might favor others because of the existing connections between the large companies concerned.’ (Translation: V.A./U.W.)
Inferential cleft constructions in translation
(26)
361
Si un grand nombre de compagnies américaines (et désormais européennes) embauchent à l’aide de variantes de l’“inventaire multiphasique de la personnalité” comportant près de six cents questions, c’est que, pour elles, la vérité doit l’emporter sur le respect de l’honneur et la sauvegarde de la vie privée. (LMD) ‘The reason why a large number of American (and now European) companies use variants of the nearly 600-question “multiphase personality inventory” in their recruitment procedures is that they put truth ahead of respect for honour and private life.’ (Translation: LMD)
(27)
Rien n’a été prévu alors que, pourtant, tout était prévisible et de longue date: dès la construction des installations nucléaires chacun savait qu’il faudrait bien, un jour, les démanteler et que cela ne serait pas simple. Si donc rien n’a été prévu c’est que l’on ne voulait rien prévoir pour ne pas risquer d’affaiblir le dogme pronucléaire. (EUROPARL) ‘There were no provisions, even though all this was anticipated long ago. We have been aware, ever since the nuclear installations were built, that one day they would have to be decommissioned and that this would not be a simple matter. So the fact that there are no provisions is indicative of the fact that no one wanted to make provisions for this so as not to undermine the pro-nuclear dogma.’ (Translation: EUROPARL)
(28)
Monsieur le Président, Monsieur le Commissaire, si les voix de ces deux femmes, Safiya Hussaini et Hafsatu Abubakar, sont montées de leur lointaine province nigériane du Sokoto et du Katsina jusqu’à notre hémicycle de Strasbourg aujourd’hui, c’est que la situation de ces deux femmes avait atteint le paroxysme de l’inacceptable. (EUROPARL) ‘Mr President, Commissioner, the only reason the voices of these two women, Safiya Hussaini and Hafsatu Abubakar, have reached this Chamber today from the distant regions of Sokoto and Katsina in Nigeria is because their situation is the very epitome of all that is unacceptable.’ (Translation: EUROPARL)
(29)
Si l’Union européenne concentre, au fond, maintenant, toute la contestation contre la mondialisation ultralibérale, c’est que tout le monde a bien compris qu’elle en était devenue le principal chausse-pied, y compris à Doha, au dernier sommet de l’OMC. (EUROPARL)
362
Vahram Atayan and Ursula Wienen
‘The reason why the European Union is now wholly opposed to ultra-liberal globalisation is because everyone has realised that it has become globalisation’s main shoehorn, including at the last WTO summit at Doha.’ (Translation: EUROPARL) Cases in which the argumentative force of the circumstances described in the si-clause is reduced in the c’est que-clause represent another type of argumentative use. Previous studies have repeatedly shown that the argumentative force of an utterance can be reduced by presenting a state of affairs that causes the circumstances in the corresponding utterance (Ducrot et al. 1980: 166; Anscombre 1996: 255; García Negroni 2003: 20–21; Atayan 2006: 327–328, 335–338.). This mechanism is shown in example (30): (30)
Thierry de Saulieu, rédacteur en chef du mensuel Les Routiers, l’avait souligné en juillet 1998: “Si RVI [Renault Véhicules industriels] a dénigré le GPL [gaz de pétrole liquéfié] et a tout misé sur le gaz naturel, c’est que ses études sur les carburants propres étaient, pour l’essentiel, financées par Gaz de France (5)” Dans la pratique, un bus au GPL n’entraîne qu’un surcoût de 100 000 francs par rapport à un bus Diesel classique. (LMD) ‘Thierry de Saulieu, editor-in-chief of Les Routiers, a monthly magazine, underlined this fact in July 1998: “RVI [Renault Véhicules industriels] maligned LPG [liquified petroleum gas] and entirely focused on natural gas because its studies on clean fuels were essentially financed by Gaz du France”. In reality, a bus that runs on LPG only costs 100 000 francs more to operate than a normal diesel bus.’ (Translation: V.A./U.W.)
The critical attitude of RVI towards liquid gases is relativized by the fact that the relevant studies were financed by Gaz de France and therefore influenced by certain interests (rather than objective knowledge). Viewed from a polyphonic perspective, the reduction of argumentative force in example (31) is attributed to Pakistani public opinion. The alleged “misbehavior” of married women is presented as the reason for their assassination and thus serves as an excuse for it. The si… c’est que-construction allows for a compact summary of this ironic argumentation: (31)
Sathya, qui travaille aussi pour Vimochana, cite le cas de ce père exaspéré qui disait à sa fille: “Tu nous fais honte, retourne chez ton mari, c’est ton devoir. Et s’il te tue, tant pis.” Ainsi, malgré les souffrances évidentes et les risques mortels, père et mère renvoient généralement leur fille chez
Inferential cleft constructions in translation
363
son époux. “L’essentiel, c’est le respect de la tradition et de l’ordre social, ajoute Sathya. Dans ce système, la femme compte peu et l’homme est rarement critiqué. S’il tue sa femme, beaucoup chercheront à l’excuser. Ils diront qu’elle était trop sensible ou trop coquette. Bref, si c’est arrivé, c’est que quelque chose n’allait pas chez elle.” (LMD) ‘Sathya, another Vimochana worker, quotes an exasperated father who told his daughter she was disgracing her family. “Go back to your husband,” he shouted. “If he kills you, too bad.” “The main concern is respect for tradition and the social order,” Sathya explains. “Women count for little and men are rarely criticised. If anything happens, the husband will say his wife was over-sensitive or flirtatious. Whatever happens is her fault.”’ (Translation: LMD) A similar technique is employed in example (32), where false predictions about the advantages of globalization are polyphonically attributed to the enemies of the United States so that the positive attitude towards globalization cannot be questioned: (32)
C’est là qu’entrent en scène Huntington et son “choc des civilisations”: si la mondialisation n’a pas produit les fruits attendus, c’est qu’elle est bridée et brimée par ceux qui, ennemis criminels de l’Amérique, veulent empêcher le reste du monde de progresser vers les douceurs civilisatrices de l’American Way of Life. (LMD) ‘This was where Huntington and his clash of civilisations came in: if globalisation had failed to deliver its expected benefits, this was because the campaign for it had been hindered and assailed by the enemies of the US, who sought to prevent the rest of the world from embracing the American way of life and its pleasures.’ (Translation: LMD)
The possibility of reducing the argumentative force in a si… c’est que-construction also seems to have some advantages insofar as the target of such argumentative attenuation (critical attitude of RVI, assassination of the wife, lack of benefits of globalization), is verbalized by means of a subordinate conditional clause and thus remains in the textual background. Summing up, we can say that deontic justification and indirect argumentative use are especially frequent in the EUROPARL corpus, while explicative function in combination with indirect argumentative use is dominant in the LMD corpus. This may be attributed to the fact that the EUROPARL corpus contains debates on parliamentary activity. In general terms, the use of c’est que in LMD, on the one
364
Vahram Atayan and Ursula Wienen
hand, reflects the general informative function of the articles in this newspaper (which aims to offer a broad view on global politics rather than focusing solely on European and Western issues). On the other hand, the indirect argumentative use correlates with the clear left-wing orientation of the newspaper. The indirect argumentative function shown in the texts of the EUROPARL corpus is due to the confrontation of political positions taking place in parliamentary debate, while most politicians’ efforts to project a positive image entail a deontic function of justification.
4 Inferential clefts in translation As already mentioned above (cf. Section 1), the main options for translation as presented in Wienen and Atayan (2013) are as follows: 1) literal translation, i.e., translation by an Inferential cleft construction in the target language, 2) translation by a generic noun + copula + conjunction, and 3) constructions with explicit markers of causality (connectives or meta-causal markers). In a wider sense, the second option can be regarded as a type of literal translation as well, as in this case a generic and semantically weak noun – e.g., il fatto (‘the fact’) in il fatto è che in Italian – assumes the function of the expletive in an otherwise parallel syntactic structure, while the third one produces a more explicit realization of the logical relation marked by c’est que. In Wienen and Atayan (2013) we pointed out that literal translation in the narrow sense (option 1) was dominant in Spanish translations, while translations using a generic noun (option 2) were particularly prevalent in Italian (il fatto è che ‘the fact is that’), Armenian and Russian. Causal metaterms (option 3), e.g. Grund (‘reason’) in German, appeared very frequently in German, Armenian and Russian. Finally, causal connectors were frequent in English, but could also be found in the other languages except Armenian. However, as to the four languages considered here, the picture is considerably different if we look at the data from our corpus:
Inferential cleft constructions in translation
365
Table 1: The distribution of the translation options Type 1+2
Type 3 (some examples contain more than one marker of different subtypes)
literal, generic, others
causal connectors
causal metaterms
conditional context
22 (weil; deshalb, weil; da; denn), among which
13, among which
13 wenn + dann
18 (deshalb), weil
5 liegt das daran, dass
9 wenn + dann / so (deshalb), weil
4 with wenn+dann
31 (because; if + 0; when + then), among which
15, among which
29 because
7 in combination with because
11 if + because
7 the (only) reason (why)… is / was that
GER (40) 5
EN (45) 6
IT (44)
3
27 (perché; non che… ma è), 14, among which among which 26 perché
7 ciò è dovuto al / determinato dal fatto che
20 se + è perché
3 il motivo (per cui… è che / risiede)
11 if
27 se
7 with se ES (38) 8 (4 of which 26 (porque; no es que… sino 4, among which es que, 1 que lo que es; si + 0), among with si) which 24 porque 21 si + es porque
24 si
3 la razón de que / por la que … es que
Table 1 shows the following: 1. There are only very few cases of literal translation in a narrow or wider sense. Only four of the Spanish translation examples fall into this category. It is striking that one of these instances in the written examples of LMD happens to be a proverb, while another is used in a quotation. As orality plays an important role in both instances, translation by es que can probably be regarded as
366
Vahram Atayan and Ursula Wienen
an effort to preserve the authenticity of spoken language (examples [33] and [34] = [31]): (33)
A moins que la réussite de quelques combats féministes inquiète ? J’aime beaucoup ce proverbe africain: “Quand on commence à lancer des pierres sur un arbre, c’est qu’il est en train de porter ses fruits.” (LMD) A menos que el éxito de algunos combates feministas resulte inquietante. Me gusta mucho este proverbio africano: “Cuando se comienza a lanzar piedras contra un árbol, es que está por dar sus frutos”. (Translation: LMD) ‘Unless the success of some feminist movements was to give rise to concern? There is an African proverb which I love: “It is only the tree loaded with fruit that the people throw stones at.”’ (Translation: V.A./U.W.)
(34)
“L’essentiel, c’est le respect de la tradition et de l’ordre social, ajoute Sathya. Dans ce système, la femme compte peu et l’homme est rarement critiqué. S’il tue sa femme, beaucoup chercheront à l’excuser. Ils diront qu’elle était trop sensible ou trop coquette. Bref, si c’est arrivé, c’est que quelque chose n’allait pas chez elle.” (LMD) “Lo esencial es el respeto a la tradición y el orden social, añade Sathya. En este sistema, la mujer cuenta poco y el hombre es criticado raramente. Si mata a su mujer, muchos intentarán excusarle. Dirán que era demasiado sensible y demasiado coqueta. En resumen, si ocurre, es que algo no funcionaba en su casa.” (Translation: LMD) ‘“The main concern is respect for tradition and the social order,” Sathya explains. “Women count for little and men are rarely criticised. If anything happens, the husband will say his wife was over-sensitive or flirtatious. Whatever happens is her fault.”’ (Translation: LMD)
2.
3.
No instance of the pattern generic noun + copula + conjunction appears in the corpus. Generic nouns are only used in Italian causal metaterms (ciò è dovuto al / determinato dal fatto che). Explicitation¹⁶ as a result of causality marking is the most frequent translation option in our corpus. Depending on the languages involved, these cases account for between 75% and 90% of the solutions. Most of the time, the standard marker of causality (weil, because, perché, porque) is used. This
16 Cf. e.g., the notion of explicitation in translation in Atayan and Kusztor (2010).
Inferential cleft constructions in translation
367
means that the causal inferential basis instruction of the Cleft construction is realized by an explicit grammaticalized marker. Particularly in translations into one of the Romance languages, the combination with the presupposed conditional context¹⁷ permits the whole range of interpretations discussed. An explicative interpretation is backed up by the presupposed status of the preceding clause, while an argumentative interpretation is supported by the focusing instruction of the Cleft construction. In examples (35) (= [16]) and (36) (= [20]), this option is justified deontically: (35)
Si j’ai accepté, pour finir, de me joindre à la résolution qui sera soumise au vote de nos collègues, c’est que j’ai pu mesurer l’émotion réelle d’une partie de l’opinion publique mondiale… (EUROPARL) Se ho accettato, alla fine, di dare il mio contributo alla risoluzione che sarà votata oggi dai nostri colleghi, è stato perché ho avuto modo di rendermi conto del reale coinvolgimento di parte dell’opinione pubblica mondiale… (Translation: EUROPARL) Si he aceptado, finalmente, adherirme a la resolución que se someterá a votación de nuestros colegas, es porque he podido percatarme del verdadero estado emocional de una parte de la opinión pública mundial… (Translation: EUROPARL) ‘If, in the end, I have agreed to associate myself to the resolution to be put to the vote in this House, it is because I have been able to gauge the real feeling of some part of public opinion world-wide.’ (Translation: EUROPARL)
In example (36), a purely explicative use is shown: (36)
“Si je suis une star, c’est que le public a fait de moi une star, pas les studios ni personne, mais le public.” (LMD) “Se sono una star, è perché il pubblico ha fatto di me una star, né gli studios né altri, solo il pubblico.” (Translation: LMD) “Si soy una estrella es porque el público ha hecho de mí una estrella, no los estudios, sino el público.” (Translation: LMD)
17 20/26 for perché and 21/24 for porque. In the translations into English, because is the most frequent solution, while in half of the translations into German (deshalb), weil is used.
368
Vahram Atayan and Ursula Wienen
‘“If I am a star, the people made me a star. Not a studio, not a single person, but the people”’ (Translation: LMD) The indirect argumentative use of the Cleft construction may also be marked by a causal conjunction, as shown for German in example (37) (= [26]): (37)
Si un grand nombre de compagnies américaines (et désormais européennes) embauchent à l’aide de variantes de l’“inventaire multiphasique de la personnalité” comportant près de six cents questions, c’est que, pour elles, la vérité doit l’emporter sur le respect de l’honneur et la sauvegarde de la vie privée. (LMD) Wenn eine Vielzahl von amerikanischen (…) Unternehmen ihr Personal mit Hilfe des fast 600 Fragen umfassenden MMPI-Persönlichkeitstests (…) auswählt, dann deshalb, weil die Wahrheit ihrer Ansicht nach wichtiger ist als die Achtung der Menschenwürde und der Schutz der Privatsphäre. (Translation: LMD) ‘The reason why a large number of American (and now European) companies use variants of the nearly 600-question “multiphase personality inventory” in their recruitment procedures is that they put truth ahead of respect for honour and private life.’ (Translation: LMD)
The reduction of argumentative force is also compatible with a translation using causal connectors (example [38] = [32]): (38)
[…] si la mondialisation n’a pas produit les fruits attendus, c’est qu’elle est bridée et brimée par ceux qui, ennemis criminels de l’Amérique, veulent empêcher le reste du monde de progresser vers les douceurs civilisatrices de l’American Way of Life. (LMD) […] se la globalizzazione non ha prodotto i frutti attesi, è perché continua ad essere imbrigliata e ostacolata da coloro che, nemici mortali dell’America, vogliono impedire al resto del mondo di avanzare verso le delizie civilizzatrici dell’american way of life. (Translation: LMD) […] si la mundialización no produjo los frutos esperados, es porque quienes la refrenan y vejan, enemigos criminales de Estados Unidos, quieren impedir que el resto del mundo progrese hacia las bondades civilizadoras del American Way of Life. (Translation: LMD)
Inferential cleft constructions in translation
369
[…] if globalisation had failed to deliver its expected benefits, this was because the campaign for it had been hindered and assailed by the enemies of the US, who sought to prevent the rest of the world from embracing the American way of life and its pleasures. (Translation: LMD) It is worth noting in this context that the majority of translations in this category are realized by means of a conditional marker… copula causal conjunction-construction (cf. [35], [36], [38]). Additionally, the focusing effect of Inferential cleft constructions is supported by the use of focusing and intensifying adverbs¹⁸ in some of the Italian and German instances and/or by the correlative construction deshalb, weil in German: (39)
a.
E se gli stati esercitano un controllo inferiore a quello che vorrebbero, è proprio perché l’immigrazione risponde ad altre dinamiche. (LMD) ‘If some countries still complain about a lack of control, then only because migration flows are influenced by different dynamics.’ (Translation: V.A./U.W.)
b. Al tempo stesso sanno che Putin ha potuto reagire così rapidamente all’11 settembre proprio perché non ha perso molto tempo in consultazioni. (LMD) ‘But they are aware that the only reason that Putin was able to react so promptly to 11 September was because he did not consult much.’ (Translation: LMD) c.
Se lo credetti per Sartre, è semplicemente perché non potevo dimenticare la sua posizione sull’Algeria… (LMD) ‘There was no doubt in my mind that Sartre would be worth the effort, simply because of his position on Algeria…’ (Translation: V.A./U.W.)
d. D’altra parte, se raccomando che venga mantenuta una differenza tra il premio di base per il settore “carne” e quello per il settore lattierocaseario, è semplicemente perché la decisione opposta starebbe a significare che l’economia lattiera ovina e caprina è priva di interesse, ma non credo che qualcuno osi ancora pensarlo. (EUROPARL) ‘Furthermore, I recommend maintaining a difference between the basic premium for the ‘meat’ sector and the premium for the milk sector because to do otherwise would leave the sheep and goat milk
18 For the correlation between the focusing and the intensifying function of such adverbs cf. De Cesare (2002).
370
Vahram Atayan and Ursula Wienen
sectors with no future, which nobody wants to see.’ (Translation: EUROPARL) e.
Signor Commissario, ho sentito la sua risposta, ma d’altro canto se ho presentato un’interrogazione è appunto perché siamo preoccupati, visto che nella relazione sul Terzo programma quadro di ricerca figura un progetto intitolato “International Thermonuclear Experimental Reactor”. (EUROPARL) ‘Commissioner, thank you for your answer. However, I asked this question because we are rightly concerned that the report on the third framework programme for research contains a project entitled ‘International Thermonuclear Experimental Reactor’.’ (Translation: EUROPARL)
f.
Ich sage dies einfach, weil ich dieses Programm für sehr wichtig halte. (EUROPARL) ‘I am speaking out like this because I believe this programme to be important.’ (Translation: EUROPARL)
g.
Wenn wir diese Frage gestellt haben, dann deshalb, weil das am Eurotunnel aufgetretene Problem, da es zwei große europäische Länder und auch unsere Sicherheits- und Asylpolitik betrifft… (EUROPARL) ‘The reason why we tabled this question is because the public sees the problem at the Eurotunnel site – since it involves two large European countries and also since it concerns our security and asylum policy – as requiring intervention by the European Union.’ (Translation: EUROPARL)
h. Wenn die Staaten dennoch ein Kontrolldefizit beklagen, so deshalb, weil Migrationsbewegungen einer anderen Dynamik folgen. (LMD) ‘If some countries still complain about a lack of control, then only because migration flows are influenced by different dynamics.’ (Translation: V.A./U.W.) Causal metaterms are also used for the translation of the more frequent categories: pure explanation (above all in the LMD corpus), deontic justification and indirect argumentation by means of a pseudo-explanation (this applies specifically to the EUROPARL corpus). These three categories are illustrated in examples (40) and (41) = (19):
Inferential cleft constructions in translation
(40)
371
Si l’ancien opposant éprouve tant de difficultés à réviser l’arsenal juridique hérité de la guerre froide, c’est que celle-ci continue d’être une réalité pour un grand nombre de Sud-Coréens. (LMD, explanation) Die Schwierigkeiten, auf die der ehemalige Oppositionelle stößt, wenn er das rechtliche Instrumentarium aus den Zeiten des Kalten Krieges reformieren will, haben damit zu tun, dass für viele Südkoreaner der Kalte Krieg heute noch Realität ist… (Translation: LMD) Se l’ex oppositore ha tante difficoltà a rivedere l’arsenale giuridico ereditato dalla guerra fredda, ciò è dovuto al fatto che questa continua ad essere una realtà per un gran numero di sud-coreani. (Translation: LMD) ‘If the former dissident is experiencing such difficulties in trying to modify the juridical arsenal inherited from the cold war, it is because for many South Koreans that war is still a present reality.’ (Translation: LMD)
(41)
Si nous sommes opposés au rapport Brok, ce n’est pas comme un signe d’ostracisme à l’encontre de nos partenaires d’Europe centrale ou de l’Est, ou de l’Europe méridionale, bien au contraire, c’est que ce rapport pose des conditions qui sont proprement inadmissibles… (EUROPARL, deontic justification) Wenn wir gegen den Bericht Brok sind, dann ist das kein Zeichen von Ablehnung gegenüber unseren mittel-, ost- oder südeuropäischen Partnern, sondern dies liegt im Gegenteil daran, dass dieser Bericht völlig unannehmbare Bedingungen gegenüber Ländern stellt… (Translation: EUROPARL) Our opposition to the Brok report is not an indication of ostracism of our partners in Eastern and Central Europe, or southern Europe, quite the contrary, it means that this report sets conditions which are absolutely intolerable… (Translation: EUROPARL) La nostra opposizione alla relazione dell’onorevole Brok non è un segno di ostracismo verso i nostri partners dell’Europa centrale e orientale o dell’Europa meridionale, al contrario, è dovuta al fatto che la relazione impone condizioni inaccettabili… (Translation: EUROPARL)
(42)
En fait, si ces directives ne sont que partiellement appliquées, c’est que les États membres n’en avaient pas mesuré les conséquences lors de leur adoption. (EUROPARL, indirect argumentation)
372
Vahram Atayan and Ursula Wienen
Wenn diese Richtlinien nämlich nur teilweise angewendet werden, so liegt dies daran, dass die Mitgliedstaaten zum Zeitpunkt ihrer Verabschiedung die entsprechenden Auswirkungen nicht genau eingeschätzt haben. (Translation: EUROPARL) Di fatto, se le direttive in questione vengono applicate solo in parte, il motivo è che quando erano state adottate gli Stati membri non ne avevano valutato le conseguenze. (Translation: EUROPARL) ‘The fact is, if the directives are only partially implemented, it is because the Member States had not weighed the consequences when adopting them.’ (Translation: EUROPARL) Finally, in the English, Italian and Spanish translations, the pattern The reason why… is that is striking. It is used to translate Si […] c’est que-constructions and enables both the realization of the cataphoric instruction of the conditional and the explicitation of the general causal semantics of c’est que. (43)
Si un grand nombre de compagnies américaines (et désormais européennes) embauchent à l’aide de variantes de l’“inventaire multiphasique de la personnalité” comportant près de six cents questions, c’est que, pour elles, la vérité doit l’emporter sur le respect de l’honneur et la sauvegarde de la vie privée. (LMD) The reason why a large number of American (and now European) companies use variants of the nearly 600-question “multiphase personality inventory” in their recruitment procedures is that they put truth ahead of respect for honour and private life. (Translation: LMD)
(44)
Mais si le Parlement européen a proposé l’interdiction des pavillons de complaisance, c’est que chacun ici sait bien […] que le pavillon de complaisance a cette particularité que, en cas de problème, il y a d’énormes difficultés pour situer les responsables. (EUROPARL) Il motivo per cui il Parlamento ha proposto il divieto delle bandiere di comodo risiede, tuttavia, nelle enormi difficoltà di identificazione dei responsabili in caso di problemi, come tutti abbiamo avuto modo di constatare nel corso delle discussioni a seguito delle catastrofi marittime. (Translation: EUROPARL) ‘[…] but the reason why Parliament has proposed that flags of convenience be banned is that we are all well aware, in the debates that we have
Inferential cleft constructions in translation
373
had following maritime disasters, that the nature of flags of convenience is that in the event of problems, it is incredibly difficult to locate those responsible.’ (Translation: EUROPARL)
5 Conclusion Our observations concerning the two corpora (LMD and EUROPARL) allow some preliminary conclusions to be drawn regarding the functionality of Inferential cleft constructions and potential translation options. We can distinguish between three main functions of the construction in the text types analyzed here, which serve both explanatory and argumentative purposes. Epistemic use and deontic argumentation concerning the necessity and justification of an action mentioned in the preceding clause are clearly argumentative. Indirect argumentative structures, in which a causal connection between the preceding clause and the c’est que-clause is created, are situated between explanation and argumentation. In these cases, however, the c’est que-clause contains clearly emotional or evaluative components, which remain relevant to the argumentation throughout the whole context. Whenever these elements are missing, the purely explicative function of the Inferential cleft construction becomes apparent. Due to the contextual conditions of both text types, it appears that direct and indirect argumentative structures are particularly frequent in the EUROPARL corpus, while explicative and indirect argumentative structures are dominant in LMD. As far as translation is concerned, the present study shows that – in contrast to our previous analyses – the explicitation of the causal inferential instruction of the marker is applied almost systematically. In the Romance languages and English, the standard causal conjunctions (perché, porque, because) are used in a large majority of cases. These causal metaterms can enable a compact codification of the cataphoric instructions of the introducing conditional and the causal information in the Cleft structure. Finally, we observed that a combination of causal conjunctions with causal metaterms or focusing adverbs is also possible. The main difficulty in studying Inferential cleft constructions is the fact that they are very rare in written texts, particularly in those which are subject to translation. For this reason, the results are provisional and require further research and testing on a larger corpus.
374
Vahram Atayan and Ursula Wienen
References Anscombre, Jean-Claude. 1996. L’opposition surtout / particulièrement et la structuration discursive. In Claude Muller (ed.), Dépendance et intégration syntaxique. Subordination, coordination, connexion, 245–256. Tübingen: Niemeyer. Atayan, Vahram. 2006. Makrostrukturen der Argumentation im Deutschen, Französischen und Italienischen. Bern: Peter Lang. Atayan, Vahram & Mónika Kusztor. 2010. Come esplicitare l’esplicitazione? Qualche considerazione sull’ipotesi dell’esplicitazione nella teoria della traduzione. In Maria Iliescu, Heidi M. Siller-Runggaldier & Paul Danler (eds.), Actes du XXVe Congrès International de Linguistique et de Philologie Romanes (Innsbruck, 3–8 Septembre 2007), vol. 1, 519–528. Berlin & New York: Mouton de Gruyter. Bally, Charles. (sine anno). Traité de stylistique française. Paris: Librairie C. Klincksieck. http:// www.scribd.com/doc/54111722/traitdestylist02balluoft (accessed February 2012). Bearth, Thomas. 1999. The inferential gap condition. Pragmatics 9 (2). 249–280. Berretta, Monica. 2002. Quello che voglio dire è che: le scisse da strutture topicalizzanti a connettivi testuali. In Gian Luigi Beccaria & Carla Marello (eds.), La parola al testo. Scritti per Bice Mortara Garavelli, vol. 1, 15–31. Torino: Edizioni dell’Orso. De Cesare, Anna-Maria. 2002. Intensification, modalisation et focalisation. Les différents effets des adverbes proprio, davvero et veramente. Bern: Peter Lang. Declerck, Renaat. 1992. The inferential it is that-construction and its congeners. Lingua 87. 203–230. Delahunty, Gerald P. 1995. The inferential construction. Pragmatics 5 (3). 341–364. Delahunty, Gerald P. 2001. Discourse functions of inferential sentences. Linguistics 39 (3). 517–545. Delahunty, Gerald & Laura Gatzkiewicz. 2000. On the Spanish inferential construction ser que. Pragmatics 10 (3). 301–322. Ducrot, Oswald et al. 1980. Les mots du discours. Paris: Editions de Minuit. Eggs, Ekkehard. 2000. Vertextungsmuster Argumentation: Logische Grundlagen. In Klaus Brinker, Gerd Antos, Wolfgang Heinemann & Sven F. Sager (eds.), Text- und Gesprächslinguistik. Ein internationales Handbuch zeitgenössischer Forschung, vol. 1, 397–414. Berlin & New York: Mouton de Gruyter. España Villasante, Margarita. 1996. Aspectos semántico-pragmáticos de la construcción es que en español. Dicenda. Cuadernos de filología hispánica 14. 129–148. Koehn, Philipp. 2005. Europarl: A Parallel Corpus for Statistical Machine Translation. MT Summit 2005, http://www.iccs.inf.ed.ac.uk/~pkoehn/publications/europarl-mtsummit05. pdf, http://opus.lingfil.uu.se/Europarl3.php (EUROPARL) Fernández Leborans, María Jesús. 1992. La oración del tipo: es que. Verba 19. 223–239. Ferrari, Angela. 2002. Valore intrinseco e funzioni testuali della frase nominale. In Hanne Jansen, Paola Polito, Lene Schøsler & Erling Strudsholm (eds.), L’infinito & oltre. Omaggio a Gunver Skytte, 171–189. Odense: Odense University Press. García Negroni, María Marta. 2003. Gradualité et réinterprétation. Paris: L’Harmattan. Guil, Pura. 1994. Es que… in italiano. In Anna Giacalone Ramat & Massimo Vedovelli (eds.), Italiano: lingua seconda, lingua straniera. Atti del XXVI congresso della Società di Linguistica Italiana (SLI 34), 111–126. Roma: Bulzoni.
Inferential cleft constructions in translation
375
Klein, Josef. 2000. Textsorten im Bereich politischer Institutionen. In Klaus Brinker, Gerd Antos, Wolfgang Heinemann & Sven F. Sager (eds.), Text- und Gesprächslinguistik. Ein internationales Handbuch zeitgenössischer Forschung, vol. 1, 732–755. Berlin & New York: Mouton de Gruyter. Koch, Peter & Wulf Oesterreicher. 2008. Mündlichkeit und Schriftlichkeit von Texten. In Nina Janich (ed.), Textlinguistik. 15 Einführungen, 199–216. Tübingen: Narr. Lambrecht, Knud. 2000. Prédication seconde et structure informationnelle: la relative de perception comme construction présentative. Langue française 127. 49–66. Le Monde diplomatique. Archives (2011) (DVD-Rom). Pusch, Claus D. (sine anno). Marqueurs discursifs et subordination syntaxique: La construction inférentielle en français et dans d’autres langues romanes. http://latina.phil2.uni-freiburg. de/pusch/Download/construction_inferentielle.pdf (accessed February 2012). Pusch, Claus D. 2003. Die es que / c’est que-Konstruktion und ihre kommunikativen Dimensionen. In Gudrun Held (ed.), Partikeln und Höflichkeit (Cross Cultural Communication 10), 295–317. Bern: Peter Lang. http://www.romanistik.uni-freiburg.de/ pusch/Download/publikationen/2003_Es-que-Konstruktion.pdf (accessed May 2013). Wienen, Ursula. 2006. Zur Übersetzbarkeit markierter Kohäsionsformen: Eine funktionale Studie zum Kontinuum von Spaltadverbialen und Spaltkonnektoren im Spanischen, Französischen und Deutschen. Bern: Peter Lang. Wienen, Ursula & Vahram Atayan. 2013. Kohäsion im Fachtext: die inferentielle c’est que-Konstruktion in Übersetzungen von Ferdinand de Saussures Cours de linguistique générale. In Monika Krein-Kühle, Ursula Wienen & Ralph Krüger (eds.), Kölner Konferenz zur Fachtextübersetzung 2010, 253–268. Bern: Peter Lang.
Volker Gast and Natalia Levshina*
Motivating w(h)-Clefts in English and German: A hypothesis-driven parallel corpus study 1 Introduction This contribution is concerned with the syntactic configuration illustrated in (1), which is often subsumed under the terms “pseudo-Cleft”¹ or “wh-Cleft” in the English literature, in contradistinction to (genuine) Clefts or it-Clefts² as in (2) (note that the term “Cleft sentence” was introduced by Jespersen 1937: 83–89, according to Fischer 2009: 169). (1)
Pseudo-Cleft/wh-Cleft³ What I miss in Mr Martens’ report are the guidelines and vision[s] of how a new cooperation could be formed. (EPEG-6/Da 1451517)
* The paper has benefited greatly from comments made by various colleagues, especially the participants of the workshop on Cleft constructions organized by Anna-Maria De Cesare in Basel (June 4, 2012). Moreover, we would like to thank an external reviewer for a very thoughtful review and valuable suggestions for improvements. Any inaccuracies are of course our own responsibility. The paper continues earlier work with Daniel Wiechmann (cf. Gast & Wiechmann 2012), Olga Rudolf and Marie Schneider. We have reused some of the data that we coded jointly with these colleagues at an earlier stage. V. Gast wishes to acknowledge financial support from the German Science Foundation (DFG GA 1288/4-1). 1 Technical terms such as “Cleft”, “Topic”, etc. are capitalized. 2 See for instance Akmajian (1970); Prince (1978); Declerck (1984); Collins (1991); Lambrecht (2001); Collins (2006); Hedberg & Fadden (2007); Dufter (2009); Fischer (2009), among other publications. 3 The examples used for this study have been taken from the EUROPARL-corpus, cf. Koehn (2005) and Cartoni & Meyer (2012). The label “EPEG-6” stands for “EUROPARL English-German, version 6”. The corpus contains the proceedings of the European Parliament. Version 6, which we used for our study, contains the proceedings from April 1996 to December 2010, approx. 55 M words in the English corpus part and approx. 49 M words in the German part. The corpus contains both original language and translations. Until 2003, the texts were translated directly from the source languages into any of the target languages. From 2003 onwards, English has been used as
Volker Gast, University of Jena Natalia Levshina, University of Louvain-la-Neuve
378
(2)
Volker Gast, Natalia Levshina
(It-)Cleft Obviously, it is the result that interests us. (EPEG-6/Fr 792462)
In German, the term “Sperrsatz” is often used to distinguish examples like (3) from those of the type illustrated in (4), which are commonly subsumed under the term “Spannsatz” (see for instance Altmann 1981, 2009). Alternatively, the terms “wCleft” and “es-Cleft” are used (e.g. by Fischer 2009; Gast & Wiechmann 2012). (3)
Sperrsatz/w-Cleft Was uns interessiert, ist selbstverständlich das Resultat. (≡ [2])
(4)
Spannsatz/es-Cleft Es ist zweifellos der Terrorismus, der verhindert, dass die Parteien an einen Verhandlungstisch kommen. (EPEG-6/It 511911) ‘It is certainly terrorism that prevents the parties from coming to the negotiating table.’
We use the term “w(h)-Cleft” as a generalization over the corresponding structures of English and German, i.e., English wh-Clefts and German w-Clefts. In addition to the two types of Cleft sentences introduced above, a third type is standardly distinguished, i.e., reversed pseudo- or w(h)-Clefts (cf. the references in Note 2). Examples of this construction are given in (5) and (6) for English and German, respectively. (5)
Reversed pseudo-Cleft/wh-Cleft Champaign is what I like. (Lambrecht 2001: 467)
(6)
Reversed w-Cleft Ist es das, was dieser Junge Mann in Indien suchte? (Altmann 2009: 17) ‘Is that what this young man was looking for in India?’
We believe that reversed w(h)-Clefts – in English as well as in German – constitute an interesting topic of its own, and that they are not merely a structural variant of uninverted w(h)-Clefts. As has been shown by Hedberg & Fadden (2007), among others, their information structural properties are probably more similar to those
a “pivot language” (Cartoni & Meyer 2012: 3), i.e., all languages were first translated into English and then into the relevant target language. In the indication of the source, the abbreviation after the slash indicates the original language of the example in question, and the following number the line in the corpus.
Motivating w(h)-Clefts in English and German
379
of it-Clefts than to those of uninverted w(h)-Clefts. We will therefore focus on uninverted w(h)-Clefts in this contribution, but a comparison with reversed w(h)Clefts is certainly an interesting topic for future research. The present study addresses the following question: (7)
Research question What determines the distribution of w(h)-Clefts in English and German?
The question in (7) is approached on the basis of data from a parallel corpus, EUROPARL (more precisely, EUROPARL is a translation corpus; cf. Koehn 2005; Cartoni & Meyer 2012; see also Note 3). The design of the study is as follows: We use properties of English wh-Clefts as independent variables that are taken to correlate with one formal property of the corresponding German sentences, captured as the dependent variable of our study, the presence or absence of a w-Cleft. The properties of the English wh-Clefts constituting the independent variables are intended to reflect motivations for (not) using a w(h)-Cleft. The rationale underlying this design is the following: We assume that all pairs of sentences from the parallel corpus are (semantically/pragmatically) near equivalent, i.e., they convey basically the same “message”, speaking in very general (information-theoretical) terms (cf. Gast forthcoming on the notion of “near equivalence” in contrastive linguistics). This assumption of interlingual near equivalence in a translation corpus is obviously an idealization, but in general, the translations of the EUROPARL-corpus are of a very high quality and certainly come close to that ideal. Given the (a priori) assumption of interlingual equivalence in our translation corpus, and given the (theoretically motivated) assumption that the properties of English sentences constituting the independent variables of our study reflect (communicative) motivations for (not) using a w(h)-Cleft, we can expect to find systematic covariation between the independent variables describing the English data on the one hand, and the occurrence or non-occurrence of a wCleft in the German data on the other, as the motivations that are reflected in the English data should have observable reflexes in the German data as well. Given that the messages underlying pairs of sentences are taken to be invariant, the question arises why there should be differences in the distribution of English wh-Clefts and German w-Clefts at all. We believe that the observable differences in the distribution of w(h)-Clefts in English and German result from the fact that English wh-Clefts and German w-Clefts are part of different systems. As a consequence, they compete with different sets of alternative structures. The (non-)availability of alternative structures, in turn, is a crucial determinant of the distribution of w(h)-Clefts and is thus expected to lead to distributional differences between English and German.
380
Volker Gast, Natalia Levshina
What we investigate, thus, is the interplay between motivations for using a w(h)-Cleft and the actual use of w(h)-Clefts in English and German, viewed as the result of a multifactorial decision based on communicative needs and intentions – the invariant or tertium comparationis – and available linguistic resources, the most important factor giving rise to distributional variation between English and German. The relationship between specific motivations underlying the use of a w(h)Cleft and the type of structure actually found in a given case will be modelled (metaphorically and informally) in terms of a cost-benefit analysis. As will be detailed in Section 4, we assume that w(h)-Clefts provide specific benefits, which are (partly) independent of their basic information structural function. The benefits of w(h)-Clefts provide motivations for using these structures, in spite of their higher costs (in comparison to canonical clauses). Accordingly, a w(h)-Cleft is expected to be used in those cases where it provides a benefit of the type discussed in the present paper which would not be provided by a canonical (“cheaper”) structure. Obviously, the various motivating factors are not mutually exclusive but may contribute jointly to the use of a w(h)-Cleft. The more motivating factors there are, the more likely the use of a w(h)-Cleft will be. The asymmetrical design of our study – we investigate covariation between English wh-Clefts and their German counterparts in a translation corpus – is due to the fact that w(h)-Clefts are much more widely distributed in English than in German. As Gast & Wiechmann (2012) have shown, the ratio of w(h)-Clefts in these languages is approximately 4 : 1 in the EUROPARL-corpus. Even though there are cases where we find a w(h)-Cleft in German but not in English – perhaps even systematic and general ones (cf. Section 4.3 for possible candidates) – it thus makes sense to take the English data as a point of departure. An investigation determining the English structures corresponding to German w-Clefts would probably provide interesting additional information about the distribution of w(h)-Clefts in the languages under comparison, but would not invalidate the findings reported in the present study. As the (aymmetrical) design of our empirical study only distinguishes two cases – (i) those where we find a w(h)-Cleft in both languages under comparison, and (ii) those where we only find a wh-Cleft in English – we will be dealing with two major types of benefits as well, i.e., (i) those benefits that are expected to motivate the use of a w(h)-Cleft in English to a (significantly) greater extent than in German, and (ii) those benefits that are not expected to lead to a distributional overrepresentation of w(h)-Clefts in English as compared to German. The first type of benefit will be called “English-specific” and the second type “general”, which we use as an abbreviation for “non-English-specific”.
Motivating w(h)-Clefts in English and German
381
Before turning to our quantitative analysis, some remarks on the structure and distribution of w(h)-Clefts are made in Sections 2 and 3. In Section 4, the assumed motivations for the use of w(h)-Clefts in English and German are discussed, and five hypotheses are formulated. These hypotheses are operationalized and tested in Section 5. In Section 6, we identify correlations between independent variables on the basis of a Multiple Correspondence Analysis. As there are clear correlations between two of the variables investigated, we identify three (instead of five) major types of English wh-Clefts, which are associated with specific motivations of use. Only one of these types is also (more or less) widely attested in German, which explains the differential distribution of w(h)-Clefts in English and German. Section 7 contains a summary and the conclusions.
2 Remarks on the structure of w(h)-Clefts W(h)-Clefts are constituted by a matrix clause headed by a copula in specificational function and a w(h)-clause filling the subject position of that copula (see Declerck 1984; den Dikken 2009; Gast & Wiechmann 2012 for the specificational – as opposed to predicational or equative – interpretation of the copula in such sentences). There is no general consensus concerning the exact function of the w(h)clause, but mostly it is regarded as a free relative clause (see e.g. Altmann 2009). Alternatively, it could be interpreted as an indirect question (cf. for instance Faraci 1971). We will not take a stance in this matter and simply use the terms “w(h)clause” and “Cleft constituent” for the two main constituents of a w(h)-Cleft. This terminology is illustrated in (8), using example (1) above. (8)
[ What I miss in Mr Martens’ report ] are [ the guidelines . . . ] [ w(h)-clause ] cop [ Cleft constituent ]
W(h)-Clefts are restricted in terms of the category of the Cleft constituent. As will be shown in this section, there are only few differences between the structural properties of English and German w(h)-Clefts. English wh-Clefts are found with DPs, sentential Cleft constituents, VPs and to-infinitives (cf. [9]–[12]). (9)
DP What is actually needed however are [DP funds and resources]. (EPEG-6/Sp 425590)
382
Volker Gast, Natalia Levshina
(10)
Finite clause (CP) What this means is [CP that we are in a position to continue the European Union’s development in line with what is needed]. (EPEG-6/Ge 281996)
(11)
VP What we have to do is [VP apply the Community acquis and not call other policies into question]. (EPEG-6/Sp 300029)
(12)
To-infinitive (TP) What we do not want is [TP to exempt from transparency obligations the small category of highly specialised financial journalists . . . ] (EPEG-6/ Du 334172)
English wh-Clefts are ungrammatical with adverbials and PPs as Cleft constituents (cf. [13] and [14]). (13)
Adverb *When I became a young revolutionary was then. (Prince 1978: 885)
(14)
PP *What/where/how many protest is against pardoning these. (Prince 1978: 885)
German w-Clefts cover basically the same range of possibilities as the English whClefts, as is illustrated in (15)–(18). (15)
DP Was die Europäische Union jetzt braucht, ist [DP ein schnelleres Wirtschaftswachstum] . . . (EPEG-6/Pl 594004) ‘This means that faster economic growth [. . . ] is needed in the European Union.’
(16)
Finite clause (CP) Was ich jetzt klar erkenne, ist, [CP dass es keine wirkliche Chancengleichheit in der landwirtschaftlichen Produktion gibt . . . ] (EPEG-6/En 595461) ‘The one thing I now see clearly is that there are no real level playing fields in agricultural production [. . . ].’
Motivating w(h)-Clefts in English and German
383
(17)
VP Was wir nicht können, ist, [VP ein Abkommen schließen, das die WTO dann als nicht kompatibel ablehnt]. (EPEG-6/Po 1543370) ‘What we cannot do is come up with an agreement and afterwards WTO tells us it is not compatible.’
(18)
Zu-infinitive (TP) Was wir wollen, ist ja, [TP den Unfallopferschutz zu verbessern]. (EPEG6/Ge 1731070) ‘Surely what we want is to improve the protection of accident victims.’
As in English, adverbs cannot be w-clefted in German: (19)
Adverb *Wie ich nach Hause kam war schnell. Lit.: ‘How I came home was fast.’
There is one syntactic difference between English and German. As is well known, English allows extraction out of PPs, so complements of prepositions can be whclefted (cf. [20]). (20)
Extraction out of PP in English Today, whati we are talking [PP about ei ], to put it simply, is the ownership of ideas. (EPEG-6/En 634287)
Such extraction is not possible in German. German, in turn, has a lexical option for these cases which English lacks, and which renders w-clefting of PPs possible. The w-forms of so-called “conjunctional adverbs”, which consist of the prefix wo(r)- and a preposition, stand for a syntactic unit of the type [PP P prowh ], and they can be w-clefted, as is illustrated in (21). (21)
W-form of conjunctional adverb in German Worüberi wir heute sprechen [PP ei ], ist – um es einfach auszudrücken – das Eigentum an Ideen. (≡ [20])
In terms of frequency, the difference between English and German that is illustrated in (14), (20) and (21) is not particularly relevant, as the corresponding structures are relatively rare, at least in the corpus used for the present study, the EUROPARL-corpus. We will thus assume that by and large, w(h)-Clefts can be used in the same range of contexts in English and German, as far as their syntac-
384
Volker Gast, Natalia Levshina
tic restrictions are concerned, and that actual distributional differences are due to differences in the motivating factors, or the interplay between those factors and the “linguistic ecologies” of English and German. A quantitative comparison of the types of Cleft constituents found in the English and German part of the EUROPARL-corpus has been provided by Gast & Wiechmann (2012).
3 The standard motivation of w(h)-Clefts 3.1 On Givenness and Topic-Comment structure In the comprehensive literature on that construction, w(h)-Clefts are mostly assumed to have an information-structural function, i.e., a function relating to “common ground management” (cf. Krifka 2007 for that term). One of the most influential information structural descriptions of w(h)-Clefts has been provided by Prince (1978), who describes their function as follows: A wh-cleft will not occur coherently in a discourse if the material inside the (subject) whclause does not represent material which the coöperative speaker can assume to be appropriately in the hearer’s consciousness at the time of hearing the utterance. (Prince 1978: 888)
There is a broad consensus that wh-Clefts are associated with stricter conditions of use than it-Clefts, as was also claimed by Prince (1978). However, the claim that the w(h)-clause needs to be Given (in a technical sense, hence capitalized) has been qualified and challenged. In their empirical investigation, Hedberg & Fadden (2007) have pointed out that the w(h)-clause of a w(h)-Cleft need not be “referentially Given”, in terms of Gundel & Fretheim (2004). What is required is that it be “relationally Given”. In a nutshell, this means that the w(h)-clause is a Topic and the Cleft constituent a Comment (about the Topic). We will adopt this terminology below, even though our analysis differs slightly from the one provided by Hedberg & Fadden (2007). The analysis of Prince (1978) – in particular, her observation that w(h)-Clefts are subject to more specific conditions of use than it-Clefts – has been challenged by Declerck (1984), who claims that w(h)-Clefts cover the same range of information structural possibilities as it-Clefts. He observes that the w(h)-clause of a Cleft sentence need not be Given, as in (22), and that there are cases where w(h)-Clefts can even be used as discourse openers (cf. [23]).
Motivating w(h)-Clefts in English and German
385
(22)
A: Those apples are good, aren’t they? B: So they are! What keeps me from eating all of them is that mother would be furious if I left none for the others. (Declerck 1984: 259)
(23)
My dear friends, what we have always wanted to know, but what the government has never wanted to tell us, is what exactly happens at secret conferences like the one you have been reading about in the papers this week. (Declerck 1984: 257)
Similar cases were discussed by Prince (1978) already (e.g. her example 14 on p. 888), and it seems clear that there is accommmodation at work. We will thus assume that the Prince-Hedberg-Fadden analysis is basically correct. Gast & Wiechmann (2012) have slightly rephrased the analyses of Prince (1978) and Hedberg & Fadden (2007) by using the concept of “Quaestio” from discourse analysis (cf. Klein & von Stutterheim 1987; Gast & van der Auwera 2011; cf. also Büring’s 2003 concept of “[current] question under discussion/QUD”). Roughly speaking, the Quaestio-theory assumes that every utterance in a discourse corresponds to some (mostly implicit) question that is under discussion. This question is called the “Quaestio” of the relevant utterance (cf. Klein & von Stutterheim 1987). The answer given to the Quaestio is called the “Responsio” (cf. Gast & van der Auwera 2011). In canonical cases, the Quaestio of an utterance can be read off the information structure of the corresponding sentence by replacing the focus with an appropriate question word. More often than not, however, the Quaestio is not immediately obvious and needs to be inferred from contextual information. The sequence of sentences in (24) (from the EUROPARL-corpus) will help to clarify the concept of “Quaestio”. (24)
a. You have presented a sound economic programme, Commissioner Almunia. I am able to assent to everything contained in it. All I should like to say is that I should very much like to see an additional dimension. b. What I should dearly like to hear come out of the economic guidelines is a constructive message to all the Member States, saying “let us now, together and in each individual country, invest in the Lisbon Objectives over the next three or four years, and let us do so at one and the same time”, for doing so simultaneously would be Europe’s secret weapon. (EPEG-6/Da 17923)
386
Volker Gast, Natalia Levshina
In (24a), the speaker mentions “an additional dimension” that (s)he would like to see. This raises a question, namely: “What additional dimension would you like to see?” This – here, implicit – question can be assumed to be the Quaestio of the following sentence. The sequence of (24a) and (24b) could be thought of as a dialogue of the following form: (25)
A: I should very much like to see an additional dimension. B: What additional dimension would you like to see? A: What I should dearly like to hear come out of the economic guidelines is a constructive message to all the Member States [. . . ]
The question asked by B in (25) corresponds to the wh-clause of the Cleft sentence in (24b). Accordingly, the effect of using a w(h)-Cleft in (24b) can be described as follows: The Quaestio of an utterance and the corresponding Responsio are merged into a single sentence as it were. The w(h)-clause of the Cleft sentence corresponds to the Quaestio, the Cleft constituent to the Responsio. This analysis is compatible with the one proposed by Prince (1978), as the Quaestio is obviously under discussion (or “in the hearer’s consciousness”, as Prince 1978: 888 puts it), and it is similar to the one of Hedberg & Fadden (2007), insofar as it is based on information structural relations rather than properties of constituents or referents (i.e., referential Givenness). Still, the Quaestio-Responsio analysis of Gast & Wiechmann (2012) and the Topic-Comment analysis of Hedberg & Fadden (2007) are not equivalent, as a Quaestio need not be an “address”, in terms of Jacobs (2001). For ease of reference, we will nonetheless adopt the (more familiar) terminology of Hedberg & Fadden (2007) and thus phrase our analysis in terms of the dichotomy “Topic vs. Comment”.
3.2 Linear synchronization of Topic and Comment According to the observations made in Section 3.1, w(h)-Clefts have the effect of rendering the Topic (as well as the Comment) of a sentence as a continuous sequence of words. Consider, once again, (1) for illustration. A corresponding canonical sentence is given in (26). The w(h)-clause/Topic is rendered in italics, the Cleft constituent/Comment is enclosed by brackets ([. . . ]C ). (1)
Quaestio: What do you miss in Mr Martens’ report? What I miss in Mr Martens’ report are [the guidelines and visions . . . ]C
(26)
I miss [the guidelines and visions . . . ]C in Mr Martens’ report.
Motivating w(h)-Clefts in English and German
387
As (26) illustrates, the w(h)-clause or Topic would be discontinuous in a canonical sentence. In a w(h)-Cleft, it forms an uninterrupted sequence of words, as does the Responsio/Comment. Topic-Comment structure and linear sentence structure are synchronized as it were. This effect of w(h)-Clefting will be called the “linear synchronization of Topic and Comment” (cf. also Gast & Wiechmann 2012), and it is regarded as the “standard motivation” for the use of a w(h)-Cleft. (27)
The standard motivation of w(h)-Clefts A w(h)-Cleft is used to render the Topic and the Comment (each) as uninterrupted sequences of words (linear synchronization of Topic and Comment).
The linear synchronization of Topic and Comment is called the “standard motivation” of w(h)-Clefts insofar as it can be recovered in most cases, even though many examples obviously imply some degree of accommodation. It is thus (basically) a necessary, though not sufficient, condition for the use of a w(h)-Cleft.⁴ In German, the linear synchronization of Topic and Comment can sometimes be achieved without forming a w(h)-Cleft where this would not be possible in English. In fact, the German example corresponding to (1) in the EUROPARL-corpus – given in (28) below – is not a w(h)-Cleft. As in the English examples given above, the w(h)-clause/Topic is rendered in italics. (28)
Im Bericht Martens vermisse ich deshalb [die übergeifenden Entwicklungslinien und Visionen für die Gestaltung einer neuen Zusammenarbeit]C . (EPEG-6 1451517, ≡ [1])
There are thus examples where we find a w(h)-Cleft in English but not in German, as the verb-second order of the latter language allows for a certain freedom of word order already. Still, in many cases a German w-Cleft seems to be motivated by the synchronization of Topic-Comment structure as well. Consider (29):
4 We use the adverbial hedge “basically” because there might be cases where the primary function of w(h)-Clefts is marginal at best and the use of a w(h)-Cleft is mainly motivated by (originally) secondary benefits like those discussed in Section 4 – perhaps cases of “exaptation”, speaking in evolutionary terms; cf. also Gast & Wiechmann (2012: 338).
388
(29)
Volker Gast, Natalia Levshina
Quaestio: Womit konnten wir die Sache für uns gewinnen? ‘What helped us win (that matter)?’ Womit wir die Sache für uns gewinnen konnten, war [eine schriftliche Erklärung der Kommission . . . ]C . ‘What helped us win was [the Commission’s pledge, in the form of a written statement . . . ]C ’
As is illustrated by (30), a corresponding canonical verb-second sentence would exhibit a discontinuous Topic: (30)
Wir konnten die Sache [mit einer schriftlichen Erklärung der Kommission]F für uns gewinnen. ‘We were able to win [with a written statement of the commission]C .’
While many examples of w(h)-Clefts can be explained on the basis of the standard motivation as described in (27) above, we will argue that they are often motivated by other, additional factors, including ones that do not relate to information structure in a narrow sense. Such motivating factors are the benefits of our study, a concept to be explained in the next section.
4 Motivating w(h)-Clefts: Benefits, costs and obstacles We assume a very simple model of language production which is obviously inspired by ideas from Optimality Theory (cf. Prince & Smolensky 1993) without making use of the formalisms provided by that framework. Speakers want to express specific meanings or functions, and they have a repertoire of structures which they can use for that purpose. Each structural option comes with specific benefits and with specific costs. The choice of a given construction is thus a matter of finding a balance between costs and benefits. Moreover, there may be obstacles to the use of a specific structure, e.g. syntactic constraints. The benefits of using a w(h)-Cleft, as well as one obstacle, are described in this section. We will not provide any quantitative measure for the costs of w(h)Clefts in comparison to canonical sentences (SVO in English and verb-second in German), but we will assume a general economy constraint on w(h)-Clefts, much in the sense of Optimality Theory. W(h)-Clefts, thus, need to be motivated in some way, otherwise they will be avoided. As mentioned in Section 3.2, the standard condition of w(h)-clefting – the linear synchronization of Topic and Comment – is
Motivating w(h)-Clefts in English and German
389
not regarded as a benefit, but rather as the primary function of that construction, and as a necessary, though not sufficient condition for its use. The actual use of w(h)-Clefts is regarded as a probabilistic choice: The more benefits a w(h)-Cleft provides in a given context, the more likely it will occur.
4.1 Creating an IP boundary for separate nuclear accents: Phonological separation of Topic and Comment The linear synchronization of Topic and Comment as described in (27) is not the only structural effect of forming a w(h)-Cleft. In addition to forming a continuous chain of words, the Topic and the Comment are contained in different constituents, and are thus also separated at the paradigmatic level. This is illustrated in (31) for example (1) (“T” and “C” stand for “Topic” and “Comment”, respectively). (31)
Whati
T
are
I miss ei
C the guidelines [. . . ]
in Mr Martens’ report One of the major benefits of putting the Topic and the Comment in different constituents concerns the syntax-phonology interface, in particular the organization of “tonality” and “tonicity”, following Halliday’s (1967) terminology (cf. Wells 2006 for a more recent overview of basic notions of intonation). The constituent boundary between the Topic and the Comment allows the speaker to align the two constituents with separate intonation phrases and, accordingly, to have separate nuclear accents in each constituent. This benefit of a w(h)-Cleft will be called the “phonological separation of Topic and Comment”. Given that this rather abstract benefit manifests itself in the presence of (at least) two nuclear accents, and given that nulear accents are associated with some type of contrast, the phonological separation of Topic and Comment will be operationalized as the presence of contrast in the Topic (the Comment is assumed to necessarily contain at least one nuclear accent; for a discussion of contrastive Topics in English and German, see Gast 2010).
390
Volker Gast, Natalia Levshina
For illustration, consider, once again, (a simplified version of) (3) and its canonical counterpart in (32): (3)
Was uns interessiert, ist das Resultat. ‘What interests us is the result.’
(32)
Uns interessiert das Resultat. Lit.: ‘Us interests the result.’
In both (3) and (32) the Topic and the Comment are (each) adjacent; but only in (32) are the two information structural components contained in separate constituents (cf. [33] vs. [34]; the Topics are rendered in italics). (33) T ist Was
uns interessiert
C das Resultat
(34) Uns interessiert das Resultat The structure in (3)/(33) allows the speaker to align the Topic and the Comment with separate intonation phrases more easily than the one in (32)/(34). As a consequence, the wh-Cleft in (3)/(33) allows for the placement of a nuclear accent in the wh-clause/Topic, and thus for the indication of contrast, more easily than the canonical clause in (32)/(34). The intonation in (35) is therefore more natural than the one in (36) (brackets indicate constituents, “|” an IP-boundary, and small caps the nuclear accent). (35)
[Was uns interessiert] | [ist das Resultat].
(36)
?Uns [interessiert | das Resultat]
In English, the constituent boundary between the wh-clause and the Cleft constituent is often used for the placement of adverbs that are little integrated into the sentence structure, e.g. “conjuncts” in terms of Quirk et al. (1985). For instance, the adverb however is often inserted into the “gap” between the wh-clause and the Cleft constituent, as in (37).
Motivating w(h)-Clefts in English and German
(37)
391
What we are talking about here, however, are medium-sized vessels [. . . ] (EPEG-6/Ge 585863)
As there is no obvious reason why the phonologically motivated use of w(h)-Clefts as described in this section should be more common in English than in German, we will assume that the benefit of phonologically separating the Topic and the Comment is a general one, i.e., a benefit which will not contribute to the use of w(h)-Clefts in English to a greater extent than in German.
4.2 Structural separation of propositional content and utterance comment The factor discussed in this section emerged upon inspection of the data used for this study. In a relatively high number of examples, the w(h)-clause does not contribute anything to the propositional content of the sentence, but merely provides a frame of evaluation, insofar as it contains information about the contextual embedding, the modal or epistemic source of the propositional content, etc. For instance, the w(h)-clause often contains a doxastic predicate like think (what I think is . . . ), a verb of saying (what I want to say is . . . ), or similar expressions not affecting the truth conditions of a sentence. In (38) and (39), the w(h)-clause links the propositional content conveyed in the Cleft constituent to the preceding utterance, characterizing it as a conclusion or consequence. (38)
What this means is that we are in a position to continue the European Union’s development in line with what is needed. (EPEG-6/Ge 281996)
(39)
Was ich sagen will, ist, dass wir zusammenstehen müssen [. . . ] (EPEG-6/ Sp 273398) ‘What I am saying is that we should act together [. . . ]’
In such cases w(h)-Clefts do not seem to primarily reflect Topic-Comment structure, but the difference between the level of propositional content (in the Cleft constituent) on the one hand, and the frame of evaluation (in the w[h]-clause) on the other. The matrix predication can be regarded as a comment on the proposition expressed in the Cleft constituent – an “utterance comment”, as we call this type of predication (cf. also Gast & Wiechmann 2012). Utterance comments are crucially characterized by the fact that they do not contribute anything to the propositional content expressed in the clause (cf. also Section 5.2).
392
Volker Gast, Natalia Levshina
W(h)-Clefts make the distinction between propositional content on the one hand, and utterance comment on the other, transparent, insofar as the two pieces of information are contained in separate constituents. This type of structural separation is illustrated in (40). (40) What Utterance comment
is that
Proposition
I want to say we should act together [. . . ] The separation of utterance comment and propositional content seems to motivate w(h)-Clefts to a greater extent in English than in German. In the latter language, we often find embedded verb-second structures instead.⁵ The English example (41), for instance, corresponds to an embedded verb-second sentence in the German corpus part (cf. [42]). (41)
What all this means is that we are in fact incapable of doing much at all. (EPEG-6/Pl 637137)
(42)
Das heißt, eigentlich können wir nicht viel tun.
The most salient property of embedded verb-second structures in German is that the embedded clause has (representative) illocutionary force. In other words, it is asserted. A comparable stucture is not available in English. While English also allows complement clauses without an introducing complementizer, the relevant clauses are not distinguished structurally from “ordinary” (unasserted) complement clauses. Moreover, such that-omission is uncommon in high-register language as represented in the EUROPARL-corpus (cf. Biber et al. 1999: 680–683). Given that German can use a more economic structure (embedded verbsecond clauses) in order to paradigmatically separate the utterance comment and the proposition, while English does not have an obvious structural alternative of this type, it is to be expected that this factor will motivate w(h)-Clefts to a greater
5 German standardly uses verb-second structure in main clauses and SOV-structure in subordinate clauses. Verb-second structure is also found in embedded clauses, however, as in (42). In this case, a complementizer cannot be used.
Motivating w(h)-Clefts in English and German
393
extent in English than in German. In other words, the structural separation of propositional content and utterance comment is regarded as an English-specific benefit or motivation.
4.3 Linear synchronization of constituency and operator scope In both languages under comparison, w(h)-Clefts often lead to a reordering of scope-bearing operators, either relative to each other or relative to their scope domains. Consider the w-Cleft in (43), which is attested in the EUROPARL-corpus, and the corresponding canonical main clause (with extraposition) in (44). (43)
Was nicht passieren darf ist, dass es eine Vorgabe der Standards durch die Kommission durch die Hintertür gibt. (EPEG-6/Ge 358731) ‘What must not be allowed to happen is standards being imposed by the Commission through the back door.’
(44)
Es darf nicht passieren, dass es eine Vorgabe der Standards durch die Kommission durch die Hintertür gibt. Lit.: ‘It must not be allowed to happen that standards are imposed by the Commission through the back door.’
There are two scope-bearing operators, the deontic possibility (permission) modal darf and the negator not. The negator takes scope over the modal. While in a wCleft, the order of elements reflects the scope-relations directly, in the corresponding main clause the element with narrow scope precedes the one with wide scope. (45) illustrates the (transparent) scope relations in the w-Cleft. Was nicht passieren darf ist, [dass . . . ] ¬ [ [ p ]] ◻
(45)
The effect that can be observed in (43) and (45) will be called “linear synchronization of constituency and operator scope”. The fact that (in German) scope relations are mirrored more directly in wClefts than in verb-second clauses is no coincidence but obviously related to the fact that the w-clause is a subordinate clause – with SOV order – and that in German, scope relations are more transparent in subordinate clauses than in main clauses (cf. König & Gast 2012: Ch. 11 for a comparison of English and German sentence structure with reference to scope relations). The effect in examples like (43) is that of a complete “outsourcing” of propositional operators. The w-clause
394
Volker Gast, Natalia Levshina
contains no lexical material at all – the verb passieren is basically a place holder for the predicate of the Cleft constituent. This separation of propositional operators and lexical material probably makes the relevant sentences more parserfriendly. In the context of political speech – a register which is characterized by a high degree of editorial elaboration and geared towards hearer-friendliness – transparency at the level of sentence operators is a welcome effect. As a matter of fact, the quantitative investigation in Section 5 will show that the linear synchronization of constituency and operator scope is a strong factor motivating w-Clefts in German. Even though English does not exhibit a difference between main clause order and subordinate clause order, synchronization effects between form and function can also be observed here, though probably to a lesser extent. Sometimes propositional operators take scope over a clausal constituent in the subject position, while following that constituent in terms of linear order. The example in (46) is a case in point. The scope relations are indicated in (47). In the corresponding canonical clause in (48), the scope-bearing operators occur towards the end of the sentence, even though they take scope over the gerund headed by being in subject position. What must not be allowed to happen is standards being imposed by the Commission through the back door. (EPEG-6/Ge 358730)
(47)
What must not be allowed to happen is [standards . . . ] ◻ [¬ [ [ p ]]]
(48)
Standards being imposed by the Commission through the back door must not be allowed.
◻
(46)
As there is no reason to assume that the reorganization of scope relations in w(h)Clefts is more important in English than in German, we will assume that it is a general benefit (remember that there is no category of “German-specific” benefit in our investigation, for which the linear synchronization of constituency and operator scope would be an obvious candidate).
4.4 Avoiding left-heavy constituent structure In some cases, w(h)-Clefts seem to be motivated in basically syntactic terms. In particular, they are sometimes used to avoid “left-heavy” constituent structure and have a function similar to that of extraposition. Consider the discourse se-
Motivating w(h)-Clefts in English and German
395
quence in (49). There is no obvious information structural motivation for the use of a wh-Cleft in this example, as remarkability is not, in any way, under discussion or topical, there is no contrast involved, and there is no scope-bearing operator, either (some discourse pragmatic motivation can of course always be accommodated). (49)
a. The first version of the report was used as a chance for them to get on their hobbyhorse and oppose nuclear energy. b. What is remarkable is that there was no mention, whatsoever, in the report of the Commission’s action programme, which contains many sensible initiatives. (EPEG-6/Fr 119772)
As Hawkins (1994, 2004) has argued, languages tend to exhibit a homogeneous branching direction in their constituent structure, as such structural homogeneity facilitates parsing and, by hypothesis, production (cf. also Gast 2007 for discussion). English is a basically right-branching language and, like most other languages of this type, it avoids sentential arguments in a pre-verbal subject position. (50) provides a simplified tree diagram of the wh-Cleft in (49b). For comparison, the structure with a pre-verbal clausal subject is given in (51). (50) CP What
TP
is remarkable
is
CP
that there was no mention [. . . ] (51) CP
TP is remarkable
That there was no mention [. . . ] In German, w-Clefts are also sometimes used to restructure left-heavy sentences, i.e., sentences with a sentential argument in initial position. An example of this type is given in (52), the corresponding verb-second structure is shown in (53).
396
Volker Gast, Natalia Levshina
(52)
Was dem Kautabak einen Vorteil verleiht, ist, daß der Nikotinabhängige ihn konsumieren kann, ohne seine Umgebung in Mitleidenschaft zu ziehen. (EPEG-6/Fr 58757) ‘What gives snuff an advantage is the fact that nicotine addicts can use it without affecting the environment.’
(53)
Daß der Nikotinabhängige den Kautabak konsumieren kann, ohne seine Umgebung in Mitleidenschaft zu ziehen, verleiht ihm einen Vorteil.
It should be noted that German, unlike English, has basic OV-order and is thus not prima facie right-branching. However, clausal constituents in the forefield (the position preceding the finite verb in verb-second languages) are also avoided, just like pre-verbal clausal subjects in English. Still, structural benefits like those illustrated in (49) and (50) are expected to have a weaker effect in German than in English. The reason is that in many cases, German does not need a w-Cleft to avoid left-heavy constituent structure as in (53). For instance, the German sentence corresponding to (the English example) (49) – here given in (54) – is not a w-Cleft. It exhibits canonical verb-second structure, with the adjective (in its predicative superlative form) in the forefield and the clausal subject in the middle field.⁶ German can thus avoid a left-heavy constituent structure without resorting to the structural choice of a w-Cleft. (54)
Am erstaunlichsten ist aber vielleicht, wie wenig Aufmerksamkeit dieses Thema bei den Eigentümern der Bank [. . . ] findet [. . . ]. Lit.: ‘Most remarkable of all, however, is perhaps how little all this is discussed by the Bank’s owners [. . . ].’
We can thus classify the (structural) motivation of avoiding left-heavy constituent structure as English-specific, as it is expected to have a stronger positive distributional effect on the occurrence of w(h)-Clefts in English than in German.
4.5 An obstacle to the use of w(h)-Clefts: Horror aequi So far, we have only considered the benefits associated with the use of w(h)-Clefts. In some cases, factors preventing the use of a w(h)-Cleft have also been men-
6 The middle field is the field between the finite verb (in second position) and the non-finite verb (if any) at the end of the sentence.
Motivating w(h)-Clefts in English and German
397
tioned, e.g. the fact that German can often achieve specific effects (like the avoidance of sentence-initial clausal constituents) without the use of a w(h)-Cleft. In this section, we will consider an additional factor that leads to asymmetries in the distribution of w(h)-Clefts in English and German, i.e., the tendency to avoid the immediate sequence of identical words (horror aequi) in sentences of the latter language. The phenomenon of horror aequi often seems to prevent the use of Clefts in cases of nominal predication. If the w-clause contains a copula, the resulting sentence contains a sequence of two copulas (because of the verb-final order in subordinate clauses), i.e., a sequence of the type [. . . ist, ist, dass . . . ]. This is, obviously, often avoided for stylistic and perhaps also linguistic reasons. One avoidance strategy found in the German corpus part is the division of the sentence into two parts separated by a colon, as in (56), which corresponds to (the English example) (55). (57) provides the German sentence realized as a w-Cleft, with the mentioned horror aequi-problem. (55)
What is even more ridiculous is that surpluses are produced even under normal conditions. (EPEG-6/Ge 1737040)
(56)
Noch viel unsinniger: Schon unter normalen Bedingungen werden Überschüsse produziert.
(57)
Was noch viel unsinniger ist, ist, dass schon unter normalen Bedingungen Überschüsse produziert werden.
As there are no comparable effects observable in the English data, the horror aequi obstacle is expected to be a factor that decreases the chances of finding a w-Cleft in German as a counterpart of an English wh-Cleft.
4.6 The hypotheses tested in this study Let us take stock of the observations made so far and formulate some hypotheses for our empirical study. We have proposed four motivations for the use of w(h)Clefts. Moreover, we have identified one obstacle which prevents the use of a wCleft in German, i.e., the adjacency of two copulas in German w-Clefts with nominal predication in the matrix clause. The benefits have been categorized into two classes, English-specific ones and general ones. English-specific benefits are assumed to motivate w(h)-Clefts to a greater extent in English than in German, while
398
Volker Gast, Natalia Levshina
Table 1: Benefits of and obstacles to the use of w(h)-Clefts English-specific benefits 1. 2.
Structural separation of propositional content and utterance comment (Sect. 4.2) Avoidance of left-heavy constituent structure (Sect. 4.4)
General benefits 3. 4.
Phonological separation of Topic and Comment (creation of an IP-boundary, Sect. 4.1) Linear synchronization of constituency and operator scope (Sect. 4.3)
Obstacle (German) 5.
Immediate sequence of two copulas (horror aequi, Sect. 4.5)
general benefits are not expected to have any such “pro-English” effect. The two classes of benefits and the obstacle are summarized in Table 1. The distribution of w-Clefts in the German data is expected to be a function of the interplay between the factors summarized in Table 1 (plus some other factors which have not been taken into account). Overall, the ratio of German w-Clefts to English wh-Clefts has been shown to be approximately 1 : 4 in the EUROPARLcorpus (cf. Gast & Wiechmann 2012). Given that in the present (asymmetrical) study we have disregarded those cases where we find a w-Cleft in German while not finding a wh-Cleft in English, the ratio is even lower in our data. Of the 722 cases of wh-Clefts used for this study, 116 correspond to German w-Clefts (cf. Table 2). The ratio of German w-Clefts to English wh-Clefts is thus approximately 1 : 5. Table 2: W-Clefts and alternative structures in the German data w-Cleft
German correspondences of English wh-Clefts
TRUE
FALSE
116
606
ratio of w-Clefts to alternative structures 1 : 5.2 (= 116/606)
The proportion of German w-Clefts in the dataset is approximarely 0.16 (= 116/722). This figure is assumed to be the result of two antagonistic forces. The Englishspecific benefits as well as the horror aequi-obstacle pull the figure towards a proportion even lower than 0.16. By contrast, the general benefits – while obviously not leading to a totally even distribution of w(h)-Clefts in English and German – are expected to have a positive impact on the occurrence of w-Clefts in the German data, in comparison to the overall proportion of w-Clefts.
Motivating w(h)-Clefts in English and German
399
On the basis of the observations made so far, we can formulate the hypotheses in (58). Each hypothesis implies a claim about the frequency of w-Clefts in the German data, under specific conditions observed in the English data. The terms “over”- and “underrepresentation” do not refer to the German as opposed to English data, but to the distribution of w-Clefts within the German data. The descriptive labels of the hypotheses are introduced for future reference. (58)
Hypotheses a. The phonological separation hypothesis If an English wh-Cleft is motivated by the placement of a nuclear accent in the wh-clause, which implies the presence of an IP boundary between the Topic and the Comment, German w-Clefts will be overrepresented in the corresponding sentences, relative to the distribution of w-Clefts in the entire German data set. b. The utterance comment hypothesis If an English wh-Cleft is motivated by the structural separation of propositional content and utterance comment, German w-Clefts will be underrepresented in the corresponding sentences, relative to the distribution of w-Clefts in the entire German data set. c. The scope hypothesis If an English wh-Cleft is motivated by the synchronization of constituency and operator scope, German w-Clefts will be overrepresented in the corresponding sentences, relative to the distribution of w-Clefts in the entire German data set. d. The structural hypothesis If an English wh-Cleft is motivated by the avoidance of left-heavy constituent structure, German w-Clefts will be underrepresented in the corresponding sentences, relative to the distribution of w-Clefts in the entire German data set. e. The horror aequi hypothesis If the wh-clause of an English wh-Cleft contains a nominal predication, German w-Clefts will be underrepresented in the corresponding sentences, relative to the distribution of w-Clefts in the entire German data set.
We will now proceed to the operationalizations and testing of the hypotheses made in (58).
400
Volker Gast, Natalia Levshina
5 Testing the hypotheses In order to test our hypotheses, we extracted 800 examples of English wh-Clefts from the EUROPARL-corpus (cf. Gast & Wiechmann 2012 for some details of the process of data extraction). During the process of coding, several instances of presumable wh-Clefts were identified as not actually representing a structure of this type. Most importantly, the original data set contained a number of examples which were instances of nominal predication with a free relative clause in subject position. Consider (59) and (60). (59)
What we need is a hero.
(60)
What you just said is a lie.
On the face of it, the two examples are structurally entirely parallel. There is an important difference in the interpretation of the copula, however. In (59), the copula is interpreted as “specificational” (cf. Declerck 1984; den Dikken 2009; Gast & Wiechmann 2012), i.e., it specifies a variable in an open proposition (“We need x, x = a hero”). By contrast, (60) attributes a property to an abstract referent described by the free relative clause what you just said (i.e., “the utterance that you just made”). While being more or less identical structurally, the two types of predication are quite different in terms of their semantics. Most importantly, predicational uses of the copula imply the attribution of a property to a referent, and are thus about this referent, whereas specificational uses provide information about a predicate or open proposition. In order to differentiate the two uses of a copula, we need a test. While a copula is systematically ambiguous between a predicational and a specificational reading, other, near equivalent predicates taking nominal complements do not display this ambiguity. We have used the predicate consider in order to determine whether a given instance of a copula is specificational or predicational and, hence, whether or not the corresponding sentence qualifies as a w(h)-Cleft. A copula is regarded as predicational iff “A is B” can be paraphrased as “I consider A B”. This paraphrase is possible for (60), but not for (59): (61)
?I consider what we need a hero. (≠ [59])
(62)
I consider what you just said a lie. (∼ [60])
Motivating w(h)-Clefts in English and German
401
On the basis of this test, we excluded some examples, thus reducing the data set to 722 occurrences of wh-Clefts in English. The data was coded for the variables described in this section. They are the independent variables of the study. Moreover, we coded the data for the type of construction found in the corresponding German sentence, the dependent variable. For the purposes of the present study, this variable was binary, i.e., we determined whether or not there was a w-Cleft in the relevant German sentence. We will now discuss each hypothesis separately, considering the operationalizations used in each case as well as the results obtained.
5.1 The phonological separation hypothesis The phonological separation hypothesis says that w(h)-Clefts are formed in order to create an IP-boundary between the Topic and the Comment, to a more or less comparable extent in English and German. Unfortunately, EUROPARL is a written corpus, so we cannot use any phonological evidence to test whether w(h)-Clefts are in fact associated with the formation of separate intonation phrases. What we have done, instead, is determine whether or not there is any type of contrast in the w(h)-clause which is recoverable from the immediate discourse environment. Such contrast is expected to be reflected phonologically in a contrastive focus accent. A contrastive accent, in turn, by definition requires the formation of a separate intonation phrase. Our operationalization of the phonological separation hypothesis is described in (63). (63)
The phonological separation hypothesis (operationalization) W-Clefts are expected to be overrepresented in the German data (relative to the entire German data set) if in the English data there is an explicit contrast between an element of the wh-clause and some other element in the discourse environment.
The notion of “contrast” was interpreted narrowly. A data point was coded as “TRUE” iff an overt constituent was found in the immediate discourse environment which contrasted with the Cleft constituent of the wh-Cleft. In typical cases, there is a binary contrast, e.g. between positive and negative polarity. In (64a), there is an explicit contrast between what [the report] says and does not say. The wh-clause of (64b) picks up the positive option (what the report does say). In this case, the contrast is also overtly marked through do-support, which is specialized for the function of verum focus.
402
Volker Gast, Natalia Levshina
(64)
a. Therefore, what remains of concern to us is not in the main the contents or analysis within the report but what it says and does not say about the performance of the structural funds programmes themselves. b. What the report does say is that there are significant and damaging delays in payments to localities and regions [. . . ] (EPEG-6/En 13261979
The results of our quantitative study are shown in Table 3. The frequencies that are expected on the assumption that the variable contrast does not have any effect are given for comparison, with the figures rounded to whole numbers. Table 3: The influence of the independent variable contrast on the occurrence of w-Clefts in German (p = 0.0004, OR = 2.18, φ = 0.14) w-Cleft in the German data observed frequencies expected frequencies contrast
TRUE
FALSE
TRUE
FALSE
48 68
148 458
23 85
165 441
TRUE FALSE
Fisher’s exact test delivers a p-value < 0.001, i.e., the impact of the variable contrast in the wh-phrase on the distribution of w-Clefts is very highly significant. The odds ratio is higher than 2, which means that, in our data set, the presence of contrast in the wh-phrase increases the chances of finding a w-Cleft in the corresponding German sentence by a factor of more than 2. Given that the odds ratio is not sensitive to absolute numbers and, hence, not a very good indicator of effect size, we also used the phi-coefficient (φ), which is calculated as follows: (65)
The phi-coefficient φ For a contingency table of the form φ=
¬A A ¬B B
a b c d
ad − bc √(a + b)(c + d)(a + c)(b + d)
In the case of 2 by 2 tables, φ is equal to the absolute value of another popular measure, Cramer’s V. Like Pearson’s correlation coefficient r, φ ranges from −1 (complete negative association) to +1 (complete positive association). 0 indicates no association. The following relation holds between φ and χ2 : (66)
φ2 = χ2 /n
Motivating w(h)-Clefts in English and German
403
For the benefit considered in this section, the presence of contrast in the whclause and the associated presence of an IP-boundary, φ has a value of 0.14. As all other φ-values determined for our data, this is considered a weak association (values lower than .25 are considered weak). Given that we are dealing with five variables, each of which has a relatively low number of TRUE cases, this is not surprising. Determining φ-values for our data will allow us to rank the variables in terms of importance, however (cf. Section 5.6).
5.2 The utterance comment hypothesis According to the utterance comment hypothesis, w(h)-Clefts are sometimes used to structurally separate the propositional content from a frame of evaluation, more commonly in English than in German. This hypothesis can be operationalized as in (67). (67)
The utterance comment hypothesis W-Clefts are underrepresented in the German data (relative to the entire German data set) if a. the wh-clause of an English wh-Cleft contains a propositional complement taking predicate, b. the Cleft constituent contains a clause complementing the matrix predicate, and c. the embedded clause is truth-conditionally equivalent to the entire sentence.
For illustration, consider (68). (68)
What I mean is that we are not starting from square one and, as a producer of several tonnes of beef a year, I know what I am talking about. (EPEG-6/ Ge 37948)
The truth conditions of the embedded clause and of the entire sentence are the same, as is illustrated by the equivalence of (68) and (69): (69)
We are not starting from square one and, as a producer of several tonnes of beef a year, I know what I am talking about.
404
Volker Gast, Natalia Levshina
Our results are displayed in Table 4. As the p-value indicates, the impact of the independent variable under analysis is highly significant, and the hypothesis can be assumed to have been confirmed. The odds ratio is 0.3. In other words, if an English wh-Cleft is used to separate an utterance comment from the propositional content, the probability of finding a w-Cleft in the corresponding sentence of our German data decreases by a factor of more than three. The phi-coefficient is −0.1. Table 4: The influence of the independent variable utterance comment on the occurrence of w-Clefts in German (p = 0.0065, OR = 0.3, φ = −0.1) w-Cleft in the German data observed frequencies expected frequencies utterance comment
TRUE
FALSE
TRUE
FALSE
TRUE FALSE
5 111
79 527
12 103
71 535
5.3 The scope hypothesis The scope hypothesis has so far been phrased generally in terms of scope-bearing operators. More specifically, we have tested whether the co-occurrence of a modal and a negator in the w(h)-clause of a Cleft sentence has any effect on the distribution of w-Clefts in the German data. The scope hypothesis is thus operationalized as follows: (70)
The scope hypothesis W-Clefts are overrepresented in the German data (relative to the entire German data set) if the wh-clause of an English wh-Cleft contains both a modal and a negator.
The number of wh-Clefts with a modal and a negator in the wh-clause was obviously rather small. Still, Fisher’s exact test shows this variable to have a very highly significant (positive) effect on the occurrence of w-Clefts in the German data (p < 0.001). The odds ratio is 10.4, and the phi-coefficient is 0.206. The data are shown in Table 5.
Motivating w(h)-Clefts in English and German
405
Table 5: The impact of the independent variable modal and negator in wh-clause on the occurrence of w-Clefts in German (p < 0.001, OR = 10.4, φ = 0.206)
modal and negator in wh-clause
w-Cleft in the German data observed frequencies expected frequencies TRUE
FALSE
TRUE
FALSE
11 105
6 600
3 113
14 592
TRUE FALSE
5.4 The structural hypothesis Left-heavy structures of the type discussed in Section 4.4 arise when a clause functions as the subject of a sentence. According to our “structural hypothesis”, w(h)Clefts are used to avoid such left-heavy structures. The benefit of avoiding leftheavy constituent structure can thus simply be operationalized in terms of the category and function of the Cleft constituent, and the structural hypothesis can be rephrased as follows: (71)
Structural hypothesis W-Clefts are underrepresented in the German data (relative to the entire German data set) if the Cleft constituent is a clause in subject function.
The results of our study are shown in Table 6. Table 6: The impact of the independent variable clausal subject on the occurrence of w-Clefts in German (p < 0.001, OR = 0.18, φ = −0.16) w-Cleft in the German data observed frequencies expected frequencies clausal subject
TRUE
FALSE
TRUE
FALSE
TRUE FALSE
5 111
124 482
17 96
107 499
Fisher’s exact test delivers a p-value lower than 0.001, so we can consider our hypothesis confirmed. The odds ratio is approximately 0.2. In other words, if the Cleft constituent of an English wh-Cleft is a clause in subject function, the probability of finding a w-Cleft in the corresponding German sentences is five times lower than otherwise. The phi-coefficient is −0.16.
406
Volker Gast, Natalia Levshina
5.5 The horror aequi-hypothesis The obstacle of having two identical words following each other emerges when the w(h)-clause contains a nominal predicate headed by a copula. We can thus rephrase the corresponding hypothesis as in (72). (72)
The horror aequi hypothesis W-Clefts are underrepresented in the German data (relative to the entire German data set) if the wh-clause contains a nominal predicate headed by a copula.
The notion of “nominal predicate” is interpreted relatively broadly, as it covers passive predications of the type what is needed/required, in addition to more prototypical cases like what is necessary. The results of our study are shown in Table 7. The variable governing nominal predicate is the only variable that we have investigated which delivers a p-value higher than 0.01, but rather minimally so (p = 0.013). The result is still significant and we consider our hypothesis confirmed. The odds ratio is 0.5, which means that the presence of a nominal predicate in the wh-clause of an English Cleft sentence decreases the chances of finding a w-Cleft in the German data by a factor of two. The phi-coefficient is −0.093. Table 7: The impact of the independent variable governing nominal predicate on the occurrence of w-Clefts in German (p = 0.013, OR = 0.51, φ = −0.093)
governing nominal predicate TRUE FALSE
w-Cleft in the German data observed frequencies expected frequencies TRUE
FALSE
TRUE
FALSE
18 98
160 446
22 87
149 457
5.6 Summary The results obtained in our study are summarized in Table 8. The hypotheses are ordered by the absolute (rounded) phi-values of the relevant distributions. As can be seen from Table 8, the presence of a modal and negation in the whclause (as an operationalization of the scope hypothesis) is the variable showing the strongest correlation with the presence or absence of a w-Cleft in German, followed by the variables clausal subject and contrast in w(h)-clause. Governing
Motivating w(h)-Clefts in English and German
407
Table 8: Effect sizes and p-values for all independent variables, ordered by importance (the absolute value of φ) p-value (Fisher)
Hypothesis
variable
φ
OR
The scope hypothesis
mod+neg in wh-clause
0.21
10.4
The structural hypothesis
clausal subject
The phonological separation hypothesis
contrast
The utterance comment hypothesis
gov. metaling. pred.
−0.1
0.3
0.0065
The horror aequi hypothesis
nom. pred. in wh-clause
−0.09
0.51
0.013
< 0.001
−0.16
0.18
< 0.001
0.14
2.18
< 0.001
metalinguistic predicate and nominal predicate within the w(h)-clause are the variables showing the weakest correlations with the dependent variable, although the association is still statistically significant. The results displayed in Table 8 obviously have to be taken with caution. It is well known that the individual effect of some independent variable can be distorted by correlations with other variables. We have therefore carried out a multifactorial analysis that is intended to show which variables are intercorrelated, and which variables constitute their own dimensions of variation.
6 Determining correlations between variables: Multiple Correspondence Analysis The method that we have used to determine correlations between variables is Multiple Correspondence Analysis (MCA) with supplementary points. MCA is a technique suitable for the visualization and exploration of multidimensional contingency tables. This is exactly what we need if we cross-tabulate all our variables, and not only pairs of variables, as in the previous section. Like Multidimensional Scaling and Principle Component Analysis, MCA represents multidimensional data in models with a lower number of dimensions, most commonly two or three. However, unlike its relatives, MCA was developed specifically for categorical (non-numeric) variables with such values as TRUE or FALSE, heads or tails,
408
Volker Gast, Natalia Levshina
male or female, smoker or non-smoker, etc. In our data, all variables are categorical and binary (TRUE or FALSE). The distances on a correspondence analysis plot represent associations between variables, formalized as chi-squared distances. The dependent variable, i.e., the presense or absense of a w-Cleft in the German data, will be treated as a supplementary variable. In the Correspondence Analysis jargon, a supplementary status means that the variable will not influence the orientation of the dimensions of variation. In this way we will preserve the special status of the dependent variable and separate the properties of the English sentences from those of the German ones, and we can map the German structures to the English ones. For more details and another application of this method in a contrastive study, see Levshina et al. (2013). The resulting map is shown in Figure 1. The two-dimensional solution accounts for 73.6 % of variation in the data. This means that 26.4 % of information is not captured by the solution. For Multiple Correspondence Analysis, this can be considered an acceptable result. The features that are near the centre of the plot are those with the highest frequencies in the data. Those are the default values, such as a lack of both negation and modality (scopeNo), the absense of an utterance comment (meta_commNo), and the lack of contrast in the wh-clause (contrastNo). The majority of observations (60 %) contain those default values.
Figure 1: A Multiple Correspondence Analysis of the variables investigated in the present study
Motivating w(h)-Clefts in English and German
409
Most variation (60.8 %) is captured by the horizontal dimension, and only 12.8 % by the vertical dimension. This suggests that intercorrelated variables are mostly spread along the horizontal dimension. Especially strongly associated are the variables clausal subject and governing nominal predicate: clause_subjYes and gov_nom_predYes are located very closely to each other in the left part of the plot, whereas their negative counterparts are found in the right part. A glance at our data reveals that the two values indeed often go hand in hand, as clausal subjects are often headed by a copula: Of the 155 cases with a clausal subject, 92 have a nominal predicate in the wh-clause (59 %). An example of this type is given in (55): (55)
[What is even more ridiculousnom_pred ] is [that surpluses are produced even under normal conditions]subj .
The presence of utterance comments (meta_commYes), which is represented in the top right corner of the map, is not strongly associated with any other features, and it co-occurs very rarely with clausal subjects (clause_subjYes) and with nominal predicates in the wh-clause (gov_nom_predYes), as can be seen from their opposite locations on the map. In a sense, meta_commYes constitutes its own functional dimension. As for the vertical axis of the map, it is formed by the presence of modality and negation (scopeYes), and to some extent by the presence of contrast (contrastYes). These features do not co-occur frequently with any other variable and therefore constitute the third major type of context. The relatively close association of these variables is not surprising, since negation, as the reader may recall from Section 5.1, can be used to create a contrast. A relevant example with this combination (modal/negation in w[h]-clause and contrast) is given in (73b). The preceding context is given in (73a). (73)
a. I can also assure the honourable Member that there is never any problem with purely local initiatives, since they can never be regarded as distorting trans-frontier competition. There is no reason why they should not go ahead. b. What we cannot allow are measures specifically designed to distort competition with neighbouring countries in particular, or measures which are claimed to be aid to employment, but which are really also designed to give firms an advantage in a given sector. (EPEG-6/ Du 1306258)
410
Volker Gast, Natalia Levshina
The contrast is between the admissibility of local iniatives and the inadmissibility of trans-frontier competition, or, more specifically, between positive and negative polarity (What we cannot allow . . . ). Note that there is, of course, also a contrast in the Cleft constituent, but we have disregarded such contrastive constituents in the Cleft constituent, as they do not (necessarily) introduce an IP-boundary. To summarize our results, it seems that we can distinguish three major types of English wh-Clefts, each of them associated with a specific benefit: (74)
a. Structural w(h)-Clefts w(h)-Clefts with clausal subjects Benefit: avoidance of left-heavy constituent structure b. Metalinguistic w(h)-Clefts w(h)-Clefts with an utterance comment in the matrix clause Benefit: structural separation of propositional content and utterance comment c. Contrastive w(h)-Clefts w(h)-Clefts with contrast in the wh-phrase Benefit: phonological separation of Topic and Comment
W(h)-Clefts of type (74a) often have a nominal predicate in the w(h)-clause, those of type (74c) (relatively) often contain more than one scope-bearing operator. Let us now examine the mapping of the dependent variable, i.e., the presence or absence of a w-Cleft in the German data, which is plotted onto the space constituted by the English Clefts in Figure 1. The positions of the two levels – cleftYes and cleftNo, see the empty triangles – is determined by the association with the independent variables. Obviously, German w-Clefts (cleftYes) are associated predominantly with the third functional dimension of the corresponding English construction, represented by the features scopeYes and contrastYes. This is not surprising if we consider that both scope interactions in the wh-clause and contrast are factors that have been subsumed under the general benefits, i.e., benefits that are expected to be reflected in English and German more or less alike. With respect to the three types of w(h)-Clefts distinguished in (74), we can thus say that only one of them – type (74c) – is prominent in German, whereas types (74a) and (74b) are characteristic of English but not of German, and are thus mainly responsible for the differential distribution of w(h)-Clefts in English and German. The linguistic interpretation of Figure 1 is shown in Figure 2. Two of the three types of w(h)-Clefts emerging from this map are labelled “English-specific”, one of them “general”, in analogy to the attributes used for benefits. Remember that
Motivating w(h)-Clefts in English and German
411
Figure 2: Major types of w(h)-Clefts as (sets of) variables
“general” means “non-English-specific”, i.e., in the present context, widely attested in both English and German.
7 Summary and conclusions In this contribution we have argued that the distribution of w(h)-Clefts is not just a function of information structural factors, in particular Topic-Comment structure, but also of other effects which are (partly) independent of information structure. We have assumed that the linear synchronization of Topic and Comment represents a necessary condition for the use of a w(h)-Cleft – allowing for some degree of accommodation – but that the probability of finding a w(h)-Cleft in a given data set is also influenced by other factors. Factors favouring a w(h)-Cleft have been called “benefits”, and four benefits of using a w(h)-Cleft have been identified: (i) the phonological separation of Topic and Comment, (ii) the structural separation of propositional content and utterance comment, (iii) the linear synchronization of scope relations and constituency, and (iv) the avoidance of leftheavy constituent structure. In addition, the distribution of w(h)-Clefts has been regarded as being determined by “obstacles”, and one particular obstacle, the im-
412
Volker Gast, Natalia Levshina
mediate co-occurrence of two copulas in German, has been taken into account. These hypotheses have been tested quantitatively, and it has been shown that all of the factors have a significant effect in our data set, but that two of the variables are closely interrelated. We have therefore distinguished three major types of English wh-Clefts, only one of which is often rendered as a w-Cleft in German. On the basis of these observations, we can make the following generalizations concerning the differential distribution of w(h)-Clefts in English and German. In addition to the five contingent (i.e., non-necessary) benefits discussed in Section 4, this summary makes reference to the standard condition for w(h)-Clefts as well (cf. Section 3): (75)
Generalizations concerning the differential distribution of w(h)-Clefts in English and in German a. Both English wh-Clefts and German w-Clefts are standardly motivated by the linear synchronization of Topic and Comment (or Quaestio and Responsio). b. Both English wh-Clefts and German w-Clefts are often motivated by the phonological separation of Topic and Comment (or Quaestio and Responsio), which is required if the w(h)-clause contains a contrastive constituent. Such occurrences of w(h)-Clefts (relatively) often contain more than one scope-bearing operator. c. English wh-Clefts are often motivated by the avoidance of left-heavy constituent structure emerging from a clausal subject in preverbal position. In German, such structurally motivated w-Clefts seem to be rare, as left-heavy constituent structure can also be avoided in canonical (verb-second) clauses. Moreover, the relative rarity of structural w-Clefts in German is probably partly due to the immediate sequence of two copulas which results when the w-clause is headed by a copula (horror aequi). d. English wh-Clefts are often motivated by the structural separation of propositional content and utterance comment. Such instances of w-Clefts seem to be rarer in German, where embedded verb-second clauses are often used instead.
Obviously, the generalizations made in (75) do not describe the (differential) distribution of w(h)-Clefts in English and German exhaustively, and it should have become obvious that there is still a lot of work to do if we want to properly under-
Motivating w(h)-Clefts in English and German
413
stand the (manifold) reasons why speakers use wh-Clefts in English and w-Clefts in German. Still, we hope to have made some progress towards a better understanding of that matter.
References Akmajian, Adrian. 1970. On deriving cleft sentences from pseudo cleft sentences. Linguistic Inquiry 1. 149–168. Altmann, Hans. 1981. Formen der Herausstellung im Deutschen: Rechtsversetzung, Linksversetzung, freies Thema und verwandte Konstruktionen. Number 106 in Linguistische Arbeiten. Tübingen: Niemeyer. Altmann, Hans. 2009. Cleft- und Pseudocleft-Sätze (Spalt- und Sperrsätze) im Deutschen. In Brdar-Szabó, Rita, Elisabeth Knipf-Komlósi & Attila Péteri (eds.), An der Grenze zwischen Grammatik und Pragmatik, 13–34. Frankfurt: Peter Lang. Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad & Edward Finegan. 1999. Longman Grammar of Spoken and Written English. London: Longman. Büring, Daniel. 2003. On D-Trees, Beans, and B-Accents. Linguistics and Philosophy 26. 511– 545. Cartoni, Bruno & Thomas Meyer. 2012. Extracting directional and comparable corpora from a multilingual corpus for translation studies. In Proceedings of the 8th International Conference on Language Resources and Evaluation (LREC). Istanbul. Collins, Peter. 1991. Cleft and Pseudo-Cleft Constructions in English. London & New York: Routledge. Collins, Peter. 2006. It-clefts and wh-clefts: Prosody and pragmatics. Journal of Pragmatics 38. 1706–1720. Declerck, Renaat. 1984. The pragmatics of it-clefts and wh-clefts. Lingua 64. 251–289. den Dikken, Marcel. 2009. Predication and specification in the syntax of cleft-sentences. Ms., CUNY Graduate Center, 24 pp. Dufter, Andreas. 2009. Clefting and discourse organization: Comparing Germanic and Romance. In Andreas Dufter & Daniel Jacob (eds.), Focus and Background in Romance Languages, 83–121. Amsterdam & Philadelphia: John Benjamins. Faraci, Robert. 1971. The deep question of pseudo-clefts. English Linguistics 6. 48–85. Fischer, Klaus. 2009. Cleft sentences: Form, function and translation. Journal of Germanic Linguistics 21. 167–191. Gast, Volker. 2007. From phylogenetic diversity to structural homogeneity – on right-branching constituent order in Mesoamerica. SKY Journal of Linguistics 20. 171–202. Gast, Volker. 2010. Contrastive topics and distributed foci as instances of sub-informativity: A comparison of English and German. In Carsten Breul & Edward Göbbel (eds.), Comparative and Contrastive Studies of Information Structure, 15–50. Amsterdam & Philadelphia: John Benjamins. Gast, Volker. forthcoming. Contrastive linguistics: Theories and methods. In Bernd Kortmann & Johannes Kabatek (eds.), Dictionaries of Linguistics and Communication Science: Linguistic Theory and Methodology. Berlin & New York: de Gruyter Mouton.
414
Volker Gast, Natalia Levshina
Gast, Volker & Johan van der Auwera. 2011. Scalar additive operators in the languages of Europe. Language 87. 2–54. Gast, Volker & Daniel Wiechmann. 2012. W(h)-clefts im Deutschen und Englischen. Eine quantitative Untersuchung auf Grundlage des Europarl-Korpus. In Lutz Gunkel & Gisela Zifonun (eds.), Jahrbuch des IDS 2011, 333–362. Berlin & New York: de Gruyter Mouton. Gundel, Jeanette K. & Thorsten Fretheim. 2004. Topic and focus. In Lawrence Horn & Gregory Ward (eds.), The Handbook of Pragmatics, 175–196. London: Blackwell. Halliday, Michael A.K. 1967. Intonation and Grammar in British English. The Hague: Mouton. Hawkins, John. 1994. A Performance Theory of Order and Constituency. Cambridge: Cambridge University Press. Hawkins, John. 2004. Efficiency and Complexity in Grammars. Oxford: Oxford University Press. Hedberg, Nancy & Lorna Fadden. 2007. The information structure of it-clefts, wh-clefts and reverse wh-clefts in English. In Nancy Hedberg & Ron Zacharski (eds.), The GrammarPragmatics Interface. Essays in Honor of Jeanette K. Gundel, volume 155 of Pragmatics & Beyond, 19–48. Amsterdam & Philadelphia: John Benjamins. Jacobs, Joachim. 2001. The dimensions of topic–comment. Linguistics 39. 641–681. Jespersen, Otto. 1937. Analytic Syntax. London: Unwill. Klein, Wolfgang & Christiane von Stutterheim. 1987. Textstruktur und referentielle Bewegung. Linguistische Berichte 109. 67–92. Koehn, Philipp. 2005. Europarl: A parallel corpus for statistical machine translation. Phuket. MT Summit X. König, Ekkehard & Volker Gast. 2012. Understanding English-German Contrasts. Berlin: Erich Schmidt-Verlag, 3rd edition. Krifka, Manfred. 2007. Basic notions of information structure. In Manfred Krifka & Caroline Fery (eds.), Interdisciplinary Studies of Information Structure 6, 13–56. Potsdam: Universitätsverlag. Lambrecht, Knud. 2001. A framework for the analysis of cleft constructions. Linguistics 39. 463–516. Levshina, Natalia, Dirk Geeraerts & Dirk Speelman. 2013. Mapping constructional spaces: A contrastive analysis of English and Dutch analytic causatives. Linguistics 51. 825–854. Prince, Alan & Paul Smolensky. 1993. Optimality Theory: Constraint Interaction in Generative Grammar. Rutgers University Cognitive Science. Prince, Ellen. 1978. A comparison of wh-clefts and it-clefts in discourse. Language 54. 883– 906. Quirk, Randolph, Sydney Greenbaum, Geoffrey Leech & Jan Svartvik. 1985. The Comprehensive Grammar of the English Language. London: Longman. Wells, John C. 2006. English Intonation: An Introduction. Cambridge: Cambridge University Press.
Subject index Adverb / Adverbial – Adverb 20, 22, 39, 242, 243, 383, 390 – Adverb phrase 12 – Adverbial 28, 127, 131, 223, 224, 227, 235, 265, 270, 293, 295, 297–300, 302, 310, 317, 329, 337, 382 – Adverbial adjunct 240, 241, 242, 262, 263, 268, 269 – Adverbial clause 12, 223 – Adverbial cleft constituent 224, 241, 242, 245, 259, 260, 263, 267, 268 – Adverbial focus 231 – Adverbial hedge 387 – Adverbial preposition phrase 234 – Anaphoric adverbial 237 – Conjunctional adverbs 383 – Focalizing / Focusing adverb 238, 245, 247–251, 264, 265, 369, 373 – Intensifying adverb 369 – Location adverb 169, 262 – Relative adverb 35 – Scene-setting adverbial 237–238 All-focus / All-new 109, 110, 112, 116, 126, 134, 329, 334, 335, 336, 341, 342 Anaphoric 124, 131, 252, 262, 267, 270, 346 – Anaphora 111, 232, 240, 257, 258 – Anaphoric adverbial 237 – Anaphoric cleft constituent 23, 238, 240, 244, 247, 249, 255, 256, 260, 262, 263, 267, 268, 326, 335 – Anaphoric focus 231 – Anaphoric function 222, 245, 267 – Anaphoric noun phrase 21 – Anaphoric pronoun 17 – Anaphoric relation 256 – Associative anaphor 258, 259 – Coreferential anaphor 259 – Non-anaphoric 219, 222, 262, 268, 269 – Resumptive anaphor 257, 259, 266 Background 101, 107, 109, 110, 111, 114, 116, 126, 134, 204, 232, 233, 240, 257, 341 – Background knowledge 221 – Background settings 260, 267
– Backgroundness 231 – Emotional background 267 – Textual background 363 Canonical structure – Canonical clause 380, 390, 393, 394 – Canonical cleft 87 – Canonical predicative sentence / construction 45, 92, 386, 387, 388, 390 – Canonical relative clause 144 – Canonical verb-second clause 388, 396, 412 Cataphoric 184, 198, 222, 244, 372, 373 Ce qui / que 14, 70, 74, 149, 346 C’est que clause 345–375 Cleft – All-cleft 18, 19, 20, 79 – All-Comment clefts 109 – Alles-clefts 18 – Cleft sentences 1, 3, 4, 9, 10–15, 17, 20, 21, 24–36, 42, 43, 45, 49, 50, 53–55, 64, 67, 79–81, 85–88, 91, 92, 94, 101–138, 140, 155, 157, 212, 217–275, 277–321, 325, 328, 330, 332, 335, 337, 339, 340, 377, 378 – Cleft sentences proper 50, 64, 79–80, 81, 86, 91, 92, 94 – Cohesive-cleft 329 – Comment-clause clefts 109 – Discontinuous clefts 108, 109 – es-clefts 378 – Explicit clefts 31, 277, 289, 291, 293, 294, 307, 309, 310, 314, 315, 316, 318 – Implicit clefts 31, 43, 86, 87, 89, 235, 277, 278, 289–294, 307, 309, 310, 314, 315, 316, 318 – Inferential clefts 3, 4, 9, 50, 63, 211, 289, 307, 313, 345–375 – Informative presupposition clefts 108, 231, 232, 237, 329, 334 – Inverted clefts 55 – Inverted pseudo-clefts 26 – it-clefts 1, 9, 10, 12, 13, 16, 17, 20, 25, 26, 35, 42, 45, 51, 90, 103, 105, 217,
416
Subject index
219, 220, 221, 222, 231, 233, 282, 326, 328–330, 333, 334, 335, 336, 338, 339, 340, 341, 342, 377, 379, 384 – Multiple foci clefts 114, 118 – Presentational clefts 9, 329, 336, 340, 351 – Pseudo-clefts 1, 3, 4, 9, 10, 11–15, 17, 18, 20–27, 29–36, 42, 43, 45, 49, 64, 79, 80, 83, 91, 103, 125, 139–176, 177–215, 220, 326, 377–414 – Pseudo-cleft sentences proper 11, 12 – Reverse clefts 4, 30, 31, 83, 85–88, 90, 91, 140, 158, 195, 198, 213, 289, 291, 307, 313, 316 – Reverse pseudo-clefts 1, 4, 9, 10, 11, 12, 17, 21, 24, 25, 27, 30, 32–34, 35, 42, 43, 49, 80, 92, 140, 147, 378 – Reverse wh-clefts 1, 9, 10, 12, 20, 25, 35, 42, 90, 91, 326, 326, 328, 333, 334, 378, 379 – Stressed focus clefts 108, 109, 237, 329, 334 – Temporal clefts 17, 63, 64, 87, 89, 108, 127 – Th-clefts 18, 20 – th-PCs 186, 187, 188, 189, 190, 191, 192, 193, 194, 196, 211 – Unstressed anaphoric focus clefts 108, 109 – w-clefts 377–414 – wh-clefts 1, 9, 10, 12, 13, 16, 20, 25, 35, 42, 45, 79, 90, 92, 103, 282, 326, 328, 330, 333, 334, 339, 340, 341, 377–414 – wh-PCs 186, 187, 191, 192, 193, 211 Cleft constituent 2, 4, 10, 11, 12, 23–25, 27–29, 31, 32, 35–38, 41, 42, 44, 45, 51, 62, 63, 64, 67, 70, 72, 78, 79, 80, 82, 83, 85, 87, 91, 93, 140, 141, 146, 148, 152, 153, 156, 170, 171, 174, 175, 177, 181, 196, 197, 199, 219, 223, 226, 233, 241–243, 245, 251, 253, 259, 262, 290, 293, 297, 298, 300, 315, 317, 381, 382, 384, 386, 390, 391, 394, 401, 403, 405, 410 Cleft clause 2, 10, 12, 13, 21, 23–32, 34–37, 39–41, 43, 44, 65, 85, 103, 104, 106, 108–112, 114–116, 118, 123, 125, 128, 130, 131, 133, 134 Compensation mechanism principle 5, 71, 94, 218, 266, 277, 315, 317
Complementizer 26, 35, 67, 104, 105, 108, 346, 392 Connective 219, 222ff., 235, 240, 259, 262, 297, 298, 299, 305, 346, 364 Contrastive 228, 412 – Contrastive accent 34, 401 – Contrastive constituents 410, 412 – Contrastive focus 38, 155, 333, 341, 401 – Contrastive function 121, 134 – Contrastive information 109 – Contrastive linguistics / studies 24, 35, 44, 52, 278ff., 281, 282, 314ff., 379 – Contrastive Topics 114, 115, 116, 134, 389 – Contrastive w(h)-clefts 410 Copula 9, 10, 12, 23, 27, 29, 31, 37, 43, 51, 63, 64, 67, 72, 82, 87, 91, 103–106, 141, 143, 151, 153, 157, 181, 219, 222, 234, 235, 237, 289, 291, 292, 325, 326, 336, 346, 347, 364, 366, 369, 381, 397, 400, 406, 409, 412 Copular sentences 103, 118, 183 Copulative sentences 140, 141, 143, 147, 149 Copulative verbs 151 Cost-benefit analysis 380 Corpora 2, 3, 50 – British corpus 81 – Comparable corpus 2, 5, 51, 52, 53, 60, 61, 139, 217, 218, 264, 278, 279, 281–284, 314–318 – C-ORAL-ROM 87, 88, 191, 328, 334 – CORIS-CODIS 52, 54, 193, 194 – EUROPARL 217, 218, 222, 223, 224, 225, 227, 234, 235, 236, 241, 243, 259, 264, 265, 266, 267, 268, 269, 270, 282, 293, 328, 329, 337, 340, 346, 348–349, 355, 363, 364, 370, 373, 377, 379, 380, 383, 384, 385, 387, 392, 393, 398, 400, 401 – FRANTEXT 52, 328 – ICOCP 10, 51, 52–62, 64, 67, 78, 80, 81, 101, 112, 139, 178, 190, 199, 278, 284, 291, 315; Italian subcorpus 59, 62, 78, 86, 192, 194, 201, 210; Swiss subcorpus 84, 86, 88, 89 – Il Manifesto corpus 278, 287 – Journalistic texts 3, 90, 101, 174, 177, 220, 277, 284, 291
Subject index
– LABLITA 88 – Le Monde diplomatique 278, 287, 346, 348–349, 363, 370, 373 – LIP 87, 88 – Monde diplomatique-Manifesto corpus 288, 289, 290ff., 314 – MR-BEAN 217, 225, 264, 265, 266 – Multilingual corpus 51, 53, 139 – Press Europ corpus 278, 288, 289, 306ff., 314 – SWISS-SMS 325ff; 330–332 – Translation / Parallel corpus 2, 5, 53, 195, 218, 267, 278, 281–284, 285, 290, 314–318, 346, 377, 379, 380 Equivalence 195, 279, 281, 300, 304, 379 – Functional equivalence 45, 174, 239 – Monoclausal equivalent 21 – Near-equivalence 379 – Non-cleft equivalence 300 – Semantic equivalence 103, 141, 188 – Structural equivalence 19, 30, 41, 112, 193, 279, 314 – Truth-conditionally equivalent 103, 403 Exhaustiveness / Exhaustive 4, 10, 237, 251, 341 Exhaustive operators 107 Focus 232, 237, 268, 329 – Anaphoric focus 231 – Emphatic focus 335 – Focalization / Focalizing 107, 108, 117, 118, 119, 140, 154, 155, 156, 157, 177, 232, 240, 267, 268, 277, 341 – Focalizer 153, 247, 248, 249, 250, 251 – Focus (of clefts) 10, 12, 42, 102, 103, 107, 109, 111, 113, 114, 115, 116, 118, 119, 132, 154, 157, 177, 196, 198, 221, 240, 348 – Focus marking / marker 232, 233, 237, 240, 256, 267, 268, 270, 328 – Focus particles 94 – Focus-background 101, 107, 109, 110, 111, 114, 116, 126, 134 – Information Focus 220, 232, 236 – Multiple foci 114, 115, 134, 329, 334
417
– Narrow focus 329, 334 – Sentence / Utterance focus 112, 154, 196 – Stressed focus 230 – Verum focus 401 Foreground 240, 257 Formality – Formal and informal registers 28, 30, 43, 44, 50, 54, 86, 88, 89, 201, 229, 264, 266, 328, 341, 342 – Clefts and levels of formality 224, 230, 234, 235, 291, 328, 330, 340 Generic – Generic (relative) form 24, 26, 28, 83 – Generic lexical head / noun 40, 41 – Generic noun / noun phrase 18, 19, 20, 21, 22, 23, 39, 40, 74, 82, 141, 142, 143, 144, 145, 146, 149, 180, 186, 187, 196, 346, 364, 366 – Generic pronoun 22, 41, 67 – Generic subordinator 235 Gerundive / Gerund 64, 22, 242, 243, 261, 262, 289, 294, 310, 317, 394 Given information 23, 107, 109, 110, 111, 131, 135, 155, 157, 163, 178, 199, 221, 230, 231, 237, 293, 337 – (Referential) givenness 101, 109, 110, 113, 114, 115, 116, 124, 384ff. – All-given 113, 114 – Co- / Contextually given 228, 245, 255 – Discourse-given 113, 209 – Given cleft (sub)clause 111, 116, 130, 134, 240, 268, 384 – Given cleft constituent 108, 114, 115, 116, 122, 123, 130, 132, 134, 240 – Given semi-active 240 – Relational givenness 384, 386 Grammaticalization 127, 212, 291 Highlighting 3, 5, 10, 23, 34, 92, 106, 107, 111, 114, 118, 119, 120, 121, 134, 156, 157, 208, 222, 252, 254, 300 horror aequi 396–397, 398, 399, 406, 407, 412
418
Subject index
Inferable information 109, 110, 114, 118, 155, 160, 231, 237, 240, 245, 253, 254, 255, 268 Inference 162, 168, 347, 351 Information structure 2, 101, 102, 103, 107–110, 114, 118, 123, 129, 134, 135, 139, 153, 154, 300, 304, 306, 326, 336, 338, 385, 388, 411 Languages – Armenian 345, 364 – Danish 1, 3, 5, 217–275, 285, 293 – English 1–3, 5, 10, 11, 13–16, 18–22, 24, 26–35, 37–41, 43–45, 49–51, 53–55, 58–62, 64–67, 71, 79, 81–84, 90–94, 101–109, 111, 112, 114, 116–118, 120–122, 125–127, 129–132, 134, 135, 141, 187, 198, 201, 211, 222, 223, 229, 230, 234, 236, 251, 267, 281, 282, 283, 285, 290, 291, 292, 308, 326, 328, 329, 331, 332, 339, 345–347, 364, 367, 372, 373, 377–384, 387–401, 403–406, 408, 410–413 – Finnish 93, 285 – French 1, 3, 5, 11, 13–17, 21, 25, 28–30, 32–34, 38–40, 49–56, 59–62, 64–67, 71, 73, 76, 79, 81, 83, 90–94, 139, 141, 143, 144, 149–153, 156, 159–162, 164, 170, 174, 175, 218, 222, 229, 259, 267, 269, 277, 278, 281–295, 300–302, 307–310, 312–315, 317, 326–337, 340–342, 345–349 – German 1, 3, 4, 5, 11, 13, 14, 18, 20, 22, 27, 28, 29, 30, 32, 34, 36, 38, 39, 40, 41, 45, 49, 50, 51, 53, 54, 55, 59, 60, 61, 62, 64, 65, 66, 67, 71, 79, 81, 83, 84, 90, 94, 95, 109, 177, 219, 220, 227, 267, 282, 283, 285, 290, 326, 327, 328, 329, 330, 331, 332, 333, 334, 336, 337, 339, 340, 341, 342, 345, 346, 364, 367, 368, 369, 377–414 – Germanic languages 1–4, 11, 41, 49, 51, 81, 84, 94, 340 – Hiberno-English 93 – Irish 50, 93 – Italian 1–5, 11, 13–17, 21, 22, 25, 29–31, 36, 38–44, 49–56, 58–62, 64–67,
71, 73, 77–79, 81–83, 84–89, 90–95, 101–109, 111–112, 114, 116–118, 120, 121, 122, 125–127, 129–132, 134, 135, 139, 142–145, 147–151, 153, 155, 156, 159, 160, 162, 174, 175, 177, 178, 181, 182, 184, 186–197, 199, 201, 206, 208, 209–213, 217, 218, 224, 225, 229–232, 234–236, 238, 241–245, 247–249, 251, 253, 254, 256, 258–262, 264–270, 277, 282, 283–293, 299, 300, 302, 304–311, 313–318, 325–334, 336, 339–342, 345, 346, 347, 364, 366, 369, 372 – Neo-standard Italian 50, 54, 86 – Polish 94, 283 – Portuguese 5, 50, 51, 267, 282, 285, 290, 328, 329 – Romance languages 1–4, 11, 14, 15, 17, 19, 21, 26, 27, 36, 41, 49, 50, 51, 67, 81, 84, 93, 94, 109, 229, 327, 336, 340, 345, 367, 373 – Russian 94, 345, 364 – Spanish 3, 4, 5, 11, 13–17, 21, 25, 26, 29, 30, 32–40, 45, 49–51, 54, 55, 59, 61, 62, 64, 65, 66, 67, 73, 78, 79, 81–84, 93, 94, 177–215, 232, 267, 282, 283, 285, 290, 326, 328, 329, 330, 346, 357, 364, 365, 372 – Relay / Pivot language 53, 283, 288, 378 – Romanian 93 – Swedish 93 – Swiss German 1, 3, 327, 330, 331, 333, 334, 335, 336, 337, 339, 340, 341, 342 – Swiss Italian 84–89 – V2 language 5, 94, 218, 269, 387, 388, 392, 393, 395, 396, 412 Lexicalization 191, 337, 338, 340 Medial position 12, 51, 93 – Medial cleft constituent 11, 37, 38, 41, 42, 62, 63, 64, 67, 78, 79, 80, 85, 101 Modal – Modal 404, 405, 406, 409 – Modal particle 94, 330 – Modal source 391 – Modal verb 222, 227, 393
Subject index
Monoclausal proposition / sentence / structure 9, 21, 22, 42, 72, 102, 106, 127, 146, 154, 326 Multiple Correspondence Analysis 381, 407ff. Object 94, 104, 105, 181, 227, 241, 243, 244, 248, 256, 257, 265, 294, 296, 304, 310, 317 – Direct object 146, 147, 150, 219, 223 – Indirect object 5, 219, 223, 225 – Prepositional object 28, 234, 241, 242, 243 Particles: modal 94, 330 Presentative – Presentative element 326 – Presentative function 197, 198, 199, 212 – Presentative / Presentational structures 77, 351 Pronoun – Free relative pronoun 13, 14, 15, 17, 34, 39, 84, 144, 179 – Fused relative pronoun 13, 80, 141, 143–145, 147, 149, 179, 186, 187, 188, 211 – wh-Pronoun 13, 41 Presupposition 103, 106, 108, 123, 154, 155, 156, 159, 231, 232, 237, 251, 268 Punctuation 157, 349 Recoverability 109 Reported speech 23, 55, 87, 88, 89, 204, 208, 213 Relative clause 4, 10, 13, 18, 65, 73, 77, 82–84, 140, 141, 144, 145, 146, 149, 152, 160, 178, 179–184, 187, 188, 193, 206, 211, 221, 222, 244, – Free relative clause 75, 82, 144, 381, 400 – Restrictive relative clause 92, 103, 143, 250 Specificational 9, 21, 22, 63, 64, 103, 118, 141, 143, 147, 151, 157, 180, 182–185, 196, 206, 209, 381, 400 Subject function 94, 146, 147, 149, 150, 256 – Clausal subject 395, 396, 400, 405, 406, 407, 409, 410, 412 – Dislocation of the subject 30
419
– Empty / Expletive / Dummy subject 37, 103, 237, 326, 336 – Implicit / Null subject 197, 200, 202, 203, 204, 205, 206, 207, 212 – Focalizing / Rhematisation of subject 221, 270 – Subject inversion 5, 159, 328 – Subject (relative) clause 4, 158 – Subject in cleft 17, 20, 72, 104, 108, 110, 114, 174, 181, 219, 221, 222, 223, 224, 227, 229, 235, 241, 243, 248, 249, 265, 267, 269, 270, 289, 291, 293, 294, 297, 298, 302, 304, 309, 310, 311, 315, 316, 317, 338, 381, 384, 394, 405ff. Spoken / Oral communication 5, 11, 14, 43, 50, 51, 54, 57, 67, 83, 86, 87, 88, 89, 93, 141, 149, 177, 191, 213, 219, 220, 225, 226, 229, 230, 245, 253, 263ff., 291, 304, 328, 332, 340, 345, 348, 366 Taxonomy – Cleft constructions taxonomy 9–48, 51, 140, 181 – Cleft sentences taxonomy 109, 110, 115 Text / Discourse – Text messages 4, 327, 328, 330, 331, 332, 333, 334, 335, 337, 339–342 – Text structuring 221, 231, 259, 270, 349 – Textual / Discourse coherence / cohesion 132, 172, 198, 259, 267, 268, 269, 277, 297, 326 – Textual functions 131, 134, 139, 156, 159, 161, 164, 168, 173, 174, 175, 177, 181, 190, 196, 198, 212, 217, 232, 241, 245, 266, 353 – Textual referent 198, 208 Topic – (Sentence) Topic 121, 122, 198, 254, 268, 384, 385, 386–388, 389–391 – Contrastive Topic 114, 115, 116, 134, 389 – Discourse / Text topic 131, 132, 135, 157, 160, 164, 165, 175, 198, 199, 201, 204, 267 – Frame Topic 111, 127, 129, 134 – Thematic organization 177, 198, 207, 212 – Topic (re)launching 121, 122, 198
420
Subject index
– Topic and Comment in clefts 110, 114, 121–122, 125, 162, 199–210, 212, 384, 398, 399, 401, 410, 412 – Topic shift / shifting 125, 221, 247, 254, 267, 268 – Topical / Topicality 238, 395 – Topicalization 329 – Topic-clause clefts 109 – Topic-Comment structure 384–386, 387, 391, 411 Translation – Covert translation 94, 279–280, 285 – Free translation 285, 286, 288 – Inferential clefts in translation 364ff. – Invisible translation 58 – Literal translation 346 – Machine translation 218 – Translationese 218, 278, 279–281 – Translated texts 50 – Translation corpus 5, 52, 195, 213, 278, 279, 281–286, 290, 306, 314–318, 379, 380 – Translation studies 278, 279 – Translation universals 278, 279–281 Typology – Cleft sentence typology 110, 116, 134, 232, 233, 236–241, 268 – Language typology 52, 90–93 – Pseudo-cleft typology 177, 178, 185–190
wh-clause 12, 386, 390, 399, 401, 403–410 Word order 94, 218, 270, 277, 311, 328, 387 Writing 90, 93, 328, 349 – Academic writing 54 – Journalistic writing 51, 59, 60, 95, 101, 177, 191, 201, 212 – Legal writing 54 – Literary writing 191 – Semi-spontaneous writing 56 Written – British written texts 50 – Written Danish 223, 225 – Written discourse / language / texts / communication 9, 34, 45, 51, 53, 55–58, 86, 87, 91, 92, 177, 191, 219, 230, 264, 265, 266, 328, 345, 348, 373 – Written English 93 – Written French 50, 288, 328 – Written German 50 – Written Italian 49, 52, 54, 77, 84, 87, 194, 229, 287, 288, 291, 309 – Written mass media 54 – Written Swiss German 331 – Written varieties 43, 86