Towards a typology of focus: Subject position and microvariation at the discourse-syntax interface

Ampersand 2 (2015) 49–60 Contents lists available at ScienceDirect Ampersand journal homepage: www.elsevier.com/locate/amper Towards a typology of focus: Subject position and microvariation at the discourse–syntax interface Ángel L. Jiménez-Fernández∗ University of Seville, Departamento de Lengua Inglesa, C/ Palos de la Frontera s/n, 41004 Sevilla, Spain highlights •Preverbal subjects in constructions involving focus fronting are possible in some varieties of Spanish. •These preverbal subjects are interpreted as Given Topics, which move to Spec-TP. •Fronting operations connected with focus include Contrastive Focus and Mirative Focus, and Quantifier Fronting. •Resumptive Preposing is an instance of Aboutness-Shift Topic. article info Article history: Received 29 November 2014 Received in revised form 15 February 2015 Accepted 5 March 2015 Available online 16 April 2015 Keywords: Contrastive focus Mirative focus Resumptive preposing Quantifier fronting Preverbal subjects Topics abstract In this work I explore the different discourse–syntax interface properties of focus fronting in Standard Spanish (SS) and Southern Peninsular Spanish (SPS) including Andalusian and Extremaduran varieties. In SSitistaken forgranted thatin focusfronting theverb isobligatorily adjacentto thepreposed constituent. I show that this is not the case in SPS, where this condition is optional. I carry out an analysis of three types of foci which involve movement to the left periphery (contrastive focus, mirative focus and quantifier fronting) and one type of topic (resumptive preposing). Discourse, syntactic, and semantic properties are taken into account to illustrate this typology. Crucially, only contrastive and mirative focus contexts allow for preverbal subjects in SPS, which are proposed to be Given Topics in this variety. On the other hand, resumptivepreposing is showntoentail acaseoftopic fronting.Iuse differentexperimentswithempirical data and judgements by native speakers to test my proposal that focus-verb (or topic-verb) adjacency is subject to microparametric variation in Spanish. ©2015 The Author. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 1. Introduction1 This paper explores the discourse–syntax interface properties of the different constituents involved in focus fronting constructions in Spanish. As is well known, one of the main traits of focus fronting is that the V(erb) must be adjacent to the focused constituent in this type of language. Building on Roberts (2012) and Biberauer et al. (2010), I argue that this condition is subject ∗Tel.: +34 954551546; fax: +34 954551516. E-mail address: [email protected]. 1Earlier versions of this paper have been presented at the Workshop on Syntactic Variation at the Universitat Autónoma de Barcelona and at the LAGB meeting at SOAS (London). I wish to thank the audience there for comments and feedback. In particular, I would like to thank Mara Frascarelli, MaLluïsa Hernanz, Ana Ojea, Mercedes Tubino, Aritz Irurtzun, Antonio Fábregas and Silvio Cruschina for fruitful and insightful discussion. The research carried out in this work has been partially supported by the Spanish Ministry of Economy and Competitiveness (FFI2013- 41509-P). to microparametric variation as some varieties of Spanish (Southern Peninsular Spanish, SPS), especially the Andalusian and Extremaduran varieties, seem to obviate the adjacency condition. I discuss the syntax of different types of focus (Contrastive and Mirative foci) alongside Quantifier Fronting and Resumptive Preposing, all of which crucially hinge on the discourse interpretation of pre- or post-verbal subjects. ItiswidelyacceptedthatlanguagessuchasSpanish(asopposed to English, but along with Catalan and Romanian) instantiate an adjacency condition in focus fronting constructions (Rizzi,1997; Uribe-Etxebarria,1991;Uriagereka,1995; among others) as the contrast in (1–2) illustrates: Standardly, it is assumed that V must be adjacent to the focused constituent (Zubizarreta, 1999: 4241; RAE-ASALE, 2009/2011), http://dx.doi.org/10.1016/j.amper.2015.03.001 2215-0390/©2015 The Author. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4. 0/). 50 Á.L. Jiménez-Fernández / Ampersand 2 (2015) 49–60 which describes the unacceptability of (2a). In more theoretical terms, this adjacency condition has been explained by proposing movement of Tense to Complementiser (hereafter, T and C respectively) – or T to the designated category Focus in cartographic analyses –, after V’s movement to T (Rizzi,1997;Barbosa,2001). Descriptive surveys such as the one carried out by RAE-ASALE do not take into account different types of focus and include all kinds of fronting within a single group regardless of the distinct salient properties at the levels of interpretation and intonation. In this group phenomena such as the following are conflated: Contrastive Focus (as in (2a)), Mirative Focus (cf. Cruschina, 2012, inspired by the use of the notion of mirativity by DeLancey (1997,2001)), Resumptive Preposing (cf. Cinque,1990; Cardinaletti,2004;Leonetti and Escandell, 2009), and Quantifier fronting/Negative Preposing (cf. Âmbar,1999,2003;Barbosa, 2001;Bosque,1980;Quer,2002).Researchershaveagreed, though, that in all these cases the adjacency condition applies obligatorily. Example (3a) illustrates Contrastive Focus, (3b) exemplifies Mirative Focus, (3c) instantiates Resumptive Preposing and (3d) is a case of Quantifier Fronting2: Studies in generative grammar have emerged that distinguish some of these information-structure phenomena (Bianchi,2012; Cruschina,2012;Haegeman,2012). In Spanish, some works have identified types of focus other than contrastive focus (Torrego, 1980; Uriagereka,1988;Quer,2002;Gallego,2007;Leonetti and Escandell, 2009), yet all these analyses agree that T-to-C is compulsory in Spanish focus fronting. In this work, I first show that interpretive and syntactic properties can be used to establish a more accurate typology of focus. Based on data from Spanish and especially its Southern variety, I propose that the adjacency condition is subject to microparametric variation in that, depending on the type of focus, preverbal subjects are readily found in some varieties of Spanish (Andalusian, Extremaduran). My proposal is crucially grounded in an experiment carried out with native speakers in which they had to judge the grammaticality of fronted focus constructions. In this experiment, speakers were presented with data in a randomised order, and they were given the context in which they had to correctly place the sentences that follow (a full description of the experiment is provided in Section 3). The informants were divided into two different groups. The first group includes areas from northern SpainandMadrid,whereasthesecondgroupcomprises the south.3 The word order variation detected in the different discourserelated phenomena is accounted for in my analysis by proposing criterial features in dedicated categories which trigger 2The examples in (3) are just used for presentational purposes to show a sample of the type of data I discuss in this work. A fully detailed description of each type of fronting is given in Section 2. 3Regarding V-adjacency in Catalan, Quer (2002, 254–255, fn.3) suggests that there may be dialectal variation since for some speakers this condition can be dispensed with (see also Vallduvi, 1993). movement of a given constituent to a specific syntactic (but discourse-linked) position in the left periphery. I assume a split Complementiser Phrase system where specific designated categories like Force, Topic and Focus project (Rizzi, 1997, et subseq.). I hold that in the type of focused constructions studied here subjects may be topics, specifically Given Topics – also called Familiar Topics – (in the sense of Frascarelli and Hinterhölzl, 2007;Bianchi and Frascarelli, 2010;Frascarelli and Jiménez-Fernández, 2013). A Given Topic entails shared information which is familiar to everybody in a relevant context, as illustrated in (4), a dialogue between two friends (A and B) in a restaurant: I assume, with Jiménez-Fernández (2011), Frascarelli and Jiménez-Fernández (2012), Jiménez-Fernández and Işsever (2012) and Jiménez-Fernández and Miyagawa (2014), that subjects which function as Given Topics in discourse may move to a Tense Phrase (TP) internal position in Spanish, thereby justifying the pattern Focus+Subject+Verb Phrase. Moreover, in this type of focused construction the head Focus may not trigger movement of V in SPS, contrary to SS. As will become clear below, this happens with Contrastive Focus, Mirative Focus and Quantifier Fronting. On the other hand, as in Italian (Cardinaletti,2004,2010;Benincà and Poletto, 2004), in Resumptive Preposing (which is identified as a subtype of topic fronting) V must obligatorily move to the dedicatedheadTopicintheCP areainall varietiesofSpanish,hence there is no slot for subjects in between the moved constituent and V. I start off with two working hypotheses: (1) The different types of focus are encoded in the lexicon as syntactic features, which will reveal that CF, MF and QF are instances of focus whereas RP is shown to be a type of topic. (2) Different types of focus (and topic) fronting may display different behaviour with respect to the relative position of subject and verb both across languages and within a single language. Within Spanish, a crucial difference as regards Verb-to-Focus is detected between SPS speakers and speakers of other varieties of Spanish, thus pointing to a microparametric distinction. The paper is organised as follows. In Section 2I discuss the nature of the fronted focus constituent. I show its interpretive and grammatical properties which I argue demonstrate that CF, MF and QF are foci but RP is an Aboutness-Shift Topic (AS-Top), thereby validating my hypothesis 1. In Section 3the methodology used in my empirical approach to the typology of focus is presented alongside sample examples of data that the informants had to judge. Section 4discusses the distribution of foci in SPS and SS with special reference to the pre- or post-verbal position of subjects, confirmingthe microparametric variation suggested by hypothesis 2. Section 5comprises empirical evidence in favour of analysing pre-verbal subjects as Given Topics (G-Tops) in SPS. I show that pre-verbal subjects are G-Tops moving to Spec-TP in SPS. Finally, Section 6concludes the paper and offers some further issues concerning types of focus which I will tackle in my future research. 2. The focused constituent: a multifactorial approach to different types of Focus The term Focus is often used to refer to phrases serving two discourse functions, namely (a) to introduce new information, which is known as Information Focus (IF), and (b) to introduce a contrast with respect to a previous assertion by denying one part and proposing another part. The latter is what is typically referred Á.L. Jiménez-Fernández / Ampersand 2 (2015) 49–60 51 to as Contrastive Focus (CF). Based on Kiss (1998), this distinction is exemplified in (5) for Hungarian. A number of scholars have argued in favour of a clear-cut distinction between these discourse categories, based on morphological, syntactic, phonological and discourse factors (cf. Kiss,1998; Zubizarreta,1998;Nespor and Guasti, 2002;Donati and Nespor, 2003).Cross-linguisticevidencesupportsthe necessity of a distinction between different types of focus, which is syntactically encoded (cf. Molnár,2006;Bentley,2007;Cruschina,2012;Bianchi and Bocci, 2012;Bianchi,2012). To illustrate, the syntactic position of CF and IF tend to be different across languages. Kiss (1998) discusses the different properties which are distinguished by two positions. Hence in Hungarian a preverbal focus is interpreted as contrastive, as in (5a), whereas a postverbal focus is assigned a purely informational reading, as in (5b)4: In Romance languages such as Italian and Spanish, it is commonly assumed that the left periphery is reserved for CF and the positionofIFissentence-internal(typicallyoccurringinpostverbal position). This is the view defended by Cruschina (2012) for Italian and Spanish, as illustrated in (6) from Zubizarreta (1998) (boldtype indicates IF): ThispositionaldifferenceworksforStandard Spanish.However, again some microparametric variation is detected in other dialects since the reply in (B′), with fronted IF, is completely natural in some varieties of Spanish. Interestingly, Southern Peninsular Spanish is one such variety, showing the availability of a leftperiphery position for purely IF, as shown by Jiménez-Fernández (in press), which constitutes another piece of evidence supporting the microvariation in SPS. In addition to syntactic properties, from an interpretive point of view, the different types of foci differ in both discourse and prosodic differences (Cruschina,2012;Frascarelli and Ramaglia, 2013). Thus, they are interpreted differently at the interfaces. Sticking to the distinction between CF and IF, Kiss (1998) claims that the contrastive interpretation is absent in IF. These are the readings corresponding to (5a) and (5b) respectively. In addition, the prosodic properties of focus will be different depending on the type of interpretation. In this connection, Belletti (2004: 43, note 12) describes CF as carrying an emphatic stress which is missing in IF. I come back to the distinct discourse and syntactic properties of focus below (though I leave aside a thorough study of the prosodic characterisation of foci; see Bianchi et al., 2014 (in press), for intonational evidence from Italian, and Jiménez-Fernández, 2013 for prosodic evidence from Spanish). Within generative grammar two different semantic approaches can be distinguished with respect to the semantic characterisation 4As pointed out by an anonymous reviewer, Hungarian exhibits a very specific type of contrastive focus, namely exhaustive focus. Therefore, the interpretation of thefocusedDPin (5a)involves ahat andonly ahat (nothingelse). Itmust beclarified that Spanish CF is not always exhaustive, and hence the corresponding sentence in Spanish does not entail that it is only a hat that Mary picked, as illustrated in (i): For the exhaustive use of CF in Spanish, see Gutiérrez-González (2010). of focus. On the one hand, there is the Alternative Semantics approach (Rooth,1992;Beaver and Clark, 2008). Under this view, focus generates a set of alternatives. For example, for a question suchas ‘‘What does Mary want?’’, there is anumber of propositions which vary in the content provided by the focused direct object ({Mary wants coke, Mary wants ice-cream,...}). Any of these sentences satisfy the information request in the question and the whole set of possible propositions make up what is known as congruent answers to the question. On the other hand, in the Structured Meaning approach (Krifka, 2006) the proposition is divided into two parts, namely a focus (e.g., ‘ice-cream’ in the example above) and a background (the presupposed denotation of the rest of the clause, i.e. the property of being something that Mary wants). In my characterisation of focus, I take into account the most salient defining parameters of both approaches. I pay attention to the set of alternatives generated and also to the information structure partition of the proposition in terms of focus and background. When the different types of focus are addressed, we are confronted with different semantic operations, which are reflected in the syntactic derivation in that, depending on the semantic interpretation, each type of focus is assigned a specific sort of discourse feature. These semantic properties are the cornerstone of the description of different types of focus fronting in the sections which follow. These features will constitute the basis to distinguish between true cases of focus such as CF, MF and QF and the ‘fake’ type of focus represented by RP, on the one hand, but also will help us elucidate the differences detected among foci, on the other hand. 2.1. Contrastive focus AccordingtoGussenhoven(2007:91), Contrastive Focus is‘‘[...] a constituent that is a direct rejection of an alternative, either spoken by the speaker himself (‘not A, but B’) or by the hearer [...]’’. Contrast,thus,implies removal ofinformationandinsertion of new information, as illustrated in (7). As a reply to (7A), B is understood as rejecting part of the previous assertion (con Ángela) and proposing a new piece of information to fill in this gap (con Jimena). Therefore, the meaning of (7B) is ‘It is with Jimena that Susana is going on vacation, not with Angela’. Note that, as previously stated in footnote 4, CF in Spanish is not necessarily exhaustive (as opposed to languages such as Hungarian). Hence the interpretation in this reply is that Susanawent with Jimena and not with Angela, but maybe she went with somebody else. In terms of the focus-presupposition articulation, Contrastive Focus implies a two-part dissection and the set of alternatives is very restricted (constrained by the semantic properties of the rejected item). In example (7) the set of alternatives includes two members, namely Angela and Jimena (maybe more members can be added to the set, though either the speaker is not interested in this addition or he/she does not know of any other specific members). 2.2. Mirative focus Inspired by DeLancey (1997,2001),Brunetti (2009) and Paoli (2009), Cruschina (2012) consider Mirative Focus as not merely 52 Á.L. Jiménez-Fernández / Ampersand 2 (2015) 49–60 informative. It provides new information and, based on the speaker’s knowledge of what the hearer expects, indicates that such information will be unexpected. Thus, unexpectedness and surprise with respect to some event are involved in this type of focus (cf. Bianchi et al., 2014): In example (8) the DP tres trozos de tarta ‘three pieces of cake’ has been preposed to indicate the speaker’s surprise and unexpectedness. Âmbar (1999: 41) terms a similar kind of fronting in Portuguese an ‘‘evaluative construction’’, involving the fronting of an emphatic and evaluative element to a dedicated functional projection (i.e., the Evaluative Phrase) in the left periphery. Âmbar associates this evaluative construction with an exclamative meaning and exclamative intonation. I take it that in Mirative Focus the speaker shows the unexpectedness of its content for himself/herself. A plausible way to test the mirativity of this type of focus is the addition of an expression such as ‘And it doesn’t surprise me’. In cases of sentences containing MF, the result is predicted to be pragmatically anomalous (Bianchi et al., in press). Let us consider (9) uttered by the same speaker: As is clear, the speaker is showing his/her surprise so it is incredible that these people have planned to go to the Caribbean. In this context it is pragmatically odd adding the information conveyed by Pero eso no me sorprende. In other words, in example (9) the adverbial al Caribe ‘to the Caribbean’ is clearly an instance of MF. Mirative Focus is not dependent on a question–answer context. Contrast is established with an element that is part of the shared knowledge of the participants and can be semantically characterised as a ‘‘proposal to negotiate a shared evaluation’’ (Bianchi, 2012). Due to the unconstrained unexpectedness involved in Mirative Focus, the set of alternatives is very large. This is totally predicted since in an example such as (9) al Caribe is just one option among many others. 2.3. Contrastive vs. mirative foci: some interpretive and syntactic differences In this section I discuss different interpretive and syntactic properties that can be employed to discriminate between CF and MF. According to Bianchi et al. (2014), in mirative contexts, the target sentence is an assertion and the context forces the mirative connotation. This is illustrated in (10), adapted from Bianchi et al. (2014), their example (8). On the contrary, instead of an assertion the sentence with CF is a reply to a previous assertion, which corrects one piece of this previous assertion. The corrected part can explicitly be mentioned in a negative coda, as illustrated in (11), also adapted from Bianchi et al. (2014). Note that MF is compatible with this negative coda, whereas CF is not, as shown by the impossibility of adding a negative coda to the assertion in (12), where the context favours the mirative interpretation (the symbol # stands for pragmatically anomalous): This also proves that MF is not contrastive. At least it is not as strongly contrastive as CF, which explains the incompatibility with the negative coda. As stated earlier, another distinction at the level of interpretation is the co-occurrence of MF with an additional expression contradicting the surprise and unexpectedness of the speaker of the previous assertion, as opposed to CF. This is further illustrated in (13) for MF and (14) for CF: As clearly illustrated in (13), MF cannot be followed by a construction conveying expectedness, whereas CF can naturally co-occur with such an expression. The reason is that CF has nothing to do with surprise or unexpectedness, thereby being compatible with an expression exhibiting either surprise or no surprise. It is just not relevant. Let us now turn to the syntactic properties. Cruschina (2012) and Bianchi et al. (2014) suggest two main distinguishing properties that describe the behaviour of CF and MF. First, based on data from Italian, these authors claim that V-adjacency is optional for CF whereas it is obligatory for MF. Á.L. Jiménez-Fernández / Ampersand 2 (2015) 49–60 53 In the present article, the main proposal is that in some Spanish dialectsV-adjacencyis simplyanoption,andthispropertydoes not discriminate between CF and MF. More precisely, we will see that in Standard Spanish both CF and MF require V-adjacency thereby blocking the interpolation of any adjunct. However, the situation is different in Southern Peninsular Spanish, where neither CF nor MF ban the occurrence of an intervening constituent. Let us consider CF in (17) in both varieties, SS and SPS: V-adjacency is obligatory in SS, as shown in (17). However, in SPS (especially in spontaneous speech) the situation is completely different and speakers produce both syntactic orderings: This is indicative that the V-Adjacency property is nullified in SPS and cannot be used to discriminate between CF and MF since again the V-Adjacency requirement is obviated in SPS for MF, as illustrated in (19) and (20): The possibility of interpolating a constituent such as an adjunct isfeasibleinSPS. In latersectionIwill show thatapreverbalsubject is also possible in this variety when the fronted phrase is MF. The second syntactic property used by Cruschina (2012) to distinguish between CF and MF is the possible occurrence of CF in an embedded sentence complement of a verb of saying and the unacceptability of MF in this syntactic context: This is a good test to apply in Spanish and will help us differentiate CF from MF regardless of the dialectal variety. The focus fronting in (24) shows that MF is unavailable in the embedded sentence, whereas (25) tells us that CF does not exhibit such a restriction. Note, incidentally, that the preposed constituent in (24) could be interpreted in contrast with other members of the set (for example, one bottle or three bottles of vodka), and hence it would be regarded as well-formed, which points to the fact that CF is entirely compatible with embedding. 2.4. Quantifier fronting Leonetti and Escandell (2009) suggest two defining properties of Quantifier Fronting constructions: (1) Fronted Quantifier Phrases do not show any emphatic intonation, in contrast with other types of focus (though they do not carry out an actual examination of intonation patterns in real data); and (2) the interpretation of the fronted element does not involve a contrast with other alternatives (Quer, 2002)5: Theexamples in(26)showthatnocontrastcanbeestablishedin QuantifierFrontingbetweenthe fronted constituentandamember of a discourse set due to the fact that polarity is the core of the contrast and as such it is the emphasised element. Two points are at issue here. First, the fronted QP does not display any contrast at all with any possible alternative, thereby disconfirming its discourse role as CF. Secondly, QF is an instance of focus, and as such it is incompatible with another focused element in the same sentence, illustrated in (27a), and does not allow the insertion of the resumptive clitic, as shown in (27b). Thus QF cannot co-occur with a narrow in-situ focus, as Quer (2002) points out: Interestingly, Leonetti and Escandell (2009) remark on the adjacency of V and the fronted element as a common property of all types of fronting except Clitic Left Dislocation, including Quantifier Fronting. For Leonetti & Escandell, Quantifier Fronting is an instance of Verum Focus Fronting,6in which the Information Structure partitions the proposition in two parts, focus and background. Focus corresponds to polarity, hence the implication of negative and positive polarity in (26a–b) respectively. The alternatives reduce to two: positive and negative. This raises the question as to why Verum Focus (VF) and QF can co-occur in the same sentence if we accept that there is one focus per clause (Rizzi, 1997). This situation predicts the ungrammaticality of sentences 5The focus status of Quantifier Fronting is quite tricky since a QP can be dislocated to the left periphery with the function of Contrastive Focus (i) or Mirative Focus (ii), or even as a CLLD-ed topic (hence the fronted QP is interpreted as specific and a resumptive clitic is available). From this perspective, it may be thought that QF should not be singled out as a distinct type of focus, in line with Cinque (1990). However, in my work I will adopt the view pursued by Quer (2002), Cruschina (2012), according to which QF stands on its own as a well-defined type of focus, with its own syntactic and semantic properties. QPs are ambiguous as for their specific or non-specific reading (Suñer, 2003). I take true QF to involve only non-specific QPs (including bare Qs such as nada ‘nothing’ or algo ‘something’). 6In line with Höhle (1992), I take Verum Focus as an instance of emphatic operation by which a proposition is connected with the truth value of the sentence, which can be paraphrased as ‘It is true that/It is the case that...’. Therefore, an example such as (8a) is interpreted as ‘It is the case that I have nothing more to add’ (see also Romero and Han, 2004). 54 Á.L. Jiménez-Fernández / Ampersand 2 (2015) 49–60 suchas(27) since theyincludetwofoci, namelyVFandQF,contrary to facts: At this point I would like to connect the co-occurrence of VF and QF with the notion of broad focus or all-focus sentences. As is commonly assumed, broad focus is involved in out-of-the-blue sentences since all the information provided is new. These sentences show that there is no incompatibility between broad focus and QF, which in fact proves that not all frontings are simple instances of narrow focus (in the traditional sense). Cruschina (2012) and Bianchi et al. (2014) have identified a similar paradoxical situation with the compatibility of Mirative Focus and Broad Focus in Sicilian: MF may be used in out-of-the-blue sentences, where the answerisgivenbyanall-focus sentence. Hence, it seems that in allfocus sentences there may be fronting if the constituent conveys an additional connotation (surprise or unexpectedness). In Spanish we find similar data for MF: B’s answer in (30) displays new information, yet Spanish selects a DP as the most relevant piece of information since it is what expresses the speaker’s surprise. One of the differences between QF and MF is that polarity is affected only in QF instances. My suggestion is that in all-focus sentences one piece of the information is more salient and it is precisely this piece of information that is preposed. This means that in QF we may have no narrow focus, but just a vacuous fronting in a sentence in which what is emphasised is the truth or falsity of the whole proposition, i.e. Verum Focus. To put it bluntly, this is exactly the connection between the polarity of the sentence and QF. The informationconveyed bythesentenceisnewandamongthepieces of information one of them, represented by QF, is highlighted. This QF may be viewed as a type of altruistic movement (Erteschik-Shir, 2007), since there is no real need for fronting to express that focus lies on the polarity. 2.5. Resumptive preposing Cinque (1990: 87) holds that in RP ‘‘the fronted phrase must either directly resume an identical phrase in the immediately preceding discourse or be inferentially linked to such a phrase’’. This is illustrated for Italian in (31): On the other hand, Leonetti and Escandell (2009) claim that, contrary to Quantifier Fronting, Resumptive Preposing (RP) occurs mostly with definite Determiner Phrases (DPs). These linguists argue that ‘‘Resumptive Preposing shows every diagnostic of whmovement: the fronted constituent has to be adjacent to the verb, which triggers subject–verb inversion, and it is incompatible with any other instance of wh-movement such as, for instance, Focus movement’’ (Leonetti and Escandell,2009: 167). Again the V adjacency condition is emphasised as a defining property of Resumptive Preposing. This is shown in (32), Leonetti & Escandell’s example (4): In addition, another property of RP is that the fronted constituent does not require any resumptive clitic, thereby patterning with focus fronting (as opposed to CLLD). I return to this issue below. Leonetti & Escandell group Resumptive Preposing and Quantifier Fronting together. The authors adduce that the two constructions share a common property, that of expressing polarity focus/Verum focus (Höhle,1992;Krifka,2007;Féry,2007). The alternatives would thus be most restricted: positive or negative. However, in light of the data from SPS shown in Section 2, this is untenable, and I stick to the distinction made by Cinque (1990) andQuer(2002).Oneofthereasonsforkeepingtothedistinctionis that, as Leonetti & Escandell argue, Resumptive Preposing involves just definite elements as in (32). However, this is not a constraint for Quantifier Fronting, as shown in (10). Leonetti & Escandell claim that no focus-background partition of the sentence is attested. Instead, the construction exhibits a focus-background partition of the sentence polarity. This may be true for Quantifier Fronting, as stated in the previous section. Nevertheless, for Resumptive Preposing it is not the case that the whole proposition is the background. In this connection, ‘‘the VFF [Verum Focus Fronting] construction is accepted only when its whole propositional content is background information, i.e., has been mentioned in the previous discourse’’ (Leonetti and Escandell, 2009: 183). This is shown in (34): 7 This type of fronting clearly contrasts with true Resumptive Preposing constructions: 7Note that the time of fronting illustrated in (34) patterns together with VP- Preposing in languages like English (see the gloss for (34)). I assume that in this type of fronting we do have a true instance of VFF, as opposed to RP, which I will argue does not exhibit any emphasis on the polarity. Á.L. Jiménez-Fernández / Ampersand 2 (2015) 49–60 55 From a discourse perspective, this sentence is felicitous as a response to a previous proposition such as El partido de la oposición ha propuesto cambiar la ley anticorrupción (‘The opposition party has proposed to change the fraud law’). Hence, there is no requirement that the whole proposition is background information. Another problem with a Verum Focus Fronting analysis of Resumptive Preposing is that contrasts can be established, thereby indicating that it is not the case that the polarity is emphasised. To the contrary, an explicit contrast with possible members of a discourse set is attested in sentences such as (36): Finally, to prove that VFF is not involved in RP, let us consider the test that Leonetti & Escandell propose to demonstrate that RP is polarity focus. The authors claim that sentences with QF and RP can be paraphrased by the same sentence without fronting but inserting the polarity-inducing element sí (que) ‘yes (that)’. If the two constructions with and without fronting have the same interpretation, this demonstrates that the version with fronting involves polarity focus. Let us start with QF: As is clear, the truth value of both sentences with and without the polarity-inducing particle is exactly the same, confirming that QF instantiates Verum Focus Fronting. Now, let us consider RP in light of the following paragraph: In the last sentence, the DP esas mismas cosas ‘those same things’ has been preposed and this involves RP. If Leonetti & Escandell are correct, we predict that the relevant sentence should be paraphrasable with the same sentence without fronting but including the polarity-inducing particle, contrary to facts: As a continuation of the previous sentence, in this context the truth value of sentence (39) is quite distinct from the corresponding sentence with fronting in (38). In other words it is not the case that we are emphasising the polarity of the sentence when RP is at issue. The question thus arises as to what type of discourse category RP represents. Resumptive Preposing partially conveys [+given] information, as illustrated in the above example, since it presupposes information already provided in the context. Thus, I will assume that it qualifies as a topic. Additionally, RP expresses some new information as well. For example, in (38) the preposed DP implies that those same things are just similar to the ones mentioned in the previous sentence. This combination of given and new information leads me to propose, following Jiménez-Fernández and Miyagawa (2014), that RP is an instance of Aboutness-Shift Topic in the sense of Frascarelli and Hinterhölzl (2007);Bianchi and Frascarelli (2010). This AS-Top serves the function of proposing what the sentence is about by providing some additional information with respect to already shared information. As for the Information Structure partition, there is a topiccomment division, and the comment includes information focus (for example, el Partido Popular ‘the Popular Party’ in example (36) or él con los primos León ‘he and the Leon cousins’ in (38)). ForreasonsofspaceIwill justprovidewith twoempiricalpieces of evidence in favour of dealing with RP as topic fronting. The Should I stay or should I go? first property which characterises RP is that it involves the feature [+given]. However, this feature can also be present in CF, as an anonymous reviewer points out to me. In my view, both CF and RP may include [+given] among their features, but they differ in other respects, such as the availability of a negative coda for CF. Note that this is not a property for RP in that a sentence such as (38) does not accept this negative alternative: Another topic-like property is that the fronted constituent is always definite and specific, as opposed to any kind of focus, which can be either specific or non-specific. This means that an indefinite and non-specific DP cannot occur in RP constructions. In a context where speaker A says that he has bought an orange shirt to wear at a party, Speaker B can reply with (41b) but never (41a). The difference is definiteness/specificity: Moreover, I have claimed that RP plays the discourse role of Aboutness-Shift Topic (AS-Top). Evidence supporting this claim comes the distinction between root and non-root properties (Emonds, 1969). It has been claimed that some main clause phenomena such as Topicalisation can only occur in root environments. Bianchi and Frascarelli (2010) have argued that an AS-Top is always restricted to main clauses. Furthermore, Jiménez- Fernández and Miyagawa (2014) have shown that Spanish AS-Tops are incompatible with factive verbs like lamentar ‘regret’, simply because the embedded clause subordinate to this type of verbs is not a root-like context (I abstract away from technicalities). The prediction is that if RP is an instance of AS-Top, it should be incompatible with this type of verbs. This is borne out by examples such as (42) from Jiménez-Fernández and Miyagawa (2014: 297), their (81): The unavailability of RP in embedded contexts points to the fact that RP is an AS-Top, given that other types of topic (instantiate by CLLD) does not exhibit this restriction8: 8One problem posed by the qualification of RP as topic is that in Romance topic fronting is linked to a resumptive clitic (when available in the language). This is CLLD. I have shown that CLLD and RP are different in many respects. It should be no surprise that they also differ in the insertion of the resumptive clitic. According to an anonymous reviewer, it sounds suspicious to consider RP as a case of topic fronting when there is no clitic. However, it should be clear that in other languages such as Portuguese, CLLD and Topicalisation of the English type coexist (cf. Barbosa, 2001), and the latter is never connected to a clitic, despite being a topic: . In the same vein, RP does not need to be linked to a resumptive clitic to qualify as a topic. 56 Á.L. Jiménez-Fernández / Ampersand 2 (2015) 49–60 Table 1 Information Focus: [+Focus] Contrastive Focus: [+Focus] [+Contrast] Mirative Focus: [+Focus] [+Unexpectedness] Quantifier Fronting: [+Focus] [+Polarity] [+Contrast] Resumptive Preposing: [+Topic] [+Aboutness] [+Shift] To conclude this section, the most salient discourse property of RPasanAS-Topisthat itstandsas ashiftoftopicwithrespectto the previous sentence that include the resumed material. To illustrate, let us return to examples such as (38). In the first sentence the speaker is talking about Lituma (AS-Top), and the comment part mentions the activities that some children are carrying out. The second sentence also has this discourse partition. However, by fronting the DP esas mismas cosas, there is a change of topic and nowthissentence isaboutthosesame thingsandnotabout Lituma. This shifting property is highlighted in every instance of RP in the examples discussed in this section. 2.6. A feature-based analysis of foci In this paper I entertain that focus is encoded as a syntactic feature [+Focus] which characterises all types of focus, and conveys purely new information. The distinction of types is realised by different combinations of features (a general line pursued by Haegeman,2010,2012;Starke,2001;Frascarelli and Ramaglia, 2013; among others). Table 1 outlines the feature composition of the types of foci that may be identified in Spanish, based on their interpretive properties I have discussed earlier: (44) Feature composition of types of foci and RP As is clear from the table, the basic type of focus is Information Focus, just carrying the feature [+Focus]; this discourse feature will be common to all types. Contrastive Focus also contains the feature [+Contrast], accounting for the contrastive/corrective interpretation of the focused element. Mirative Focus involves some sort of surprise on the part of the speaker in line with Cruschina (2012), which is captured in the system by adding the feature [+Unexpectedness]. Quantifier Fronting carries the feature [+Polarity], meaning that it emphasises the polarity of the sentence. Resumptive Preposing is separated from types of foci because, as I have argued, it is not a true focus but a topic and hence carries a [+Topic] feature. In addition, Quantifier Fronting (though not RP) affects the polarity of the sentence, and as such QF contains the afore-mentioned [+Polarity] feature, but RP shows no contrast at all. On the other hand, implementing Frascarelli’s (2007) and Frascarelli and Hinterhölzl’s (2007) proposal, RP carries a [+Aboutness] feature, indicating that its import is what the sentence is about, but also a [+Shift] feature, suggesting that the topic is newly proposed in the relevant sentence. 3. Methodology and sample data from tests Two tests have been created for a systematic interface analysis of focus fronting (for Mirative Focus, Contrastive Focus, Quantifier Fronting) and Resumptive Preposing in two varieties of European Spanish: Southern Peninsular Spanish (SPS, specifically Andalusian and Extremaduran varieties) and Standard Spanish (SS, mainly Asturian,theSpanish of Catalonia,BasqueSpanishand Madrilenian varieties).9As suggested in the RAE-ASALE (2009/2011), Standard Spanish is very difficult to define despite the unifying character of Spanish in general, and corresponds to the educated variety shared in all dialects. In this work, I concentrate on European Spanish, leaving American Spanish aside. In RAE-ASALE (2009/2011), Andalusian and Extremaduran varieties are geographically grouped together with common linguistic properties. It is within this approach that the division between SPS and SS should be taken. Hence, the main reason why I concentrate on Andalusian and Extremaduran Spanish is that they share the linguistic phenomenon that I am describing here (non-adjacency of V to Focus), which is not found in the rest of European varieties of Spanish. On the other hand, the varieties representing SS have been chosen motivated by the geographic origin of informants. In the first test all types of focus discussed above have been takenintoconsideration and focusfrontingwithpreverbal subjects was systematically compared with focus fronting with postverbal subjects in both varieties. The second test has concentrated on focus fronting with preverbal subjects in SPS, controlling the possible confusion with the postverbal subject constructions. Due to the fact that my goal is a full description of focus fronting in SPS, the number of informants is significantly higher in SPS (32 speakers) than in SS (24 speakers). From a sociolinguistic point of view, informants have a high education level (BA, MA and PhD students at the University of Seville and professors at the University of Seville and other universities in Spain). They all have a good knowledge of language and linguistics. As mentioned earlier, speakers have been classified as belonging to one variety or another depending on their geographical origin.10 The study was divided into two phases. In the first step, informants were given a set of sentences preceded by a context which induced a specific focus reading of the preposed constituent and asked to judge them (judgements could be expressed as */??/OK; informants were given a full description of what these conventional symbols stand for). The tests were presented in writing and in randomised order. Written stimulus is supposed to be neutral in that no auditive coercion with a specific intonation is exercised on speakers, who can thus freely interpret the relevant sentence with a specific discourse reading. The presentation of a context for each sentence also created the environment to induce a Given-Topic interpretation of both preverbal and postverbal subjects.11 The examples in (45) make up samples of the data that informants had to judge. Sentences A constitute the stimulus context; focused constituents are in caps and topics are underlined for my present purposes, though informants did not have this information. Sentences in 1 are instances of preverbal subjects, 9From mow onwards I will use the neutral term ‘types of fronting’ to refer to the three types of foci and to RP, though it should be clear that RP is not a type of focus, but rather a type of topic. 10 One of the questions in the test was ‘What variety of Spanish do you speak?’, and informants had two options, either Standard or Southern. This helped me in classifying speakers within one of these varieties. The main basis for this division is geographic, never sociolinguistic. Actually, the sociolinguistic factors characterising my informants are very similar in both varieties. 11 As argued in Frascarelli (2007), Given Topics instantiate given/familiar information, sit in the low C-domain and can be found in any type of embedded clause (Bianchi and Frascarelli, 2010); this can describe the preverbal subjects that I focus on in this work. On the other hand, postverbal subjects can also be interpreted as Given Topics in languages like Spanish (as opposed to Italian), as shown by Frascarelli and Jiménez-Fernández (2013). These Given Topics are not right-dislocated topics in Spanish; they remain in their original position, spec-vP. Á.L. Jiménez-Fernández / Ampersand 2 (2015) 49–60 57 Table 2 Fronting with a preverbal G-Topic. Test 1 Fronting with preverbal Given Topic DP subject (SPS) Type of fronting CF MF RP QF Grammaticality 56.25% (19/32) 56.25% (19/32) 50% (16/32) 40.6% (13/32) Grammaticality 37.5% (9/24) 41.6% (10/24) 29.1% (7/24) 16.6% (4/24) Table 3 Fronting with a postverbal Given Topic. Test 1 Fronting with postverbal Given Topic DP subject (SPS) Type of fronting CF MF RP QF Grammaticality 90.6% (29/32) 87.5% (28/32) 93.7% (30/32) 90.6% (29/32) Grammaticality 100% (24/24) 58.3% (14/24) 83.3% (20/24) 79.1% (19/24) whereas those in 2 contain postverbal subjects. Examples in B are for Contrastive Focus, examples in C are for Mirative Focus, examples in D illustrate Resumptive Preposing and examples in E instantiate Quantifier Fronting. 12 In the Appendix at the end of this article I include the whole set of sentences that informants were asked to judge. It was crucial to give the speakers more than one case of each type of fronting so I could make sure that their answers were systematic. 4. Statistical analysis of data In this section, I discuss the results of the different tests completed by native speakers. For the sake of clarity, I use tables so that the results can easily be followed. The figures in the different tables are based on the sentences judged to be ‘OK’ only. 4.1. Test 1: focus fronting with preverbal and postverbal ‘subjects’ In my first experiment, both SS and SPS speakers were presented with the data containing not only postverbal subjects but 12 Topicality of the subject is achieved via context bridging (Roberts, 2003), hence it is weakly familiar (Roberts 1993). It can be claimed that in the economic situation Spain is going through at the moment the involvement of the government in this type of events is expected. It is activated by the speakers’ knowledge about the new employment laws the Spanish government is proposing at present. also preverbal subjects, whose interpretation has been crucially induced to be that of a Given Topic. The results are presented in two tables, where I use abbreviations CF for Contrastive Focus, MF for Mirative Focus, RP for Resumptive Preposing and QF for Quantifier Fronting.Table 2 presents the judgements of SPS and SS informants with respect to preverbal subjects, whereas Table 3 comprises the results regarding postverbal subjects. Figures are accompanied by the number of speakers (in brackets) who rated the relevant sentences as acceptable: When SPS informants were faced with the systematic comparison between preverbal and postverbal subjects, they judged the lattertobefullygrammatical(approximately 90%). The same result obtained for SS with respect to Fronting with postverbal subjects, as clearly shown in Table 3. However, figures in Table 2 indicate that in the relevant discourse context SPS speakers found preverbal subjects acceptable (approximately 50%, except for Quantifier Fronting, with just 40.6%). Conversely, the SS informants rated preverbal subjects in constructions involving Fronting quite low. The partial conclusion to be drawn at this point is that SPS speakers prefer postverbal subjects with Fronting, but the preverbal option is at least acceptable. SS speakers dubbed the preverbal subject construction ungrammatical. However, the figures are somehow borderline. In order to have a clearer picture, a second test was designed based just on preverbal subjects, whose results are discussed in the next section. 4.2. Test 2: fronting with a preverbal given topic in SPS So far it is clear that SS does not include preverbal subjects as an option in Focus Fronting, whereas SPS speakers seem to accept this construction, though its acceptability is rather tricky with figures which do not point to full grammaticality. One of the main problems SPS speakers found was that if they were to choose between pre- and postverbal subjects in sentences with Fronting, they critically preferred postverbal subjects. This led me to create a second test, just for SPS speakers (though see below for SS), focusing exclusively on preverbal subjects so the informants could not relate the relevant sentence with one having a possible postverbal subject. The results are displayed in Table 4. When confronted with Fronting with preverbal subjects alone (with no explicit mention of the post-verbal constructions), SPS speakers showed a great tolerance with respect to Resumptive Preposing, yet we must consider this type of fronting just marginal. On the other hand, grammaticality increases with Contrastive Focus and it achieves its highest ranking with Mirative Focus and Quantifier Fronting. To have a full picture of the contrast between SPS and SS with respect to the acceptability of preverbal subjects in sentences with Fronting, I include Table 5. Here I show the low rates obtained in SS for preverbal subjects: