Past talks

Below you can find the talks which were given in the past.

(October) 2024 – (June) 2025

Easter Term 2025

Tuesday, 13 May 2025 (1-2pm, UK time)

Gabriel Martínez Vera (Newcastle University) – ​Remarks on evidentiality and illocutionary force

Abstract: This presentation examines Faller’s (2019) proposal to account for discourse commitments and reportative evidentiality in declarative clauses based on (i) the illocutionary force of presentation and (ii) the Collaborative Principle (Walker 1996). Key for (i) are a principal and an animator (Goffman 1974). The latter presents a proposition but need not be committed to its truth; the commitment lies in the principal—this characterizes reportative evidentials. In their absence (i.e., the default case), the animator and the principal are the same individual; assertion follows from this. As for (ii), the absence of overt disagreement towards the at-issue proposition means that the animator intends to resolve the question under discussion with it. Based on evidence from the reportative dizque in American Spanish and the direct -∅ in Southern Aymara, I propose that the only thing that presentation does is to put an issue on the table—importantly, there are no “defaults” tied to the denotation of presentation in the sense of Faller. I further revise the Collaborative Principle such that the discourse participants’commitment to the truth of the at-issue proposition, as well as to having adequate evidence for it followfrom this principle—ultimately, this seems reducible to Grice’s maxim of quality. In this approach, assertion arises as a result of presentation and pragmatic strengthening. 

Tuesday, 20 May 2025 (1-2pm, UK time)

Cristina Flores (Universidade do Minho) & Esther Rinke (Goethe-Universität Frankfurt am Main) – ​Overspecification of subjects and underspecification of objects: subject-object asymmetries in heritage Portuguese as a window into divergent diachronic pathways

Abstract: Heritage speakers of null subject languages have been shown to overuse overt subjects in comparison to monolingual or majority language speakers (see Keating et al., 2016; Montrul, 2004). This tendency to be overexplicit can even be observed in bilingual speakers who speak two null subject languages (Rinke & Flores, 2018; Sorace et al., 2009). In contrast, when it comes to object realization, heritage speakers of null object languages such as European Portuguese and Polish tend to omit objects more frequently than their monolingual peers (Rinke et al., 2018).  In this talk, we examine data from several studies illustrating the tendency to overuse subjects and to omit objects in Portuguese as a heritage language in Germany. In addition, we present new findings from a recent study on a corpus of written narratives of 31 Portuguese-German bilingual children and 30 monolingually-raised EP children, aged 10 – 12 years. The bilinguals produce higher rates of overt subject pronouns and, simultaneously, lower rates of overt objects. Thus, these opposed tendencies—overspecification of subjects and underspecification of objects—are observable within the same population and often within individual speakers. We discuss the underlying reason for this asymmetry and suggest that the opposed tendencies for subjects and objects are two sides of the same coin.  We argue that the omission and realization of subjects and objects in null subject and null object languages is mediated by discourse and referential factors. From a discourse-pragmatic perspective, subjects tend to be topics and omitted, whereas objects tend to be focussed and realized. From a referential perspective, however, subjects typically represent more referential (=animate, definite) referents and are realized, while objects represent less referential (=inanimate, indefinite) referents and tend to be omitted (see Cyrino et al., 2000). Bilinguals’ behaviour reflects a sensitivity to universal referential hierarchies, while simultaneously exhibiting a relaxation of discourse-pragmatic constraints. We propose that these asymmetries are not random but represent systematic and predictable tendencies shaped by reduced input and structural complexity. Notably, a similar development has been observed in Brazilian Portuguese, supporting our earlier hypothesis that heritage speakers’ grammars reflect ongoing diachronic development (Rinke & Flores, 2014).

Tuesday, 27 May 2025 (1-2pm, UK time)

Lena Baunaz (Université Côte d’Azur) & Anne-Li Demonie (Masaryk University) – ​Is expletive negation a unitary phenomenon? A cross-linguistic investigation of fear vs until-clauses

Abstract: Recent research has sought to account for the distribution and interpretation of so- called expletive negation (ExN)—a formally negative marker that appears to lack negative force—across various clause types. Many proposals have aimed, implicitly or explicitly, to offer a unified analysis of its diverse manifestations. One line of work argues that ExN carries a modal meaning, specifically an epistemic one (e.g. Yoon 2011; Makri 2013, 2015; Mari & Tahar 2020; Tsiakmakis et al. 2022; Tsiakmakis & Espinal 2022; Tahar 2023). Others, however, have argued that these markers still correspond to regular negators, and that deviant behaviour should be attributed to different scoping positions at LF, or even at a purely pragmatic level (e.g. Abels 2005; Cépeda 2018; Delfitto et al. 2019; Delfitto 2020). A further strand of research challenges the assumption of a unitary phenomenon, proposing instead a split between ‘weak’ and ‘strong’ ExN (Greco 2019, for Italian), or between ‘apparent’ and ‘real’ ExN (Tsiakmakis 2025 for Modern Greek). In this paper we build on the latter intuition, that there are different types of ExN and adduce morphological support for this claim. By comparing these markers in so-called emotive doxastic environments, i.e. fear-clauses, and temporal environments, i.e. until-clauses, we also aim to show that the distinction between these types is not one of ‘apparent’ vs. ‘real’ expletiveness, but rather one of ‘tensed’ vs. ‘modal’ expletiveness. We propose that this distinction is captured lexically, adopting an analysis within the nanosyntactic framework (Caha et al. 2025).

Tuesday, 3 June 2025 (1-2pm, UK time)

Adnana Boioc & Valentina Cojocaru (Universitatea din București & Institutul de Lingvistică al Academiei Române „Iorgu Iordan – Al. Rosetti”) – ​Code-switching in Romanian bilingual communities 

Abstract: Code-switching, a linguistic phenomenon characterized by the alternation between two or more languages within a single discourse, has long been a focal point of sociolinguistic (see Heller 1988; Gonzales-Velásquez 1995, etc.) and psycholinguistic research (see Myers-Scotton 1993). Early studies (e.g., Gumperz 1977; Poplack 1980) emphasized its sociolinguistic underpinnings, exploring the contexts and social factors that drive this bilingual behavior. Subsequent investigations delved into its psychological dimensions, focusing on the cognitive mechanisms and neuroanatomical structures governing bilingual language control (Desmet & Declerq 2006). More recently, researchers have turned their attention to its structural implications, analyzing the lexical, phonological, morphological, syntactic, and semantic alterations that emerge from language alternation (Bullock & Toribio 2009).

This presentation focuses on the structural dimension of code-switching, analyzing the types of changes that occur when a bilingual speaker alternates between two or more known languages (Zirker 2007: 8) in a speech act (see Myusken 2000). This alternation involves using elements from both languages in the same sentence or phrase (Ihemere 2003). Utilizing a corpus-based approach, this research examines both intrasentential and intersentential code-switching (Myers-Scotton 1993), focusing on several contact situations:

(i) The oral discourse of bilingual speakers of Romanian and Russian, the latter being used for interethnic communication. A symbolic distinction, formulated by Gumperz (1982: 66), is made between ‘we’ and ‘they’, Specifically, Russian represents the group’s language (home/group language, “we language”; generally, minority), while Romanian, not native to the group (out-group language, ‘they language’), is used in more formal contexts.

(1) La mnie prahodila odna… tot o tânără, și tot a zis că e din București, tot a făcut facultate. Și mă filma și am zis eu što ne hočy filmatsa, că eu sunt bolnavă (…).

‘another young woman came to me and said she also is from Bucharest, she also graduated college. And she filmed me and I said I didn’t want to be filmed, that I was sick (…)’(Romanian – Russian)

(ii) The oral discourse of bilingual speakers of Romanian and English in Romania.

(2) Now, I am gonnaschimb în treflă.

‘Now, I’m going to change to clubs.’ (Romanian – English)

(iii) The oral discourse of bilingual speakers of Romanian and French, Spanish, English living outside Romania, where Romanian is primarily used within the family/community.

(3)Regarde maman: c’estsuperfrumos.

‘Look, mom: it’s super beautiful!’ (Romanian – French)

The theoretical framework includes current approaches to linguistic contact and resulting interferences, specifically the code-switching phenomenon (see Matrix Language Frame Model). The method involves grammatical analysis of phenomena inventoried from a corpus of oral conversational texts. The corpus comprises recorded oral conversations: Romanian-Russian contacts include a corpus collected during fieldwork in Dobrogea, Romania (where there is a compact community of Lipovan Russians) and Moldova; Romanian-English, French, Italian, and Spanish contacts include conversations of bilingual children and adults. Topics varied and reflected informal interactions, with most examples extracted from spontaneous conversations.

The primary objective of this study is the grammatical analysis of code-switching in the discourse of Romanian and other Romance, Germanic, or Slavic language speakers. The study aims to identify and present:

(i)                  how code-switching occurs in bilingual discourse, with features varying significantly depending on the type of bilingualism (symmetric or asymmetric);

(ii)                factors driving this phenomenon in the analyzed discourse;

(iii)               the predominant type of code-switching (intrasentential vs. intersentential), considering bilingualism levels for structural classification.

Tuesday, 10 June 2025 (1-2pm, UK time)

Veronica Girolami (Università di Verona) – ​Combining Unaccusativity Diagnostics in Romance: Preliminary Evidence from Italian and French

Abstract: Since Perlmutter’s (1978) formulation of the Unaccusative Hypothesis, a central issue in syntactic
theory has focused on how languages mark the distinction between unergatives and unaccusatives.
Romance languages have long played a pivotal role in this debate, with diagnostics such as auxiliary
selection (Burzio 1986; Sorace 2000), past participle agreement (La Fauci 1988; Loporcaro 1998),
word order (Benincà 1988; Pinto 1997, Sheehan 2010) and ne-cliticization (Burzio 1986; Belletti &
Rizzi 1981) providing key evidence. While these tests are often applied independently, in this talk I
explore their interaction across three Romance languages: Italian, Catalan, and French particularly in
relation to ne/en-cliticization. Indeed, ne/en-cliticization appears as a particularly informative
diagnostic for investigating the interaction among these tests, as it is sensitive to argument structure
while also interacting with word order, auxiliary selection, and information-structural properties. To
do so, I investigate the distribution of partitive ne/en with unergative verbs as shown in sentences like
(1) from Italian, Catalan and French which should be ruled out by standard assumptions.
(1)
a. Ne cammina tanta, di gente, su quei marciapiedi
NE walk.3PS.PL many of people on those sidewalks
“There are many people walking on those sidewalks” (Lonzi 1986:121)
b. En dormiran tres a l’ habitació de la dreta
EN sleep.FUT.3PS three in the room of the right
“three of them will sleep in the room on the right” (Cortés & Gavarró 1997:41)
c. Il en a sauté beaucoup par la fenêtre (d’otages)
EXPL EN have.3PS jump.PST.PRT many from the window (of prisoners)
“Many prisoners have jumped out of the window” (Legendre 1990:94)
The availability of ne/en-cliticization with unergatives has long been debated in the literature (Burzio
1986; Lonzi 1986; Legendre 1990; Cortés & Gavarró 1997; Mateu 2002; Bentley 2004, a.o.) putting
at stake its status as a reliable diagnostic for unaccusativity. In this talk I will focus on the acceptability
of unergative verbs with ne/en-cliticization in three Romance languages providing preliminary results
from acceptability judgments in Italian, French and Catalan conducted as part of my thesis project.
Preliminary results indicate that while ne-cliticization with unergatives is variably accepted in Italian
and Catalan under specific syntactic and discourse configurations, it is rejected by native speakers of
French. These findings seem to suggest that the acceptability of ne/en-cliticization may be influenced
by other parameters that go beyond the unaccusative/unergative distinction.

Tuesday, 17 June 2025 (1-2pm, UK time)

Jonathan Kasstan (University of Westminster) & Coppe van Urk (Queen Mary University of London) – ​tbc

Tuesday, 24 June 2025 (1-2pm, UK time)

Jan Casalicchio (Università degli Studi di Siena) – ​Verb second in main and embedded clauses in Ladin

Abstract: In this talk, I discuss the presence of a Verb second rule (V2) in different varieties spoken in North-Eastern Italy. This area shows an interesting case of microvariation, because the contexts in which the V2 rule applies change from one variety to another. The talk will be based on the data that have been collected within the PRIN PNRR 2022 project Resilient Syntax in Contact: Assessing Minority Languages (RESYNC). I will focus on the V2 rule in embedded clauses of Northern Dolomitic Ladin (Gardenese Ladin and Badiotto Ladin, spoken in the province of Bolzano/Bozen), comparing the data from these varieties to what we find in “major” languages like German, Danish and Yiddish, and in the German islands of Sauris and Timau (province of Udine) which are also studied in the RESYNC project. The data from Ladin are particularly interesting because they show a pattern that has not been documented so far: in languages like German or Danish embedded V2 depends on the absence of an overt complementiser, or on the presence of a bridge verb in the main clause. In the Ladin varieties, instead, a crucial factor for the presence of embedded V2 is the type of fronted element: if an object is fronted, embedded V2 is much more restricted than when an adverbial is fronted. I discuss two possible approaches that could potentially explain these data: the pruning hypothesis (cf. Haegeman 2006) and the hypothesis that the left periphery of embedded clauses hosts a factive operator in FinP (cf. Haegeman 2012). According to the latter, a direct object cannot be fronted because it would cross the operator (if we assume the latter has a nominal nature), because it would violate relativised minimality.

Lent Term 2025

Michaelmas Term 2024

For those interested in Kim’s work and/or in her presentation, please contact her directly at kim.groothius@ugent.be.

Abstract: Se analizarán y presentarán las vías y la cronología de introducción, el alcance textual y las características morfológicas, gramaticales y semánticas de algunas de las 614 estructuras de origen catalán que se documentan en el Diccionario del castellano del siglo xv en la Corona de Aragón (DICCA -XV). El siglo xv cobra especial interés para un estudio de estas características, por dos motivos: tal y como señalan Nadal y Prats (1987: 483 y ss.), se trata del período en que se inicia la catalanización paulatina de la poesía y el desplazamientoconsiguiente del provenzal, que había sido la lengua de la poesía durante la Edad Media.  Por otro lado, y desde el punto de vista de la historia de la lengua española, el siglo xvejerce un papel fundamental en la gestación de lo que será la lengua clásica de los siglos xvi y xvii, ya que la variedad híbrida y fronteriza entre las vertienteslingüísticas occidental y oriental, tan característica del castellano del siglo xv en la Corona de Aragón (cfr. Lleal 2011: 51), destaca por su gran cantidad de neologismos (cfr.Lleal 2009 y 2016, Dworkin 2012 o Raab 2015).

El estudio aborda transferencias léxicas (1), morfológicas (2) y sintácticas (3):

1. Despues los turcos dieron mucha priessa en batir la torre de Sant Nicolas pensando muy presto ganar la ciudadsi con·ella diessen primero en·el suelo: entra el molle trezientos passos el mar adelante que cierra el puerto y lo asegura por subtil manera contra la parte de occidente: es asentada encima la torre contra la parte de transmontana donde solia los tiempos antiguos estar el coloso siquiere memoria de·la qual se hablo sobre la figura de todos pintada que fue de·lassiete marauillas vna y la principal en todo el mundo: el qual a .liij. años despues queende se puso cayo en pedaços por vna tierratremol o terremoto [Viaje 167r] 
2. E por reuerencia de Dios e por contemplacion nuestra los administradores del dito spital han feta graciae merçe de·la meyetat de·la dita quantidat a Maria filla d·en Ferrando de Solis quondampubilla e en ayuda de su casamiento. [Cancill-3110:022r]
3.  Ya se sea que hombre deua sembrar en campos temprados. conujene a·saber quenon seanmucho enaguados njn mucho secos. [Agricultura 011v] 

En resumen, la mayoría de catalanismos léxicos son tecnicismos y se documentan en textos científicos y técnicos, sobretodo en traducciones del catalán, mientras que los textos jurídicos y administrativos recogen un mayor número de elementos gramaticales procedentes de la lengua catalana.

Abstract:This paper investigates three discourse markers in Mauritian Creole, a French-lexified Creole spoken in Mauritius, a multilingual speech community where most speakers switch between French, Creole and English on a regular basis. The three discourse markers that we investigated, dizon, koumadir and savedir, are all derived from the verb ‘to say’ dir, based on French dire. The focus of our investigation is two-fold. First and foremost, we wish to profile the totality of usage of each marker, paying close attention to any potential ‘common denominators’ (cf. Östman, 1995; Lansari, 2020). Secondly, given the high degree of code switching, we are interested in whether these markers are used in the same way when speaking Creole or French. To this end, we analyse the three markers in dialogic contexts between pairs of Mauritian speakers in predominantly Creole conversations, predominantly French conversations and a few conversations with a high degree of code switching between the two. The framework that we adopt for our analysis is loosely based on a canonical typology approach, whereby each marker is rated according to whether certain features are present or absent in their usage, allowing for both a qualitative analysis of the markers in context, and a small-scale quantitative approach. Given that all three markers are ultimately based on French dire, the starting hypothesis is the existence of a core meaning, shared by the markers, possibly related to the etymology of dir, that in each marker also co-exists with individual uses that have developed in the Creole. Our investigation reveals that the three markers are at different stages of grammaticalization/ pragmaticalization and have therefore developed different functions. The verb dir which entered Mauritian Creole in the 18^th century to combine with comment to create koumadir and develop into a uniquely Creole pragmatic marker is quite different from the dir in dizon and savedir, the markers that have appeared at a later stage. The systematic investigation and comparison of these markers show that koumadir is only used in Creole contexts, dizon matches very closely the pragmatic marker usage of French disons, and savedirshows a higher degree of integration into Creole, even if it is mainly used with its original, literal meaning.

(October) 2023 – (June) 2024

Easter Term 2024

Tuesday, 7 May 2024 (12-1pm, UK time)

Xavier Villalba (Universitat Autònoma de Barcelona) – ​Mirative markers in Romance exclamative sentences

Abstract: Exclamative sentences in Romance involve a rich set of markers whose syntactic position and semantic/pragmatic meaning is unclear. Consider the markers (mrk) in boldface in the exclamative sentences in (1):

(1) a.  Si      (que) n’és,       de bo!                                                               (Catalan)

          mrk  mrk  of.it-is    of good

          ‘How good it is!’

     b.  Maque   és bonic això!                                                                 (Valencian Catalan)

          mrk  mrk  is  nice   this

          ‘How pretty this is!’

     c.  ¡Vaya,   qué    delirio!                                                                      (Spanish)

          mrk       what  delusion

          ‘Wow, what a delusion!’

The hypothesis we will defend is that the meaning encoded by each of these particles, while contributing to creating an exclamation speech act, is neatly different, and spreads across Rizzi’s left periphery (Rizzi 1997) and Krifka’s Speech Act Layer (Krifka 2021).

Sentence-type. At the lower level we find the marker que, encoding the exclamative sentence-type (Villalba 2003), not only in (1)a-b, but in wh– and definite exclamatives as well (2):

(2) a.  Que   bo     que   és!     b.  ¡Las/Menudas cosas      que       come!

          how   good mrk  is           the.f.pl/small things     mrk      eats

          ‘How good it is!’               ‘The things she eats!’

This marker is homophonous with optative and interrogative que (Prieto & Rigau 2007; Sánchez López 2017a), so I will suggest that it encodes sentence-type, and generates in the head of ForceP. We will discuss the consequences of such a move for analyses of exclamatives following (Benincà 1996), who argues this element is focus marker.

Degree quantification. Exclamative sentences must include a degree operator creating the necessary domain extension which surpasses the speaker’s expectations (Zanuttini & Portner 2003; Rett 2011; Castroviejo 2021): this is the case of si (1a) (cf. si in Spanish exclamatives (Sánchez López 2017a)), the null operator (1b), the wh-words qué (+N)(1c) and que (+A) (2a), and the definite DP las/vaya cosas (2b). All these elements move to the Spec of ForceP. Schematically:

(3) a. [ForceP si/OP [Force’ que [FinP [TP n’és de bo ]]]]

     b. [ForceP que bo [Force’ que [FinP [TP és ]]]]

Mirativity. In a higher position than ForceP we find mirative markers, like Catalan mira/ma, goita, òndia, Italian guarda, Portuguese olha, nossa, or Spanish mira and vaya. These elements modify the whole proposition, marking that the information is unexpected for the speaker, and can appear either at the beginning of the utterance, or at the end:

(4) Tu,    goita les          pentinetes del       monyo!            Ma!

     you    mrk  the.f.pl  combs       of.the   bun                  mrk

     ‘Look at the bun’s combs! Wow!”

Here, the proposition is modified by both initial goita and final ma. For final mirative markers, we follow (Espinal, Real-Puigdollers & Villalba 2022) in considering that a propositional anaphor bound by the previous proposition is involved. Since mirative markers encode the speaker’s surprise towards the situation expressed by the proposition they modify, we will generate them in the layer devoted to the speaker’s epistemic states by (Krifka 2021), namely J(udge)P.

     (5) [JP [Jº ma] [ForceP que bo [Force’ que [TP és ]]]]]]

     We consider mirative markers in exclamative sentences only, but note that they may occur in declaratives or interrogatives as well (6) (Olbertz 2012; Cruschina, Giurgea & Remberger 2015; Sánchez López 2017b):

(6) a. O João, nossa, eu não sabia que ele era tão esperto. (Br. Port. (Moreira 2017))

          ‘John, wow, I didn’t know he was so smart.’

     b. Nossa, mas o que aconteceu?

          ‘Wow, but what happened?’

Hence, we will treat mirativity and exclamativity as two distinct categories (cf. (Michaelis 2001; Unger 2019)).

Proposal. We obtain the following typology of markers:

These elements are distributed in the left periphery of sentence as follows:

(7) a. [ActP [Act ! ] [ComP [Comº ⊢ ] [JP [Jº ma] [ForceP que bo [Force’ que [TP és ]]]]]]

     b. [ActP [Act ! ] [ComP [Comº ⊢ ] [JP [Jº ma] [ForceP OP/si [Force’ que [TP ésbo ]]]]]]

The illocutionary operator ! in ActP converts the proposition into an exclamation (these are exclamative sentences, but the same analysis would apply to declarative sentences).

     The placement of mirative markers above ForceP is confirmed by their interaction with vocatives. As argued by (Slocum 2016; González López et al. 2023), utterance initial vocatives typically fulfill a call role. Hence, they precede mirative markers (ex. (4) above), so we will suggest this vocative projection (Voc(call)P) is above ActP:

     (8) [Voc(call)P tu [ActP [Act ! ] [ComP [Comº ⊢ ] [JP [Jº goita ] [ForceP … ]]]]]

In contrast, lower vocatives are typically used for maintaining contact between speaker and listener, and appear typically after mirative markers, but before degree operators:

(9) a.  Ma,   chica,    quina falla      més      bonica.

          mrk   girl         what  bonfire so        pretty

          `Wow, girl, what a pretty bonfire!’

     b. Bueno,   hombre,     vaya  genio…

          well       man            what  temper

          ‘Ok, man, what a bad temper…’

Accordingly, these vocatives will appear in a Voc(addr)P just above ForceP (as suggested by (Hill 2007; Espinal 2013)):

(10)   [Voc(call)P [ActP [Act ! ] [ComP [Comº ⊢ ] [JP [Jº ma ] [Voc(addr)P chica [ForceP quina falla…]]]]]

The global picture shows that markers in Romance exclamatives have a strict division of labor in creating an exclamation speech act, which syntax reflects transparently.

Tuesday, 14 May 2024 (12-1pm, UK time)

Roberta D’Alessandro (Utrecht University) – Honorificity, phi-agreement and an inside-view on pronouns and referentiality. The case of Italian voi and lei

Abstract: Honorifics are not as widely present in Romance languages as they are in several Asian languages. However, they are equally relevant when trying to identify the internal structure of pronouns. Italian features two main honorary pronouns: lei and voi: by examining their agreement patterns and syntactic behavior, it will be shown that their rather complex (and non-homogeneous) agreement patterns result from the interaction of an HON feature with other phi-features. It will be shown that HON is related to gender, and that it is not a purely deictic feature. Finally, by examining ongoing changes in the use of honorific pronouns, conclusions will be drawn regarding the directionality of language change and the relevance of phase edges as loci where narrow syntax “communicates” with its interface systems. Honorificity will be shown to be similar to prosody in targeting phase edges and being external to the core pronominal (or verbal, in the case of prosody) structure.

Tuesday, 21 May 2024 (12-1pm, UK time)

Vlad Jipa (University of Bucharest) – Romanian invariable nouns and the neuter gender 

Abstract: It has been claimed that a distinct, third neuter gender might be currently emerging in Romanian (Maiden 2016). This is alleged to be the case with Romanian invariable nouns, more precisely recent loans (mango, kiwi) which are not yet (fully) morphologically adapted to the system of Romanian. Not only do these nouns have the same singular and plural form, but they also display genus alternans (masculine agreement in the singular, feminine in the plural), despite there being no transparent morphology to license this behaviour. Thus, it could be concluded that they belong to a different (controller) gender due to their peculiar agreement pattern.

I will try and address the issue of an emerging third gender in Romanian by investigating invariable nouns: masculine, feminine, and ‘neuter’. I do this for three main purposes. First, I intend to find all ‘neuter’ nouns which are invariable and could constitute an ‘emergent’ gender in Romanian, following Maiden’s (2016) claims. Secondly, I wish to check if ‘neuter’ invariable nouns are quantitatively significant (especially compared to masculine and feminine invariable nouns). Thirdly, I want to observe if there are ambiguous noun endings with respect to gender – especially in the singular. This will be useful because most ‘neuters’ take the plural ending -uri in everyday use, thus justifying the feminine plural agreement (therefore, inflexion motivates agreement in such contexts). If these nouns receive -uri in the plural, can we still talk about an emergent ‘neuter’ gender in Romanian?

My goal is to catalog invariable ‘neuters’ in DOOM³ in order to investigate their morphology (especially full final vowels and -uri) and their agreement pattern. Some affixes (like -o in mango) are foreign to Romanian, therefore it is even more difficult for them to be adapted to its morphophonological system. However, -uri can be attached irrespective of the phonology of the root (and this frequently happens in day-to-day use), which explicitly licenses feminine plural agreement; masculine singular can be easily explained as a morphological default (nonwords typically take masculine singular agreement in Romanian – e.g. un bum ‘a boom sound’ –, thus recent loans could be similarly perceived by speakers).

Tuesday, 28 May 2024 (12-1pm, UK time)

Cristina Guardiano (University of Modena and Reggio Emilia) – Testing Parametric Comparison against ultralocality

Background. The Parametric Comparison Method (PCM) is a system of phylogenetic reconstruction that uses syntactic parameters as taxonomic characters. Since Longobardi and Guardiano (2009), it has been exploited at three levels of historical depth:

1. Meso-comparison. The method is able to identify the internal articulation of established historical families (Longobardi et al. 2013, Ceolin et al 2020) and separate them from one another, leading to the neat conclusion that parameters encode a signal largely reflecting vertical transmission. The small proportion of inconsistency between parametric taxonomies and the accepted wisdom is entirely imputable to heavy interference: syntax is not immune from secondary convergence effects, but the latter become visible on parameters only when extensively observed on other linguistic levels.
   
2. Macro-comparison. Capitalizing on these results, the PCM has begun to explore the depth of the historical signal of parameters testing the statistical strength of cross-family aggregations (Ceolin et al 2021). The first analyses suggest that the method distinguishes between statistically plausible relationships and cases in which accidental similarity cannot be excluded.
   
3. Micro-comparison. Since Guardiano et al (2016), the PCM has become to be tested on subsets of languages closely intertwined geographically, genealogically, and sociolinguistically. The results support the conclusions that the historical signal encoded by parameters is strongly vertical (when comparing historically close languages, parametric distances become lower, as expected if we assume that recent splits cannot have developed a great number of syntactic differences) and that instances of secondary convergence, although visible especially in selected subdomains, do not have dramatic effects in terms of classification.

Beyond micro-comparison. The fact that the signal encoded by parameters does not saturate at the local level hints that parametric similarities are likely to reveal aspects of syntactic microvariation able to shed light on the fine mechanisms of contact and interference which shape parameter change. To better investigate this domain, the PCM has begun to be tested against a shallower level of historical depth: dialect classification in a monophyletic domain (Devoto 1953). As a preliminary attempt to quantitatively assess dialect similarity using syntactic parameters, we analyzed the dialectal structure of a pilot dataset of 36 Romance dialects of Italy belonging to 5 acknowledged macro-groups (Ascoli 1882, Merlo 1924, Rohlfs 1937, Pellegrini 1977): Extreme Southern (14), Upper Southern (11), Northern Gallo-Italic (6), Venetian (3), Gallo-Italic of Sicily (2).

Tools and methods. To explore syntactic diversity across the dataset, we started from the list of 94 DP-parameters used in Ceolin et al (2020, 2021) and described in Crisma et al (2020), that was integrated with a set of 14 additional parameters capturing aspects of internal variation among the relevant dialects in the nominal domain. The data were collected using a list of 257 manifestations, corresponding to the observable patterns generated by the 108 parameters. For each pattern, a minimum of three examples were elicited from native speakers (including ungrammatical patterns, when possible). Each manifestation was associated to a YES if the speaker produced the relevant examples, to a NO if the pattern was not available (or ungrammatical). Using this information, the 108 parameters were set in accordance with the parameter setting algorithm presented in Crisma et al (2020) and taking into account the implicational structure of the system (Guardiano and Longobardi 2017). The output is a list of 36 strings of parameter values, corresponding to the 36 selected dialects.

Comparison and classification. In line with the usual PCM procedures and with dialectometric approaches (see, e.g., Nerbonne et al 2011, Nerbonne and Kretzschmar 2012, Leinonen et al 2016), cross-dialectal similarity was measured through the assessment of distances, by computing the number of identities and differences in the parameter states of each pair of dialects and reducing them to a monadic figure through the Hamming formula (d/(i+d)). The distribution of parametric distances was then evaluated using a selection of computer-based techniques imported from quantitative phylogenetics (hierarchical clustering) and dialectometry (PCoA, maps). The exploratory analyses show that parametric distances do not saturate at this “ultralocal” level: hence, in principle, it is possible to identify aggregations of dialects (i.e., syntactic isoglosses) based on parameter similarity. The next step is checking whether such aggregations have any historical significance. To this end, we tested parametric distances against two sets of variables: the linguistic data traditionally used for dialect classification, and geographic distances. To check whether parametric taxonomies are consistent with accepted dialect classifications, we used 30 binary characters extracted from the 33 isoglosses of Pellegrini’s (1977) Carta dei dialetti d’Italia: after collecting the relevant data from the dialects of our dataset, we represented each dialect as a list of 33 binary values and we computed their distances using the Hamming formula. The correlation between the distance matrix so derived and parametric distances was performed through quantitative and dialectometric tools (Mantel tests, heatmaps, PCoA, GabMaps). The Mantel test reveals a good correlation (Pearson = 0.72): in fact, the aggregations of dialects based on parametric similarity largely overlap with acknowledged dialect groups. The inconsistencies mostly correspond to independently known areas of discontinuity, such as the border between northern and central-southern Salento (Parlangeli 1953), or between southern and central dialects of Calabria (Rohlfs 1972, Alessio 1963/1964). In this respect, parametric similarity is likely to reflect aspects of variation which fail to be identified by traditional morphophonological isoglosses. Correlation with geographical distances (0.64) quantitatively corroborates the expectation that syntactic similarities, even in this domain, are not entirely predictable from areal factors. Correlations with two further matrices of linguistic characters extracted from Pellegrini (1970) and SSWL (Koopman and Guardiano 2022) return more spurious results, which suggest that different types of linguistic data are not equally (im)permeable to homoplastic effects.

Tuesday, 4 June 2024 (12-1pm, UK time)

Josep Ausensi Jiménez (University of Alicante) – Hybrid subjects in Spanish and Catalan: Halfway between agents and patients

Abstract: We analyze an intransitive construction involving verbs like Spanish matarse ‘kill’ whose subjects appear to have both internal and external argument properties, e.g., Juan se mató en un accidente de coche ‘Juan got himself killed in a car accident’, in which the subject’s referent shows hybrid behavior between agent and patient as it needs to be engaged in an action leading to its accidental death. We propose that the subject’s internal and external argument properties can be accounted for if subjects can bear two semantic roles by virtue of being associated with more than one distinct head in the syntax (Pineda & Berro 2020). We argue that such intransitive uses involve a distinct argument structure from transitive reflexives despite sharing the same surface form, cf., El sospechoso del homicidio se mató al estar rodeado por la policía ‘The suspect killed himself when he was surrounded by the police’. 

Tuesday, 11 June 2024 (12-1pm, UK time)

Iulia Anghel (University of Bucharest) – The main features of the verbal and nominal phrase in Romanian Sign Language

Abstract: In 2020, Romanian Sign Language (RSL) became recognised by law as the mother tongue of deaf and/or hard of hearing people. However, to date, there are no linguistic studies addressing this topic. In the context of a contemporary world that places more and more emphasis on its inclusivity, it is essential to have a linguistic study dedicated to this topic, as it can put linguistic principles in a new light, unraveling a new way of functioning. Thus, in this presentation, I aim to study the morphology and syntax of the verb phrase and the nominal phrase. In order to elaborate this presentation, we have consulted the entire Dictionary of Romanian Sign Language, which gathers some of the signs, namely 3500 out of about 7000. In addition, from the digital textbook created by the ANIALMG team and the Orange Foundation, we have selected twelve texts of varying degrees of difficulty in order to contextualise the lexical-grammatical elements under analysis. The corpus will be the basis in elaborating the three parts of the presentation. In the first part, we will argue that RSL is a natural language. The second part will be devoted to the verbal phrase, while the third one will focus on the nominal phrase (with both noun heads and pronoun heads), which we will approach from the perspective of derivational morphology and inflectional morphology. Also, we will concentrate on how grammatical categories such as time, aspect, manner, person and numberare saturated in the VP and how number, gender, case and definiteness appear in the NP. Therefore, we will answer the following research question: How different is RSL compared to the standard spoken/oral Romanian or to other sign languages?

https://prezi.com/view/ofUZVFgAPJ34eqNb6KIZ/

Lent Term 2024

Tuesday, 30 January 2024 (12-1pm, UK time)

Vitor Nóbrega (University of São Paulo/FAPESP) – Structuring the syntax of expressivity: The case of Brazilian Portuguese binomial DP

Abstract: In this talk, I explore the descriptive and expressive meanings of two classes of nominal modifiers in Brazilian Portuguese (BP): (i) epithets (e.g., mala lit. suitcase ‘annoying’; cachorro/a lit. dog.M/F ‘shameless’; banana lit. banana ‘lazy’) and (ii) slurs (e.g., droga lit. drug; porcaria lit. garbage ‘no-good’). These nominal modifiers are licensed either an attributive position (1a) and in phrases that follow the DP₁-of-DP₂ scheme (1b). In both cases, they display an idiomatic interpretation, as can be seen with the readings of N₂ in (1).

a. D                             N₁                               N₂

   O/esse/um                vizinho                         mala/porcaria

   the/this/a                  neighbor                     suitcase/garbage

b. D₁                           N₂                               of-D₂                           N₁

   O/esse/um                mala/porcaria             d–o/esse/um                vizinho

   the/this/a                  suitcase/garbage         of- the/this/a               neighbor

   ‘That/a annoying/no-good (of) a neighbor’

Nouns in predicative position are subject to two types of interpretation in BP. They can either display a descriptive interpretation, and hence contribute to the ordinary descriptive content of the DP, as illustrated in (1), or they can convey an expressive interpretation, where they are semantically opaque to descriptive content, and simply express a negative attitude of the speaker toward the proposition. Interestingly, an expressive interpretation is only obtained when the predicate (N₂) inverts its position with the subject noun (N₁) (1b) (Basso, 2020). Once the DP bears such configuration, expressiveness can either have a narrow scope, predicating over the DP, when the predicate noun is prosodically marked (2a), or it can have a wide scope and take as its semantic argument the entire sentence, when the main verb is prosodically marked, as in (2b) (see Potts, 2007; Gutzmann, 2019).

  A     Maria  está namorando_EX2[o/aquele  mala/porcaria_EX1  do/aquele   vizinho].

      the   Mary   is     dating                the/that    suitcase/garbage    of-the/that  neighbor

1. Narrow scope (EX₁): The speaker feels negatively about the neighbor.
2. Wide scope (EX₂): The speaker feels negatively about the fact that Mary is dating the neighbor.

To account for the idiomatic aspects of the descriptive layer of meaning, I examine how the syntactic structure of predication contribute to create a domain for idiomatic interpretation. Following the approach for DPinternal predication developed by den Dikken (2006), I argue that predicative positions, defined in terms of a relator head (R) within the DP, trigger root allosemy when hosting a nominal head in BP. This empirical observation puts forth some theoretical implications for the syntax of word meaning: it indicates that root allosemy can be conditioned not only by locality with a specific functional head, as largely observed in the literature (Arad, 2003; Marantz, 2013), but also by a dedicated syntactic configuration (viz., [R, Compl]).

With respect to the expressive layer of meaning, I provide a set of syntactic tests to show that expressiveness in BP is structure dependent, since it is only licensed in DP₁-of-DP₂ structures. Thus, expressiveness in BP is a structural rather than a lexical property (contra Bastos-Gee, 2011, 2013). With this in mind, I claim that expressiveness is the result of an agreement interaction of a syntactic expressive feature [Exp] (Gutzmann, 2019), codified in a linker head internal to the DP (den Dikken, 2006) —which triggers predicate inversion—, and an interpretable, but unvalued [Exp] feature in the speaker head composing a Speaker-Addressee Phrase (saP), which is available in the superordinate structure of the nominal extended projection, i.e. above DP. This is what gives rise to the narrow scope reading in (2a). Once such an agreement relation is obtained, an expressive interpretation is assigned to the whole DP, consequently overwriting the descriptive content of the inverted predicate noun. The wide scope reading in (2b), in turn, is argued to be the result of an agreement relation between the [Exp] feature of the saP layer in the DP with the saP layer of the clausal spine, i.e. above CP (Wiltschko, 2021; Miyagawa, 2020). Finally, I discuss a set of relevant counter-arguments against the predicate inversion analysis for (1b) recently put forth by Saab (2022).

Tuesday, 6 February 2024 (12-1pm, UK time)

Monica Irimia (University of Modena and Reggio Emilia) – DOM and impersonals: some notes from Romance

Abstract: Impersonal and passive configurations are an important testing ground into the nature of differential object marking (DOM), especially under theories that link its dedicated morphosyntax to a strategy to disambiguate objects from subjects (Comrie 1989, Aissen 2003, Richards 2010, see also the extensive discussion in Aranovich 2009, a.o.). Despite the prediction that the suppression of the external argument should lead to DOM being blocked in these contexts, passives and impersonals show instead a mixed behaviour across the Romance domain (just like in other language families). If the periphrastic passive tends to block DOM (Jaeggli 1986, a.o.), the SE-passive (including the so-called impersonal SE and the SE-passive) is more variable – it allows DOM in some languages (varieties of Spanish, dialects of southern Italy, etc.), but not in others, such as Romanian (Dobrovie Sorin 1998, 2021, Giurgea 2019, a.o.). Additionally, a less discussed fact is that SE-DOM languages are not uniform, many diverging points being salient, especially when it comes to the interaction with (accusative) clitic doubling and the possibility of overt agreement. 

This presentation has two main goals: first, it contains a reassessment of various diagnostics which have been taken to indicate the projection of an implicit external argument in passives and impersonals with DOM. The data come from both Western Romance and Daco-Romanian (SE constructions in standard Romanian and the so-called clausal supine in Moldovan, Hill and Alboiu 2016, Costea 2021, a.o.). Secondly, it summarizes the various patterns of variation seen with DOM under passives and impersonals: whether DOM can be clitic doubled or not, whether it can trigger agreement or not, whether it allows full pronouns or not, whether full nominal DOM is permitted, while (third person) accusative clitics being blocked, whether overt dislocation is necessary or not (building on and extending Mendikoetxea and Battye 1990, Mendikoetxea 2008, Ordóñez and Treviño 2016, Ormazabal and Romero 2022, a.o., as well as novel data). The presentation will also address theoretical aspects that are relevant to an understanding of how DOM interacts with arbitrary constructions:  whether v can license structural accusatives in the absence of a syntactically projected external argument (Collins 2005, Pujalte and Saab Saab 2012, Saab 2014, Šereskaitė 2021, a.o.), and thus whether accusative Case is available in SE-passives (D’Alessandro 2007, Ordóñez and Treviño 2016, MacDonald and Melgares 2021, a.o.) and the Moldovan clausal supine, or how arbitrary constructions interact with (the lower vP) discourse-related layer(s).

Tuesday, 13 February 2024 (12-1pm, UK time)

Gigi Andriani (University of Hamburg) & Manuela Pinto (Utrecht University) – Determiner omissions in the Heritage Italo-Romance varieties of New York City

Abstract: This talk presents cases of determiner omission in heritage Italo-Romance varieties spoken in the New York area and puts forth an account of the structural representations that may underlie the differential heritage grammar. We examine the interplay among different sources of linguistic input and how these may contribute to the shaping of novel DP-structures which are not the result of direct transfer, but of an independent reorganization of the DP-internal constraints which are absent from the input.

The NYC corpus was built for control purposes within the MicroContact project (D’Alessandro 2018 et seq.; https://microcontact.hum.uu.nl/#contributions) and includes a heritage varieties from Italy in contact with English: Italo-Romance (Nònes Trentino, E. Abruzzese, Cilentano, Apulo-Barese, Sicilian) and Rhaeto-Romance (Friulian) varieties, as well as spoken Italian, i.e. the NYC koine, an Italian-based contact variety used as the shared ‘community language’(Haller 1987). The spontaneous speech, elicited from 58 speakers (first generation G1: 32 vs second generation HS: 26) by means of semi-guided interviews, reveals an incipient tendency in the omission of some core functional heads (Andriani & D’Alessandro 2022), among which definite articles, i.e. D-heads. This tendency is only attested in second-generation HSs of all Italo-Romance varieties considered here, while G1 speakers show no signs of attrition in the use of definite articles. Instead, HSs produce ‘non-target’, article-less DP-structures (Ø_D) in the contexts in (1)-(5a), except for the opposite tendency with possessives modifying kinship terms in (5b):

• Plural NPs in subject position: Ø_D–NP_[PL]

            stanno a parlare Ø_D llingue sue (=loro)… ma io non gapisco, Ø_D italiani non vogliono capire… non vogliono pratticare Ø_D italiano adesso             [Barese (F, 65)]

‘they’re all speaking their own languages… but I don’t get it, Italians do not want to understand… they don’t want to practice Italian now’

• Singular bare NPs (± generic): Ø_D–NP_[SG]

i ai dudj (Ø_D)credits, ma i na i ai Ø_D diploma                                  [Friulian (m, 93)]

            ‘I have all the credits, but I do not have the diploma(/certificate)’

• Locative (and other) PPs: PP-Ø_D-NP

Però, mó, se vai in–Ø_D ristorante, pagano.                                        [Abruzzese (f, 47)]

‘But, now, if you go to the restaurant, they do pay.’

• Numerals (4a), Quantifiers (4b), and ‘-sective’ APs (same, other): Ø_D-AP-NP/Q-Ø_D-NP

a.    Ø_D primma vota, jì avev’òtt’annə, Ø_D primma votə ca so gghiutə   [Cilentano (f, 30)]

‘(The) first time I was 8 years old, the first time I went (to Rimini to visit my cousins)’

b.    perché el l’ha fat el panetier tuta Ø_D sua_D vita in Italia                   [Nònes (f, 71)]

       ‘because he was a baker for all his life in Italy’

• Possessives with common nouns (5a) vs kinship terms (5b): Ø_D-Poss-NP vs D-Poss-NP

a.    prima stavan’i tagliani, come Ø_D mia_D compagna A.                       [Barese (f, 42)]

            ‘before there were Italians, like my friend A.’

b.    cə sta la famijja purə də mamm’e ppapà, e anche del mio marito    [Abruzzese (f, 47)]

       ‘(In Abruzzo) there is also mum and dad’s family, as well as my husband’s’

Crucially, parts of these tendencies had already been highlighted in Haller’s (1987 et seq.) work on NYC Italian, as well as in Sydney (Bettoni 1991) and Montreal (Reinke 2014) Italian. While transfer cannot be excluded for some structures, the general behaviour of the heritage DP can be understood as the result of a differential ‘feature reassembly’ (Lardiere 2008) of D-related features, responsible for the (c)overt distribution of heritage definite articles compared to the input languages (cf. Longobardi 1994; Chierchia 1998; i.a.). We suggest that this feature-reassembly process is holistic and taps into the many varieties and variation of syntactic options present in the input (cf. Cardinaletti & Giusti 2018) and leading HS to reanalyse the structural variation in the input against a principle of relative economy of derivation. As a result, HSs allow a more permissive distribution of null Ds, even though this increases interpretative ambiguity whenever the (bare) noun is unmodified.

These changes might be linked to the encoding of number in the heritage Italo-Romance DP, as well as in the verbal domain (e.g. loss of subject-verb agreement on T, e.g. cose_[PL] cambia_[SG]; Andriani & D’Alessandro 2022). We argue that, depending on their specific/referential vs generic/kind-interpretation, those nouns surfacing ‘unexpectedly’ bare either undergo N-to-D (Longobardi 1994) vs N-to-Num (Alexoupoulou et al. 2013) movements, respectively.

From a broader diachronic and typological perspective, definite articles are the product of grammaticalisation and are only present in roughly half of the world’s languages (308 on 620 surveyed languages on the WALS). Being heritage contexts the perfect circumstances for linguistic changes to become visible in synchrony (Kupisch & Polinsky 2022), these Italo-Romance heritage grammars point towards an incipient shift to a new parametric (re)setting in these contact grammars (if transmission continues). According to the principles ‘avoid indeterminacy/silence’ discussed in Andriani et al. (2022; and references therein), we would expect the overproduction of articles to prevent ambiguity or to avoid ‘overload’ at the interface with pragmatics (cf. the overgeneralization of subject pronouns; cf. Sorace 2004); in contrast, we observe that HSs favour the opposite tendency, showing a more permissive occurrence of articleless, bare nouns. If we assume that syntax needs a morpho-phonological trigger to change – otherwise it tends to be ‘inert’ (Longobardi 2001, a.o.) – evidently, silence, is one of the cues that these HSs began to gradually generalise, leading to this incipient parametric change in their grammars which requires the lexicalisation of D in increasingly less contexts.

Tuesday, 20 February 2024 (12-1pm, UK time)

Antonio Fábregas (Norwegian University of Science and Technology) – Person features and non-specific readings of speech participants

Abstract: The goal of this talk is to explore the connection between formal features and pragmatics through the study of three types of non-specific subject structures in contemporary Spanish: 3pl non-pronominal subjects (1a), 2sg subjects (1b) and uno ‘one’ (1c).

(1)       a.         Dicen que estás enfermo.

                        say.3pl that are.2sg sick

                        ‘It is said that you are sick’

            b.         Si tienes padres ricos, todo es más fácil.

                        if have.2sg parents rich, all is more easy

                        ‘If one has rich parents, everything is easier’

            c.         Uno piensa que estas cosas no van a pasar.

                        one think.3sg that these things not going to happen

                        ‘One does not expect such things to happen’

We will concentrate on two properties of these structures, both related with speech participants. 2sg forms can get a non-specific interpretation, but 2pl forms cannot:

(2)       #Si os casáis en una iglesia, es todo más bonito.

              if you marry.2pl in a church, is all more nice

            Intended: ‘If any two people get married in a church, the wedding is nicer’

            Actual reading: ‘If you two get married in a church, the wedding is nicer’

The indefinite in (1c) can be used as a covert 1sg, as in (3), but not as a covert 2sg.

(3)       Uno ya no sabe qué hacer.

            one already not knows what to.do

            ‘I do not know what to do any longer’

            *’You do not know what to do any longer’

Through the analysis of these distinctions, we will argue in favor of a syntactic system that specifically formalizes anchoring to a logophoric context (presumably through FinP or an equivalent projection), and show that such an approach can integrate the pragmatic and cognitive differences between these three types of non-specific subject within a formal analysis, at the same time that it complements the cognitive explanations provided in pragmatic studies. The main conclusion is that syntax should not formalize pragmatic notions through individual features; rather, the interpretation should emerge globally from the complex interaction of formal features and the operations that they trigger.

Tuesday, 5 March 2024 (12-1pm, UK time)

Francesco Gardani & Chiara Zanini (University of Zürich) – A baseline for object clitic climbing in Italian: Preliminary results

Abstract: Object clitic climbing (OCC) is the phenomenon whereby a clitic pronoun realizing a direct object (do) selected by a complement verb moves out of its local domain and climbs to the matrix clause attaching to a host verb. This talk focuses on OCC as it occurs in the presence of restructuring verbs, in Italian. Consider the following data.

1) a. Marco vuole comprar=lo

Marco wants buy.inf=do

‘Marco wants to buy it.’

b. Marco lo=vuole comprare.

Marco do=wants buy.inf

‘Marco wants to buy it.’

In both (1a) and (1b), the do is expressed by lo: in (1a), lo is enclitic to the non-finite verb, while in (1b), which instantiates OCC, it procliticizes to the verb head, viz. the inflected verb vuole ‘wants’, of which it is not an argument—lois, in fact, the argument of the embedded infinitive (inf) comprar(e) ‘to buy’. The sentences in (1a) and (1b) have the same interpretation, both semantically and pragmatically.

In spite of a considerable amount of research on clitic climbing (e.g., Cardinaletti & Giusti 2022; Cardinaletti & Shlonsky 2004: 574; Cinque 2004; Monachesi 1993; Pescarini 2021: 235–266), we know very little about variation, in terms of where in the Italian-speaking areas and to what extent OCC occurs. We know virtually just as much as that (a) “in Italian, clitic climbing is optional” (Monachesi 1995: 6, fn4), and (b) the type “Carlo lo andrà a comprare è più usata nell’italiano toscano e centro-meridionale; [while the type] Carlo andrà a comprarlo è più usata in quello settentrionale” (Calabrese 2001: 586).

Our goal is to fill this massive descriptive lacuna, detecting a robust baseline for OCC in (standard) Italian. To this end, we designed a questionnaire containing eleven questions targeting the informant’s sociolinguistic profile and a production task. The production task contained 144 stimuli, of which 48 experimental stimuli and 96 fillers (1:2 ratio). The dependent variables were E (occurrence of enclisis), P (occurrence of proclisis), O(ther). The independent variables were verb type (including modal, motion, conative, aspectual), number value of the clitic, gender value of the clitic, animacy value of the clitic’s referent, type of subject (NP vs null) in the clause; age group, current place of living, place of living before age 14, geographic area of Italo-Romance variety (IRV), frequence of use of IRV, origin of parent1, origin of parent2.                                                   

The questionnaire was administered online via the LimeSurvey app. We collected data from over 300 informants having Italian as L1 and resident in 82 Italian provinces (homogeneously distributed across the country) and the Italophone areas of Switzerland. For the purpose of the present presentation, we analyzed 238 questionnaires. Analysis was conducted by means of the R software (R Core Team 2020), conditional inference trees and random forest models; for the latter two we used the party package in R (Hothorn et al. 2006; Strobl et al. 2007; Strobl et al. 2008).

The analysis of the conditional inference tree model shows that, in terms of predictive power over the sample, the most relevant predictor is (a) age, with older generations having higher probability of using OCC than younger generations; (a) is followed by (b) verb type (for younger generations), with  aspectual, modal, and motion verbs having higher probability of licensing OCC than conative verbs, on par with (c) frequence of use of IRV (for older generations), with those using an IRV more frequently having higher probability to produce OCC. The analysis of the random forest model (2000 trees) is coherent with the tree model results, in that it shows that the linguistic variables are not predictive, whereas the sociolinguistic variables are. In the talk, these and other results are discussed, by keeping in mind that they are preliminary and that for a robust generalization, we need to reach the 300-questionnare threshold.

Calabrese, Andrea. 2001. I pronomi clitici. In Lorenzo Renzi, Giampaolo Salvi & Anna Cardinaletti (eds.), Grande grammatica italiana di consultazione, vol. 1: La frase. I sintagmi nominale e preposizionale, 563–606. Bologna: Il Mulino.

Cardinaletti, Anna & Giuliana Giusti. 2022. Dependency, licensing, and the nature of grammatical relations. In Adam Ledgeway & Martin Maiden (eds.), The Cambridge Handbook of Romance Linguistics, 604–636. Cambridge University Press.

Cardinaletti, Anna & Ur Shlonsky. 2004. Clitic positions and restructuring in Italian. Linguistic Inquiry 35, www.jstor.org/stable/4179295.

Cinque, Guglielmo. 2004. “Restructuring” and functional structure. In Adriana Belletti (ed.), Structures and beyond. The cartography of syntactic structures, volume 3 (Oxford Studies in Comparative Syntax), 132–191. Oxford & New York: Oxford University Press.

Hothorn, Torsten, Peter Bühlmann, Sandrine Dudoit, Annette Molinaro & Mark J. van der Laan. 2006. Survival ensembles. Biostatistics (Oxford, England) 7(3). 355–373.

Monachesi, Paola. 1993. Object clitics and clitic climbing in Italian HPSG grammar. 437–442.

Monachesi, Paola. 1995. A grammar of Italian clitics. Tilburg: Universiteit Tilburg PhD dissertation.

Pescarini, Diego. 2021. Romance object clitics. Oxford: Oxford University Press.

R Core Team. 2020. R: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria, https://www.R-project.org/.

Strobl, Carolin, Anne-Laure Boulesteix, Thomas Kneib, Thomas Augustin & Achim Zeileis. 2008. Conditional variable importance for random forests. BMC Bioinformatics 9. 307.

Strobl, Carolin, Anne-Laure Boulesteix, Achim Zeileis & Torsten Hothorn. 2007. Bias in random forest variable importance measures: illustrations, sources and a solution. BMC Bioinformatics 8. 25.

Tuesday, 12 March 2024 (12-1pm, UK time)

Angela Castiglione (University of Messina) – Il sistema di genere nelle varietà di minoranza gallo-italica in Sicilia.

Abstract: Le minoranze galloitaliche di Sicilia (San Fratello, Novara con Fondachelli-Fantina, Nicosia, Sperlinga, Aidone e Piazza Armerina) sono ciò che resta di un vasto flusso migratorio che, tra la fine dell’XI e la metà XIII secolo, dall’Italia nord-occidentale portò all’insediamento di gruppi di coloni in svariati centri e casali dell’isola (Petracco Sicardi 1969; Varvaro 1981: 185-196; Pfister 1988; Trovato 1998).

Lo studio oggetto del seminario intende esaminare il sistema di genere di tali minoranze. Considerato il definitivo distacco dalla terra d’origine e la plurisecolare convivenza con il siciliano che, almeno fino a qualche decennio fa, ha rappresentato il codice di prestigio e la lingua socio-culturalmente dominante (Tropea 1966; 1970; 1974), le indagini sulle grammatiche delle varietà gallosiciliane non possono prescindere dalla considerazione del contatto linguistico e delle dinamiche di assimilazione vs. resistenza rispetto alla lingua egemonica. In ragione di questo prolungato contatto e in situazioni di diffuso bilinguismo anche di carattere monocomunitario, le varietà galloitaliche si sono avvicinate strutturalmente al siciliano attraverso un processo di advergenza (Mattheier 1996; Berruto 2006), che ha però investito in maniera disomogenea i diversi livelli del sistema linguistico. Come ha recentemente spiegato De Angelis (2023), mentre la sintassi – via “dominanza linguistica” e “agentività” della lingua fonte (nei termini di Van Coetsem 1995; 2000; vd. anche Winford 2005) – si è completamente rifoggiata sul modello siciliano, soprattutto la fonetica e la fonologia hanno mantenuto una facies di tipo altoitaliano. Tale conservazione viene spiegata, nel quadro ricostruito da De Angelis, alla luce di fattori di natura più squisitamente sociolinguistica, e più precisamente sociofonetica, riassumibili nel valore identitario assegnato dalle comunità alloglotte alle specificità fonetiche della propria parlata, come reazione a una condizione in cui il siciliano tendeva ad assimilare a sé la lingua di minoranza.

Ora, l’ambito – ancora inesplorato per questi dialetti – relativo al sistema di genere fa emergere ulteriori aspetti meritevoli di attenzione sia dal punto di vista strutturale sia dal punto di vista delle dinamiche del contatto, mettendo in luce significative differenze all’interno delle stesse varietà di minoranza e permettendo di individuare nel dialetto di San Fratello alcuni pattern estremamente conservativi riguardo alla supposta area di provenienza e all’epoca di emigrazione. Tali pattern risultano divergenti rispetto al modello siciliano, ma anche rispetto agli altri dialetti gallosiciliani, facendo della varietà sanfratellana il principale baluardo di resistenza dell’alloglossia. In estrema sintesi, impiegando gli strumenti teorici e operativi messi a punto dal filone di studi che fa capo a Corbett (cfr. almeno Corbett 1991; Corbett/Fraser 2001; Loporcaro 2018), è possibile rilevare – attraverso l’analisi delle proprietà di accordo dei controllori e dei bersagli – che, al pari del siciliano, i dialetti di minoranza galloitalica (ad eccezione del sanfratellano) configurano stabilmente un sistema di genere di tipo binario convergente al plurale:

Il sanfratellano, al contrario, presenta un sistema di tipo binario pienamente parallelo, prevede, cioè, con particolare compattezza e regolarità, anche al plurale forme di accordo dedicate dei suoi bersagli e, in più, conserva molto bene un sistema di genere “alternante” per quel gruppo di nomi che continuano i plurali neutri II decl. del latino (es. sanfr. li_fpl assi, carni, dinuogi, li burieḍi, uovi ecc. ‘le ossa, corna, ginocchia, budella, uova’) e che nell’Alta Italia ha progressivamente perso terreno fino a dissolversi quasi del tutto (cfr. Loporcaro 2018: 87-91).

 A tal riguardo, il sanfratellano rispecchia condizioni settentrionali più antiche, di cui oggi è rimasta qualche traccia residuale in dialetti di area emiliana e ligure, ma che in passato avevano più ampia diffusione, in quanto esso conserva il tipo ‘l’osso/le osse, il braccio/le bracce ecc.’ documentato in testi medievali di varia provenienza: es. ant.lomb. le osse, legne, brace; ant.padov. le brazze, buelle, cegie; ant.venez. le ose, legne, castelle, ovvero nelle stesse zone in cui ha in seguito prevalso il tipo ‘il braccio/i bracci’ (Rohlfs 1968: § 369).

Nel corso del seminario si illustreranno nel dettaglio e in maniera più approfondita i dati relativi all’accordo di varie tipologie di bersagli (articoli, aggettivi, possessivi, dimostrativi, il quantificatore ‘due’ ecc.) nel sanfratellano rispetto al siciliano e alle altre varietà galloitaliche della Sicilia e della presunta area altoitaliana di origine. Dal confronto emergeranno una notevole conservatività e una tenace opposizione al contatto di questo dialetto in merito al sistema di genere. Considerazioni di ordine sociolinguistico e testimonianze di varia natura indicano che una maggiore resistenza strutturale del galloitalico proprio nella comunità di San Fratello possa essere stata favorita da un maggiore grado di consapevolezza della propria diversità linguistica e da un senso, più spiccato che altrove, di “fedeltà” nei riguardi della varietà alloglotta.

Tuesday, 19 March 2024 (12-1pm, UK time)

Fabienne Martin (Utrecht University) – The ‘no-agent’ scalar implicature triggered by anticausatives is stronger when the causative alternative is structurally-defined

Joint work with Florian Schäfer, Despina Oikonomou, Felix Gölcher and Artemis Alexiadou

Abstract: Weak scalar expressions like English It is possible that P (defeasibly) implicate the negation of their stronger alternatives such as It is certain that P (Grice 1967, Horn 1972, Gazdar 1979 a.m.o). When the implicature triggered by the weak scalar item is not satisfied in the context, infelicity may arise (Noveck 2001 a.m.o.). Schäfer and Vivanco (2016) propose that anticausative  expressions such as in (1a) form scales with their (lexical) causative counterparts as in (2) (<break(y), break(x,y)>). Under this view, AC statements should exhibit a similar behavior as other items triggering scalar implicatures: they  should be felt less natural in a context fulfilling the stronger alternative  (e.g., to describe a broken window and a smiling boy  with a slingshot in the hand in front of it), because the AC alternative is too weak in such contexts as it triggers a `no-agent’ scalar implicature (SI), i.e. an inference that there is no agent involved in the event denoted by the AC (see (1a/b)).

(1a) The window broke.

(1b) NOT(Someone broke the window.)

 In English,  AC and  causatives, however, are not of equal formal complexity: While ACs involve a vP denoting a set of events endured by the theme argument, causatives have on top of this vP a Voice-projection introducing an external argument variable. English AC therefore do not have  causative counterparts as  structurally-defined alternatives  (Katzir 2007, Katzir and Fox 2011).  More complex structures do not count as alternatives,  unless they are salient in the discourse (i.e. are contextual alternatives). 

Languages like French differ from English in that a subset of their ACs receive morphological marking (se  in French), either optional or compulsory, depending on their morphological class. We  report experiments testing the hypothesis that the no-agent scalar implicature triggered by AC-statements  is stronger when the corresponding causative statement is a structurally-defined alternative. Results show that indeed,  French  marked ACs  trigger a stronger no-agent scalar implicature  than unmarked ones (in French or in English).

Michaelmas Term 2023

Tuesday, 5 December 2023 (12-1pm, UK time)

Cecilia Poletto – Subjects and Topics: a complicated relationship

Abstract: The problem of the position of lexical subjects in the pro drop Romance languages has often been addressed by arguing that since pro always occupy the SpecT position, the lexical subject occurs in a Topic position. Arguments in favor of this analysis are: a) subjects are often doubled in the languages that have subject clitics similar to agreement markers (Northern Italian and colloquial French) b) the Subject criterion proposed by Rizzi in various articles is similar to other criteria, all of the occurring in the left periphery. On the other hand, the idea that subjects are automatically topic elements, and as such occur in a left dislocation position cannot be correct since quantifiers, which are notably not prone to become topics, can be perfectly well-formed preverbal subjects. On the other hand, various authors have proposed that the higher subject position SubjP is located in the high IP area, but this hypothesis is not corroborated by the fact that in some Romance languages like Occitan, Guascon, Piedmontese and Ligurian lexical subjects must indeed occur higher than complementizers. Furthermore, there is clear evidence from Germanic agreeing complementizers that some subject features must reside in the left periphery. In this talk I propose that subjects are located in the CP area, as the data on complementizer ordering clearly show, but are not necessarily in the typical Topic positions where left dislocated elements are. Furthermore, I will show that preverbal quantificational subjects have a different agreement pattern (both in terms of subject clitic doubling and verbal agreement) with respect to definite subjects and are located in the same left peripheral subfield where wh-items and Focus are re-merged. To explain these fact, I will propose an analysis similar to the one originally proposed by Beghelli and Stowell (1997), who show that the position of quantifiers is related to their semantic interpretation, with definite nominal expressions raising higher than all types of quantifiers. The same can be shown to be the case for different types of subjects, so there is not only one SubjP, there are several positions where different types of subjects are semantically interpreted. This will open up the possibility that all nominal expressions have different positions, mirroring their distinct interpretation, as has already been noticed for DOM objects in SOV languages like Turkish, which are higher than non-DOM objects or quantifier objects like “was” in languages like German and Dutch, which must remain lower than definite objects. This will pave the way to reconsidering Case and Agreement as highly ambiguous categories, which blur semantic distinctions that are indeed reflected in the various positioning of different types of subjects/objects. Nonetheless, Agreement can be ultimately considered to be related to the interpretation of subject/objects and is not a procedure void of any meaning which cannot project its own functional projection.       

Tuesday, 28 November 2023 (12-1pm, UK time)

Ziwen Wang – Individuation and genus alternans in Romance

Abstract: Of all the phenomena related to nominal inflectional morphology in Romance languages, genus alternans can undoubtedly be considered the most problematic and controversial one. The term genus alternans (i. e. alternating gender) refers to an inflectional pattern such that nouns select masculine agreement in the singular but feminine agreement in the plural. This phenomenon is widely attested across Eastern Romance. For example, Romanian is well-known for having a substantial group of nouns displaying this unusual pattern, as exemplified below by the noun scaun (‘chair’) (cf. Maiden 2016b: 113): 

The rise and fall of genus alternans has been discussed in Romance diachronic linguistics almost exclusively from a comparative perspective or with reference to the history of individual languages. There is consensus that Romancegenus alternans stemmed from the Latin neuter gender, which in most Eastern Romance languages adopted feminine agreement in the plural (e. g. Lat. BRACCHIUM/A > It. il braccio/le braccia ‘arm/arms’), but masculine agreement in other Romance varieties, the latter involving the complete merger of neuter nouns with masculines (e. g. Lat. BRACCHIUM/A > Sp. el brazo/l-os brazo-s ‘arm/arms’). Nevertheless, a recent typological survey by Loporcaro (2018) provides evidence that genus alternans, which first emerged in Late Latin, survived not only into Eastern Romance, but also into the older stages of Raeto-Romance and Gallo-Romance, thus suggesting that the preservation of alternating agreement is an early development common to all Romance languages. This hypothesis is further supported by new evidence discovered by myself from early Ibero-Romance varieties, where genus alternans, contrary to conventional reconstruction, is consistently detected across the 9th-10th century Iberian peninsula (e. g. m. sg. animal/f. pl. animalias ‘animal’). These facts strongly suggest that “le neutre comme partie intégrante du système a persisté jusqu’à la veille de la phase romane et même au-delà” (Väänänen 1967: 109). 

    The common preservation of genus alternans across Romance languages seems to be a necessary consequence of the scale of individuation proposed by Grimm (2018). In this talk, I will provide evidence showing that, in the diachrony of Italian and Spanish, genus alternans nouns denoting lesser-individuated entities commonly remained resistant to morhological levelling and retained feminine plurals due to the unmarked status of the plural (e. g. It. m. sg. il corno/f. pl. le corna), whereas others denoting highly individuated entities massively adopted masculine plurals after levelling in favour of the unmarked singular. I will also tentatively assess whether such generalization holds true throughout the diachrony of other Western Romance varieties, such as Gallo-Romance, Sardinian, Occitan, Italo-Romance dialects, etc., with further data collected from texts or corpora.

Tuesday, 21 November 2023 (12-1pm, UK time)

Anne Abeillé- Closest conjunct agreement in French (De certaines règles et usages du français)

Abstract: Recent work (corpus based and experimental) has shown that contemporary French has kept (feminine) closest agreement, (certaines régions et départements ‘certain.fpl region.fpl and department.mpl) (Abeillé et al 2021, 2022), despite the prescriptive norm towards masculine agreement in case of mixed gender conjuncts. We examine the factors favoring one or the other, building on Corbett (2023)’s agreement hierarchy. We show that other factors (directionally, humans) play a role as well, as in other languages.

Tuesday, 14 November 2023 (12-1pm, UK time)

Clarissa Facchin – Empty head vs verb movement: an empirical study on Complementizer Deletion in Spanish

Abstract: The existence of subordinate clauses lacking an overt C element is a well-known fact for many languages. Different proposals have been advanced to account for these structures, and many of them involve different degrees of defectivity assumed for the CP layer. On the one hand, some authors claim that the absence of a complementizer is due to the absence of the CP in the derivation (see Bošković (1997) and Stowell (1981) for English; Brovetto (2002) for Spanish). This hypothesis is known as the IP-hypothesis. By contrast, other works defend the CP-hypothesis, which comprises very different proposals depending on the functional projections assumed in the CP area. The C layer has been either analysed as a pruned domain, which projects only its lower part (see Giorgi & Pianesi (1996) and Giorgi (2010) for Italian), or it has been characterized by a syncretism of projections (Poletto (1995) for Italian; Llinàs i Grau & Fernández Sánchez (2013) for Spanish; Kishimoto (2006) for Kansai). According to the CP account, the embedded verb works as an alternative checker of the relevant feature of the complementizer. Therefore, complementizer deletion is allowed by an I-to-C movement of the verb to check the feature of the missing complementizer. The existing proposals on Spanish highlight that the embedded CP cannot host left-peripheral material if the complementizer has been omitted, a restriction found also in English and Italian complementizer-less clauses. However, a constraint which only Spanish manifests is the adjacency between the main and the embedded verb, which means that complementizer deletion is incompatible with overt preverbal subjects (Torrego 1983; RAE 2010; Brovetto 2002; Antonelli 2013; Llinàs i Grau & Fernández Sánchez 2013).  

1. a. Espero *(que) los problemas causados por el huracán       se solucionen        pronto.     
   I-hope   that   the problems   caused     by  the hurricane  CL solve-SBJV-3pl soon    
   ‘I-hope that the problems caused by the hurricane will be solved soon.’ 

b. Espero (que) se  solucionen         pronto los problemas causados por el huracán. 
I-hope   that  CL solve-SBJV-3pl  soon    the problems   caused     by  the hurricane   
‘I-hope that the problems caused by the hurricane will be solved soon.’ (Brovetto 2002) 

In this talk, after discussing the main analyses on the complementizer deletion phenomenon in Spanish, I will explore the extent to which they can account for empirical data, which I gathered from the Corpus del Español del Siglo XXI and the Corpus de Referencia del Español Actual. I will show that subjects occupying the preverbal embedded position of a complementizerless clause are actually attested. Therefore, the main goal of this study is, on one hand, to analyze the interpretation that preverbal elements receive, and on the other, to determine whether the embedded verb has moved from the I-domain, in an attempt to sketch a configuration of the C-field in these environments.  Basing on the empirical evidence that an overt subject can occupy its basic position within the embedded clause, and low adverbs preceding the verb are attested, I will suggest that the embedded verb of a complementizerless clause remains in a low position, and no I-to-C movement occurs. I will then show that the left-peripheral elements, as well as high adverbs (Cinque 1999), cannot occur before the embedded verb, and that only an unmarked intonation can be associated to the subjects and adverbs appearing in preverbal position.  The picture that emerged from the data leaves us two different ways of explaining the configuration of the C-field of Spanish complementizer-less clauses: either by assuming a pruned empty C-layer preserving only FinP (Rizzi 1997), or by treating these environments as asyndetic constructions lacking their CP (Brovetto 2002). Both analyses are able to justify the impossibility for left-peripheral material to occur, but only by assuming the presence of FinP we are also able to explain the mood selection by the governing predicate.  

Tuesday, 31 October 2023 (12-1pm, UK time)

Raluca Braescu & Irina Nicula – Invariable adjectives which undergo morphological recategorisation

Abstract: In the current article the authors tackle the enrichment of the class of invariable adjectives. Given the fact that invariability is considered a weakness in the inflectional system, the extension of this class of units with a marginal status raises the impredictibility area in the inflection of the adjective. We focus on the common path of certain units in spoken Romanian (such as criță ‘steel’, sloi ‘ice block’, bâtă „bat”, lună  ‘moon’,  bec ‘bulb’, lulea ‘cigar’, pușcă ‘shot gun’, măr ‘apple’, cuc ‘cuckoo’ , cobză ‘kobsa’, scândură ‘plank’, belea ‘trouble’, bombă ‘bomb’,etc.)  and we delineate more stages: first, we deal with nouns which occur in “false”-comparative structures; then they undergo a process of partial desemantisation, after which they become inflectionally rigid, and, due to a process of metaphoric transfer, they become intensifiers of certain adjectives; in the third stage, the aforementioned nouns enter recategorization, they become invariable adjectives; finally, some of these forms become adverbs and they acquire a purpose reading (accompanying verbs) or they enter adverbial patterns (adverb + de + adjective).  

The pattern with nouns which recategorize and acquire the superlative meaning is vivid and very productive in current Romanian and is characteristic of the colloquial, spoken language. 

Tuesday, 24 October 2023 (12-1pm, UK time)

Michelle Sheehan – Clitic placement in European Portuguese: the picture from non-finite clauses

Abstract:

 European Portuguese, Galician and Asturian are well known to be unusual among modern Romance varieties in having a complex pattern of clitic placement. With inflected verbs, proclisis is triggered in the presence of: (i) quantifier phrases (including wh-phrases, negative phrases, contrastive foci) (ii) overt complementisers, (iii) negation, (iv) light adverbs (see Madeira 1992, Martins 1994, Barbosa 1996, Duarte & Matos 2000, Raposo and Uriagereka 2005, Fernandez-Rubiera 2009, Luís and Kaiser 2016, and many others). Enclisis occurs in all other contexts and has been argued to be the default pattern (Costa, Fiéis & Lobo 2015). 

(1) Como te posso ajudar? 

how you= can.1SG help 

‘How can I help you?’ 

(2) Acho que te vou ajudar. 

think.1SG that you= will.1SG help.INF 

‘I think I’ll help you.’ 

(3) Não te vou ajudar. 

NEG you= will.1SG help.INF 

‘I won’t help you.’ 

(4) Já te ajudei. 

already you= helped.1SG 

‘I’ve already helped you.’ 

Adapting Fernández-Rubiera’s (2009, 2010) analysis, I argue that these languages have what I call an anti-V2 requirement whereby either a phrase or head must raise to the C-domain. Where nothing else raises, the verb itself raises to C, resulting in enclisis. This captures the patterns in (1-(4) as all involve material either base generated or moved to the C-domain. In my talk, I will defend this syntactic account of clitic placement, further arguing that EP clitics originate in argument positions and raise to spec TP because of a [uT] feature. 

The evidence comes from uninflected infinitives. With these verb forms (in EP, at least) enclisis becomes possible as an option in contexts (i)-(iii), whereas proclisis remains obligatory in (iv) (see Raposo & Uriagereka 2005, Martins 2013): 

(5) Não temos a quem nos= queixar / queixar = nos}. 

NEG have.1PL DAT who ourselves= complain complain= ourselves 

‘We don’t have anyone to complain to.’ (Martins 2013: 2283) 

(6) [Para { a= ver / vê= la} outra vez], … 

for her= see see= her another time … 

ʻIn order to see her one more time ….ʼ (Raposo & Uriagereka 2005:677) 

(9) [Não { a= convidar /convidá =la} para a nossa festa] 

NEG her= invite invite =her for the our party 

seria uma boa ideia. 

would.be a good idea 

ʻNot to invite her to our party would be a good idea.ʼ (R&U: 2005:669) 

(7) Apesar de já {se= encontrar / *encontrar =se aberto ao público… 

despite of already itself= find find =itself open to.the public 

‘Despite already being open to the public…’ 

The class of adverbs with this behaviour is those which have been argued to be syntactic heads and which can function as the answers to Y/N questions (Castro & Costa 2003, Santos 2009). Furthermore, the addition of a light adverb to any of the contexts in (i)-(iii) also leads to obligatory proclisis. This pattern, I argue, clearly shows that EP clitic placement is syntactic, rather than prosodic. I sketch an analysis based on the idea that EP clitic pronouns are base generated in argument positions with a [uT] feature. This means they must either move to spec TP as XPs or cliticise onto a verb which then moves to/through T. Following Rapsoso and Uriagereka (2005), low attachment to verbs is possible with uninflected infinitives. The difference between (iv) and (i-iii) is that, as heads below T, light adverbs block head movement to T, making clitic movement to spec TP the only convergent derivation (resulting in proclisis). 

Tuesday, 17 October 2023 (12-1pm, UK time)

Fabian Helmrich – Aspectual Suppletion in Istro-Romanian – Fact or Fiction?

Abstract: In Istro-Romanian, there are two main groups of verbs, one inherited from Romance, one borrowed from the co-territorial Chakavian Croatian variety. The Croatian-derived verbs are specified for aspect, while the Romance ones are not. Generally, these groups do not interact. However, for a number of verbs previous scholars have claimed that they form suppletive aspectual pairs consisting of a prefixed Croatian-derived perfective verb and a simplex Romance imperfective. Given the anaspectual status of most Romance verbs such an analysis deserves scrutinous empirical verification. We surveyed a comprehensive corpus of Istro-Romanian spanning from the second half of the 19^th century to the 21^st century with respect to seven alleged suppletive pairs. After describing their behaviour with respect to Vendlerian verb classes, as well as the semantics of their arguments, we find a murky and varied picture with respect to aspect, telicity and semantic specialization. This talk will shed light on some of the issues surrounding the notion of aspectual suppletion in Istro-Romanian and suggest why the notion may need to be abandoned altogether.

Tuesday, 10 October 2023 (12-1pm, UK time)

Leonardo Russo Cardona (University of Cambridge) – Two types of tough-constructions in Italo-Romance

Abstract: In tough-constructions (TCs), the surface subject of the matrix predicate (typically a copular structure) appears to function also as the object of the embedded infinitive. The canonical TC in Standard Italian can be exemplified by (1):

1)    Questi libri sono difficili da leggere. 

‘These books are tough to read’.

I collected data from a broad range of Italo-Romance varieties to investigate the variation in the syntax of TCs, focusing on two main aspects: 

i)               The introductory element and the size of the embedded clause, assessing the presence of CP and TP projections;

ii)             The availability and licensing of a resumptive object clitic or an object-promoting structure in the embedded clause.

The observed pattern of variation can be summarised in terms of 2 main types. In Type 1 TCs, the embedded clause has avery reduced functional structure and an infinitive verb without passive markers/object clitics (despite its passive syntax). On the other hand, the embedded clause of Type 2 TCs has a richer functional structure and an obligatory passive marker or object clitic. In other words, the two logically independent parameters (i-ii) seem to be robustly correlated: I provide a syntactic explanation for this connection, capitalising on the role of Voice. This analysis can also account for why a structure with a rich functional structure and no passive marking nor object clitic is consistently ungrammatical.

(October) 2022 – (June) 2023

Easter Term 2023

Tuesday, 25 April 2023 (12-1pm, UK time)

Alice Corr (University of Birmingham) & Onkar Singh (University of Cambridge) – Adjective-Noun orders in Turkish Judeo-Spanish: a case of conservatism or innovation in the nominal domain?

Abstract: By means of novel diachronic and synchronic dialectal data from Judeo-Spanish dialects -a variety for which there exists a notable paucity in dialectal data and proposals within the scholarship-, this paper explores the case of Adjective-Noun orders attested in the Turkish dialect of Judeo-Spanish. Frequently attested in speakers of Turkish Judeo-Spanish post-1930, these nominal patterns contrast with the expected Noun-Adjective orders otherwise generally attested in southern Romance, whereby it is assumed that these varieties display high N-to-D movement of nominals within the DP (viz. Strong D-languages à la Ledgeway 2015; see also Ledgeway 2020). The orders found in Turkish Judeo-Spanish are striking whereby: (i) they are indicative lower noun movement within the nominal spine à la Cinque (1999), (ii) they are not attested in other dialects of Judeo-Spanish such as those spoken in the Balkans and those in Northern Africa, and (iii) modern Turkish has a default Adj-.N order in the DP. By means of empirical data and qualitative analyses, this paper thus examines whether these Adj.-N. Orders are instances of calquing from Turkish or conservatism from old Spanish, whereby old Romance has been suggested to have had a less systematic noun movement (Ledgeway 2007, 2009). We propose these orders to be lower noun movement as part of a syntactic calque from Turkish among L1 Turkish speakers after 1930, i.e., the date when Turkish became the obligatory language of instruction in the newly-formed Republic of Turkey. Strikingly, the data from before this date shows the expected Romance pattern of higher noun movement in Turkish Judeo-Spanish, also patterning with 15th century old Spanish prior to the expulsion of the Jews from the Iberian Peninsula (Singh, in prep.). All in all, this paper adds to a growing literature that bridges the gap between central Ibero-Romance and Ladino varieties, whilst also adding to an increasing scholarship on internal variation in noun movement within the DP.

Tuesday, 2 May 2023 (12-1pm, UK time)

Víctor Acedo-Matellan (University of Oxford) & Isabel Oltra-Massuet (Universitat Rovira i Virgili), Prepositional verbs in Catalan: From event to argument structure

Abstract: Drawing on Catalan, we show how the aspectual classification of intransitive prepositional verbs—pensar en ‘think of’, al·ludir a ‘allude to’, patir de ‘suffer from’, etc.– is, partially, a predictor of their argument structure properties. Prepositional verbs qualifying as stage-level form a heterogeneous class, comprising both unergative and unaccusative verbs. By contrast, those prepositional verbs that qualify as individual-level predicates are homogenous as regards argument structure: their prepositional complement is robustly obligatory and their subject is not an external argument, suggesting that they involve an underlying small clause configuration whose predicate is the prepositional complement. This result supports a more general constraint against the encoding, as verbs, of individual-level property predicates.

Tuesday, 9 May 2023 (12-1pm, UK time)

Alina McLellan (University of Manchester) – ‘Sak’ relatives in Reunion Creole: light-headed or free?

Abstract: In this talk, I investigate the structure of free relative clauses in Reunion Creole (RC), a French-based Creole spoken on Reunion Island, using new data from corpora and fieldwork. The most frequent strategy for forming free relatives in RC is with sak, which has phonological variants sék, sat, sét and sad.

• Ti-Pierre     té       i    agard   trankiman         sak      té           i     espas.

Little-Pierre  ipfv   fin  watch   peacefully        what    ipfv       fin        happen

‘Little Pierre was peacefully watching what was happening.’  (Corpus, short story)

In the literature, free relatives are often described as wh-clauses because, in many languages, free relative pronouns are identical to interrogative pronouns. However, sak and variants are not interrogative pronouns, and in fact, interrogative pronouns receive low acceptability ratings (subject to syntactic, and possibly semantic, constraints) in free relatives in RC (McLellan 2023a). Instead, the form sak originates from the combination of a demonstrative (sa) and a complementiser (k), which formed a light-headed relative clause (a relative clause with a pronominal antecedent, cf. Citko (2004) among others). I argue that this once light-headed structure is developing into a true free relative structure, where sak and variants are free relative pronouns. However, at the same time that this new free relative pronoun was forming in RC, so too was a new demonstrative pronoun with the same five variant forms (sak, sék, sat, sét and sad).

• Pran     loto-la,             sat          papa    lé                     kasé.

take     car-dem            dem      dad     cop       broken

‘Take that car, Dad’s is broken.’ (Armand 2014)

The existence of this demonstrative form is relevant to the relative clause context because I have found evidence of a second light-headed relative structure, where the new demonstrative pronoun is the head of the relative clause. The structure in (3), where sat is followed by a relative complementiser, is accepted by native speakers:

• Sat       ke   mwin    la      vi             yèr                      lété                              shèr.

dem      rel 1sg        prf       see       yesterday         be.ipfv expensive

‘The one that I saw yesterday was expensive.’ (Constructed; accepted in interviews)

Structures like that in (3), with a phonetically realised relative complementiser, did not occur in the corpus. However, their acceptance by native speakers raises an analytical question about sak-relatives without an overt relative complementiser (e.g. (1)): whether such relatives are light-headed relatives with a zero-marked relative clause (which is very common for headed relatives in RC (McLellan 2023a, 2023b)), or whether they are truly free relative clauses with sak (and variants) as free relative pronouns. Using Role and Reference Grammar (Van Valin & LaPolla 1997; Van Valin 2005, 2008; Bentley et al. forthcoming), I illustrate the differences in syntactic structure between light-headed relatives and true free relatives. I argue that sak-relatives of three different types exist or have existed at a previous stage of RC: a light-headed relative headed by sa (which is now largely obsolete), a light-headed relative headed by sak (and variants), and a true free relative. I argue that the two light-headed structures constitute two channels of grammaticalisation working towards the same result: the emergence of sak as a free relative pronoun. The diverse set of data and the co-existence of these three structures is reflective of the high degree of variation in RC and suggests that these structures are in the process of grammaticalisation, which might be at different stages for different linguistic groups or individuals on the island.

Tuesday, 16 May 2023 (12-1pm, UK time)

Imanol Suárez-Palma (University of Florida) – On dative possessors and middle-passive constructions in Spanish

Abstract: In this talk, I will revisit the phenomenon of inalienable possession (and lack thereof) and external possession by looking at data from different types of Spanish middle constructions. I will address issues pertaining to information structure and argument structure. First, I will look at middle-passive sentences, i.e., generic predicates denoting intrinsic properties of the verb’s internal argument, which surfaces preverbally as the grammatical subject due to its status as a sentential topic (Sánchez López 2002). When the theme in these sentences is a relational/body-part noun, it can be inalienably possessed internally or externally, by means of a possessive determiner (1a) or a dative possessor (1c-f), respectively; the dative DP in these contexts tends to occur preverbally (1c-e). 

(1)     What happens? 

          a.  (Que) sus cicatrices se ven fácilmente. 

          b.  ?(Que) se ven sus cicatrices fácilmente. 

          c.  (Que) a Marta_i se le_i ven [las cicatrices]_i fácilmente . 

          d.  (Que) [las cicatrices]_i a Marta_i se le_i ven fácilmente. 

          e.  (Que) a Marta_i [las cicatrices]_i se le_i ven fácilmente. 

          f.   ?(Que) [las cicatrices]_i se le_i ven a Marta_i fácilmente. 

These configurations, therefore, can be used to provide further insight into the position preverbal datives and subjects occupy in Spanish. I will show data suggesting that the dative DP in (1c) occupies the preverbal subject position –presumably Spec,TP–, forcing the theme to remain inside the VP; this is also the case for the preverbal dative in (1d), whose theme appears left-dislocated. However, in (1e) the dative DP is the constituent that is clearly left-dislocated, while the theme appears to sit in the subject position. If that is the case, an analysis whereby the dative possessor DP originates in the specifier of a low applicative projection responsible for establishing the static relation of possession with the theme in its complement position (Low-Appl_AT, following Cuervo 2003) will violate minimality when accounting for the derivation in (1e). In this scenario, an empty category of some sort would have to merge in Spec,ApplP to stand for the possessor, and this null pronominal –being structurally closer to Tº than the theme– would intervene when Tº tries to probe the latter to its specifier. In order to avoid this technical problem, I will propose that these preverbal dative DPs and theme DPs are in fact clitic left dislocations coindexed with resumptive empty pronominals inside the sentence; this proposal aligns with Barbosa’s (2009) analysis for preverbal subjects in Romance consistent null subject languages. 

Additionally, I will address the fact that the inalienable possession construal that commonly arises between a dative DP and a body-part noun becomes optional in the context of verbs denoting a change of state (2a). I will offer an analysis to structurally account for these data that combines the notion of possessor raising (Deal 2017) and middle/affected applicatives (Cuervo 2003, 2010). This argument introducing functional head merges “sandwiched” between two subevents in change-of-state contexts, conferring the dative DP originating in its specifier the interpretation of affected by the theme’s resulting state (2b). Furthermore, the dative DP may be understood as the accidental causer of the change of state if Appl merges on top of a dynamic subevent (v_GO, in Cuervo’s terms). In these two scenarios, the dative DP can also be interpreted as the alienable possessor of the theme. Crucially, I will show that the inalienable possessor reading of the dative in change-of-state contexts containing a body-part noun is only available when the possessor is also affected by the body-part, i.e., when ApplP merges below vP, but never when the dative is also the accidental causer, i.e., when Appl is on top of the outermost vP. I will explain how these contrasts can be accounted for by a hybrid approach to inalienable possession in Spanish. 

(2)     a.  A Marta_i, [las_i/arb/tus_k verrugas] se le_i congelan fácilmente. 

              ‘Marta’s warts freeze easily.’ 

              ‘(Your) warts freeze easily on Marta.’ 

              ‘Marta accidentally causes (your) warts to freeze easily.’ 

          b.  A Juan_i se le_i quemó la hamburguesa. 

              ‘The/his burger burned and Juan is affected by it.’ 

              ‘Juan accidentally caused (the/his) burger to burn.’ 

          c.  A Juan_i se le_i congelaron [los dedos]_i 

              ‘Juan’s fingers froze, and he’s affected by it.’ 

              Less transparent: ‘The fingers were frozen for Juan.’ 

              Impossible: ‘Juan accidentally froze his own fingers.’

Tuesday, 23 May 2023 (12-1pm, UK time)

Chiara Gianollo (University of Bologna) – From “alter”/”alius” to “altro”: ambiguity and context-dependence of “other” in Latin and Italian

Abstract: The cross-linguistic investigation of the expression of ‘other’ is made challenging by a series of factors, among which the variety of lexicalization patterns, the widespread polysemy, and the strong context-dependence of the disambiguation strategies play a particularly relevant role (Beck 2000, Breban 2003, Eguren & Sánchez 2004, Oxford 2010, Charnavel 2015, Cinque 2015, Brugè 2017, 2018 a.o.). In Italian, as in many other languages within and outside Romance, a fundamental ambiguity is observable between the additive (1) and the substitutive (non-identity) reading (2):

(1) Il tracciamento dei contatti è essenziale per sconfiggere la pandemia. Un’altra misura da mettere in atto è, inoltre / # invece, il distanziamento sociale.

‘Contact tracing is essential to overcome the pandemic. A further measure to be applied is, moreover / # instead, social distancing.’

(2) Si ritiene che il virus sia stato trasmesso agli esseri umani dal pipistrello. Un’altra ipotesi riconduce il salto di specie, invece / #inoltre, al pangolino.

‘The virus is believed to have been transmitted to humans by the bat. A different hypothesis traces back the spillover, instead / # moreover, to the pangolin’

(Gianollo & Mauri 2020: 132)

In the disambiguation between the two readings, contextual factors such as discourse relations, definiteness / indefiniteness and grammatical number are crucial. In this talk, after discussing the main dimensions of variation and the contextual factors involved, I will explore to what extent the additive / substitutive ambiguity was present in Latin and how it interacted with the lexicalized difference between the dual determiner alter ‘other (of two)’ and alius ‘other (of more than two)’ and with the absence of articles (Bertocchi et al. 2010, Bortolussi 2015: 173-243). Historically, alter is continued as lexicalization of ‘other’ in Romance (e.g. Fr. autre, Sp. otro, Pt. outro, Rom. alt, Sard. àtteru , Rhaet. auter), which is in principle surprising in view of its more specialized meaning. Bertocchi et al (2010: 160-161) propose to connect the loss of the difference between Latin alter and alius to the development of articles. I will elaborate on this observation and discuss the contribution of the Late Latin in this respect.  

Tuesday, 30 May 2023 (12-1pm, UK time)

Elena Isolani (University of Cambridge) – A parametric analysis of verb movement and non-standard questions in NIDs

Abstract:

Background: The Parametric Comparison Method (PCM) is a tool for language comparison developed over the last two decades (Longobardi, 2003, 2018; Longobardi & Guardiano, 2009; Longobardi et al., 2013, 2015; Guardiano & Longobardi, 2017; Ceolin et al., 2020) which takes sets of syntactic parameters from different languages as input to computations and output phylogenies of those languages. One of the main benefits of the PCM is the possibility of its expansion in terms of the languages under analysis and the parameters included. While Longobardi et. al focused on the nominal parameters and Baker & Roberts (2021) on the clausal parameters, this further expansion involves the complementizer domain (CP). As for the languages tested, whereas the studies on the other functional domains embraced the traditional language family (Romance, Germanic, Slavic, etc.), this research attempts to provide a more language-specific parametric comparison addressing the Romance varieties spoken in the Italian peninsula. This presentation aims to present some preliminary results retrieved from the application of the PCM to the CP on Northern Italian dialects (NIDs).

The parameters: The selection of the parameters was carriedoutby adopting the cartographic framework; for each head of the split CP (Rizzi, 1997), a sub-list of parameters was generated, for a total of 101 parameters. Several parameters were formulated in terms of formal features on functional heads. Following Gianollo et al. (2008), the grammaticalisation, checking, spreading, and strength of functional features were verified to account for numerous phenomena characterizing the CP. Other parameters, instead, account for further salient patterns of variations in CP. Once the value of each parameter was assigned, languages were compared in pairs by means of the syntactic distance: the distance between two languages is δ (0 > δ > 1) determined by the following formula for the ordered pair <i, d> (where i = the number of identities in parameter values and d = the number of differences): d/(i+d).  

The parameterization of Verb Movement: In order to describe verb movement to the C-layer, it is vital to make a distinction between (at least) two possible target positions, namely FinP and ForceP, located at the two extreme edges of the CP (Rizzi, 1997).  

Pc1 and Pc3 ask whether a feature [F] carried by any element can move from a lower position to ForceP or FinP respectively. Pc2 and Pc4 regulate whether [F] can move while endowed by the verb. Finally, Pc5 determines an empirical context where V-FinP occurs, namely cases of Residual V2 that in Romance languages are characterized by the so-called Aux-to-Comp constructions. While this schema shows the specific relations between parameters, it fails to empirically represent V-ForceP. Traditionally, V-ForceP languages have a poor verbal pre-field resulting in a low occurrence of V3 and V4 (Wolfe, 2016). In this study, we find a similar pattern in Fiorentino compound structures where the complementizer ‘che’ (that) is omitted.Fiorentino stands out with respect to other Italo-Romance varieties as it accepts complementizer deletion (CD) in several syntactic contexts with, apparently, the sole requirement of an intervening clitic-nature element (a preverbal clitic, a preverbal negator or an auxiliary) between the matrix and embedded verbs (Cocchi & Poletto, 2002). According to Cocchi & Poletto (2002), this clitic element moves from an embedded position to Force°, traditionally filled by the declarative complementizer. In this presentation I will propose that the clitic forms a unique unit with the verb, meaning that embracing the idea that CD involves clitic movement to Force (Cocchi & Poletto, 2002), verb movement along with the clitic need to be hypothesized as well. 

The Parametrization of Non-Standard Questions (NSQ): For NSQs, we indicate some interrogative structures that are not uttered in out-of-the-blue contexts, but that require a specific semantic/pragmatic context in order to be felicitous. More than two decades of literature on this topic provided a tripartite classification of NSQs (Munaro & Obenauer, 1999; Obenauer & Poletto, 2000; Obenauer, 2012; Hinterholzl & Munaro, 2021): Surprise-Disapproval Interrogative, Can’t-find-the-value-of-X Interrogative and Rhetorical Interrogatives. For each type of NSQ, four parameters were considered regulating the phenomena involved in their realizations.  

• PC6: Are NSQs syntactically realized in the same way as standard questions? 
• PC7: Are NSQs realized through a discourse particle? 
• PC8: Are NSQs realized through wh-movement? 
• PC9: Are NSQs realized through a cleft structure? 

In this presentation, I will illustrate the parametrization of NSQs in several NIDs, primarily focusing on three Veneto dialects: Bellunese, Trevisano and a variety of Polesano. Although these three varieties have similar syntactic properties, they behave differently with respect to the realization of NSQs, indeed they set a positive value of PC6 and PC8 depending on the type of NSQ analysed. The analysis of NSQs is particularly relevant for the outcome of the study as it confirms a comparative pattern among these three varieties that is conclusively established by the calculation of the syntactic distance.

Tuesday, 6 June 2023 (12-1pm, UK time) – POSTPONED (TBC)

Wyn Shaw (University of Oxford) – The History of Subject Verb Inversions in French

Tuesday, 13 June 2023 (12-1pm, UK time)

Susann Fischer (University of Hamburg) – On experiencer doubling and other internal arguments in Spanish and Catalan

Abstract: Psych verbs show a different behaviour in different languages. The differences are (among others) whether the experiencer of these verbs show nominative (1) or oblique case (2-6) and whether they are analyzed as (oblique) subjects (3-5) or objects (6), and whether the oblique subject needs to be doubled by a clitic (3 and 5) or not (1, 2, 4, 6).

(1) I like rock music                     English

(2) Mér líkar rokktónlist                 Icelandic

(3) A Marcos le gusta la música rock      Spanish

(4) A Gianni piace la musica rock          Italian

(5) A en Jordi li agrada la música rock     Catalan

(6) La musique rock plait à Marie         French

In the Old Romance and Old Germanic languages, most experiencer verbs are attested with oblique experiencers that look like structural subjects; even in English (7) and French (8).

(7) Me hungreþ                    OldEng

     meOBL go3SG hungry    ModEng   ‘I am hungry.’

(8) Moi/Me souviens          OldFre

     meOBL remember3SG   ModFre    ‘Je me souviens’ ‘I remember.’

The availability and properties of oblique structural subjects have been seen to depend on factors such as lexical case and case- or verbal-morphology and how case is assigned (Belletti 1988, Chomsky 1995, Sigurðsson 1992, 2004, among others). The loss of oblique subjects has been explained by a change concerning the argument- and event-structure of the verbs, the loss of case- or verbal morphology and how case is assigned (Lightfoot 1979, Seefranz-Montag 1983, Allen 1995, Mathieu 2006, van Gelderen 2014, 2019, Batllori et al. 2019 among others).

In this talk, we will provide data showing that Old Romance (even Old French) allow structural oblique experiencer subjects that pass most subject tests. We will argue that the change concerning the experiencer in Spanish and Catalan cannot be explained with a loss of case- or verbal-morphology, but seems to be connected to a change concerning the left periphery, i.e. a change from a topic prominent to a subject prominent language. Not only can this change be seen in the change concerning the variability of word order and the status of the oblique experiencers, but also in the fact that oblique experiencers in psych verb constructions need to be doubled obligatorily in Modern Catalan and Modern Spanish.When the significance of the subject in Romance became more relevant, subjects needed to be identified. We argue that the obligatory doubling of experiencer (a highly accessible constituent) can be considered a disambiguating mechanism. English and French are strict SVO lgg, the structural subject is easy to identify. In Catalan and Spanish some of the object experiencer verbs are reanalysed as subject experiencer verbs (s. also Batllori et al. 2019). However, some object experiencer verbs are still in use, and since Catalan and Spanish still allow word-order variation, the object experiencer verbs need to be differentiated from subject experiencer verbs, and that is why they are doubled

Tuesday, 20 June 2023 (12-1pm, UK time)

James Corbet (University of Newcastle) – The perfect road and imperfect reassembly: The use of Spanish tense-aspect morphology by heritage speakers in the UK

Abstract: Silva-Corvalán (1994) observes that intergenerational changes in bilingual grammars can obtain through grammar internal or external motivations. Previous research on the Preterite-Imperfect contrast among Spanish heritage speakers in the USA (e.g., Silva-Corvalán, 1994, 2014; Montrul, 2002; Cuza et al., 2013) has generally shown instability in their association of the Imperfect with imperfective meaning, influenced by the dominant English grammar. Putnam and Sánchez (2013) model this cross-linguistic influence as an activation-linked instance of feature reassembly (see Lardiere, 2009) in a steady-state heritage grammar. However, these previous studies in the USA have exclusively focused on speakers of specific Latin American varieties of Spanish, none of which exhibits extension of the Present Perfect into canonically perfective contexts – the so-called “perfect road” (e.g., Schwenter & Torres Cacoullos, 2008).  

In this presentation, I examine the viewpoint aspect system of Spanish heritage speakers in the United Kingdom, many of whose parents come from areas where the varieties of Spanish have progressed along the “perfect road”. In common with previous studies in the USA, I find substantial reassembly of form-meaning associations in specific imperfective contexts for speakers from all backgrounds, linked to activation as claimed by Putnam and Sánchez. Moreover, a subset of heritage speakers whose parents have “perfect road” varieties extend this incipient change into distant perfective contexts. Despite this change not obtaining as a result of cross-linguistic influence, Putnam and Sánchez’s activation-based model of feature reassembly also explains this pattern of behaviour, suggesting a possible extension of their model to earlier stages of acquisition. 

Tuesday, 27 June 2023 (12-1pm, UK time)

Marco Fioratti (University of Cambridge) – Microvariation in Romance inversion: from northern Italian dialects to a pan-Romance perspective

Abstract: With this talk I intend to present some preliminary observations resulting from the first months of my PhD research project, which focusses on non-veridical verb-subject inversion in (mainly) Romance varieties. In (not only) Romance varieties, verb-subject inversion (or verb-subject clitic, henceforth SCI) can trigger a series of non-veridical polarity readings (Poletto 2000; Munaro 2002; Obenauer 2004; Munaro 2010; Ledgeway 2015) – not always classified according to the same criteria in the literature – which, interestingly, seem to belong to the same set in every language (i.e., interrogative, exclamative, optative, etc.). Research thus far has mainly focussed on whether inversion is grammatical with a given reading in a given variety (Munaro 2002, 2010); nevertheless, a first exploration of the literature and some data from northern Romance varieties show that the actual productivity of inversion cannot be determined binarily even within the same interpretive context. Many restrictions are indeed at work in limiting the availability of verb movement to the CP layer (or simply, inversion linearly visible, as in the case of verb-subject clitic inversion (SCI) in NIDs), and two of them have been preliminarily explored during these first months, namely, verb class (lexical vs functional) restrictions, and person-driven restrictions.

For each restriction, a specific generalisation seems to possibly capture the data:

a)    If for a given non-veridical reading a variety allows inversion with lexical predicates, then it also admits inversion with all functional predicates for the same non-veridical reading.

b)    If for a given non-veridical reading a variety allows inversion with only one person of the paradigm, this will be the 3rd person; if it allows inversion with two, these will be the 3rd and the 2nd person. Consequently, if inversion is allowed with the 1st person, then it will also be grammatical with the 2nd and 3rd persons.

The main hypothesis that is being investigated is that the availability of V-to-C movement may correlate with the height of the predicate involved in the inversion process in declarative contexts, and the loss of inversion might be explained by assuming that the requirement for a Probe-Goal relation between C and T is not being satisfied anymore, due to a reinterpretation of the featural content of either C or T. We can then expect that the loss of inversion generally starts with the highest projections of the CP layer – possibly following Munaro’s (2010) hierarchy – and with the lowest predicates of the TP layer. Acceptability judgements by northern and southern Italian speakers, from regional varieties of Italian with different verb movement (Schifano 2018) seem to suggest that this correlation holds.

Some ad interim conclusions can also be drawn with respect to the status of enclitics in some northern Italian varieties. In this talk, I show that in the Polesine varieties analysed, clitics might not be analysed as clause markers, as they still show interactions with the argument structure, although a V-to-C movement analysis can be maintained, and the distribution of the availability of inversion across the varieties analysed seems to follow both Munaro’s (2010) projection hierarchy and the generalisations in a) and b).

Selected references

Ledgeway, A. (2015) ‘Parallels in Romance nominal and clausal microvariation’, Revue Romane de Linguistique, LX, 2–3, 105–127.

Munaro, N. (2010) ‘Toward a Hierarchy of Clause Types’, in P. Benincà and N. Munaro (eds.), Mapping the Left Periphery: The Cartography of Syntactic Structures, 4, Oxford: Oxford University Press, 125-162.

Obenauer, H.-G. (2004) ‘Nonstandard wh-questions and alternative checkers in Pagotto’, in H. Lohnstein, and S. Trissler (eds.), The Syntax and Semantics of the Left Periphery, Berlin and Boston: Mouton De Gruyter, 343-383.

Poletto, C. (2000) The Higher Functional Field. Evidence from Northern Italian Dialects, Oxford: Oxford University Press.

Schifano, N. (2018) Verb Movement in Romance. A Comparative Study, Oxford: Oxford University Press.

You can register for the talk via the following link: https://zoom.us/meeting/register/tJIuf-uqrTMvE9J3wDScAGgFvTzks03qWYfN. **We kindly ask you to not share this link with third parties.**

Lent Term 2023

Tuesday, 21 March 2023 (12-1pm, UK time)

Valentina Cojocaru (University of Bucharest and Romanian Academy, Institue of Linguistics) – Language contact and discourse markers. A case study on the Romanian variety spoken in Moldova

Abstract: The Romanian variety spoken in The Republic of Moldova, also known as Moldovan Daco‑Romanian, has recently aroused many linguists’ interest (Ștefănescu 2016, Costea 2017, 2018, a.o). For more than 200 years, Moldovan Daco-Romanian has been under intense language contact with Russian, being isolated from standard Daco-Romanian. These factors led to two unavoidable phenomena: (1) introduction of Russian elements into various areas of vocabulary, phonology, morphology, and syntax; (2) preservation of numerous features from old Romanian at all levels of the language.

Given sociolinguistic considerations, it is little surprising, therefore, that the syntax or morphology preserved certain features, but what is quite surprising instead is finding old Romanian features at the discourse level, the level of spontaneous speech, subject to change and innovation. Alongside Russian discourse markers, which are quite numerous and very frequent in spontaneous speech (koroche  ‘in other words’/‘long story short’/‘so’/‘anyway’, ladno ‘ok’, polyubomu ‘anyway’, pohodu ‘looks like’/‘it seems that’/‘apparently’, tipa  ‘seemingly’/‘sort of’/‘like’ (just to give very few examples with their approximate translations in English)), there are several “local” discourse markers, some of them being attested in the historical region of Moldavia since the end of nineteenth century: mătincă‘I think (that)’, psinică‘it seems (that)’/‘probably’, de-amu ‘already’/‘from now on’ (Iordan 1950, Dimitrescu 1958, Ivănescu 1980, Arvinte 2002).

In the present contribution we focus on the analysis of linguistic interferences at the level of discourse markers following two major lines of research: pragmatics and sociolinguistics. The pragmatic approach considers the types of discourse markers used, their functions, the context in which they appear, the type of speech acts they relate to, etc.; whereas the sociolinguistic approach explores the background of discourse markers production considering linguistic and extralinguistic factors. The socio-pragmatic analysis of discourse markers is based on a corpus that contains samples of text messages, chat conversations and recorded spontaneous speech (some of them are published in Bochmann 2002; and others (those recorded by myself) are to be published).

Tuesday, 14 March 2023 (12-1pm, UK time)

Norma Schifano (University of Birmingham) – Innovation and preservation in UK Ecuadorian Spanish: a preliminary morphosyntactic investigation

Abstract: Overlooked by official statistical data and largely absent from academic investigations, the Latin Americans living in the UK represent one of the least visible migrant communities of the country, despite being one of the fastest growing groups of the capital. While progress has been made in sociology, surprisingly little is known about the grammatical properties of their languages. Focusing on the Spanish-speaking population, the project ‘Heritage Speakers of Spanish: the case of Latin American Londoners’ aims to fill this gap by producing the first grammatical documentation and analysis of their varieties of Spanish. In this presentation, I will concentrate my attention on the preliminary results based on a first set of investigations centred on the Ecuadorian variety, with specific reference to the expression of evidential and (ad)mirative values via synthetic (canté ‘I sang’) vs analytic perfect (he cantado ‘I have sung’) verb forms. From a contact perspective, it has already been shown that these innovative values which characterise the he cantado form of Ecuadorian Spanish can be lost under the pressure exerted by other varieties of Spanish, as in the case of Ecuadorian speakers living in Madrid for a sustained period of time (Palacios 2007). The question therefore arises as to whether the same is happening in the UK and whether the attested changes (if any) can be ascribed to contact with English and/or with other Spanish varieties and/or are due to a process of spontaneous change. Preliminary results from 14 speakers (including one control group from Ecuador and one from Spain) will be presented, gathered via an aural preference task and a production task. It will be shown that patterns diverging from the Ecuadorian norm can be found not only among Ecuadorian speakers who have lived in Spain for a sustained period of time before reaching the UK, as expected, but also among speakers with a short-medium period of UK residency who have moved directly from Ecuador. Notably, my preliminary sample also includes speaker with a medium-long UK residency and middle-sustained English contact who have remained remarkably close to the Ecuadorian norm. The interplay between sociolinguistic and linguistic factors will be discussed, as well as some of the consequences for existing theories on language contact.

Tuesday, 7 March 2023 (12-1pm, UK time)

Guglielmo Cinque (“Ca’ Foscari” University of Venice) – A movement approach to DP-internal ellipsis

Abstract: In this talk I consider a particular generalization concerning ellipsis within the extended nominal projection: ellipsis can target a nominal modifier only if all constituents below it are also elided. Building on an analysis of ellipsis grounded on movement to left edges I suggest that this generalization follows from a condition on DP-internal movement proposed in Cinque 2005 to derive Greenberg’s Universal 20.

Tuesday, 28 February 2023 (12-1pm, UK time)

Maria Cristina Cuervo (University of Toronto) – The Problem of Tritransitives

Abstract: This research describes and analyzes a particular construction involving inalienable possession of a body part by an external dative possessor. The relevant sentences comprise four participants, three of which appear as internal arguments: a theme DP, a locative PP and a dative possessor DP, as illustrated in (1).

(1)   a.    Jean lui a mis un chapeau sur la tête. (French)

                ‘Jean put a hat on his head’

b.    María le puso el bebé en los brazos a Emilio. (Spanish)

c.     Maria i-a pus copilul in brate lui Emil. (Romanian)

       ‘Maria placed the baby in Emilio’s arms’

These tritransitives share properties with related constructions discussed in the literature on external possession but their specific properties have not been investigated. Tritransitives present several puzzles that are not straightforwardly explained, either by studies of verb classes, current theories of argument licensing or by typological approaches to external possession. In particular, tritransitives raise questions about argument licensing, syntactic relations between arguments and their interpretation, and the notions of valence and transitivity.

As the first step of a crosslinguistic investigation, the main focus is on Spanish, in which, similar to Romanian but in contrast with French, dative possessor constructions are less restricted along several dimensions, such as the types of nouns and possession relation involved, the referentiality of the possessum, the syntactic expression of the dative as a full clitic-doubled DP, and the types of verbs that are compatible with this construal. A descriptive overview of Spanish TTRs is presented.

The problem with tritransitives goes beyond the classic challenge of accounting for the apparent thematic licensing of the dative DP as the possessor of the locative and its status as a clausal, affected argument. Unlike other dative possessors in Romance, these datives cannot be analyzed as taking the possessum as its complement, in a double-object or Figure-Ground structure, as this seems to be the structural relation between the direct object and the locative PP to the exclusion of the dative DP. Further, corpus data show that TTRs are not restricted to inalienable possession, as evidenced in (2), which makes a raising analysis less attractive.

(2)   Emilio le instaló una canilla nueva en la cocina a Gabi.   [buscar en corpus]

        ‘Emilio installed a new faucet in the kitchen for Gabi/in Gabi’s kitchen’

These constructions are better explained by an applicative analysis whereby the dative DP structurally relates to the small-clause type relation between the theme and the locative. As such, this configuration—typically expressing a change of location event—is to be distinguished from the better studied Romance external possession cases in which the possessor relates directly to the other internal argument (French Jean lui a lavé les chevaux; Spanish Emilio la besó en la frente, Romanian Maria se spălat pe mâini). Instead, the dative DP in TTRs shares properties with datives in change of state constructions (both causative and inchoative) in which it is interpreted as the “possessor” of the the new state of the object (Emilio le rompió la radio a Carolina ‘Emilio broke Carolina’s radio on her’—> Carolina has a broken radio). This relation underlies the affectedness interpretation arising configurationally from arguments related to a state (or location) embedded under a dynamic event. This suggests, in line with much recent work on argument structure, that the underlying structure of TTRs is not exclusive of external possession, and, conversely, that external possession sentences can rely on distinct underlying configurations.

Some broader implications are considered for theories of valence, argument licensing and interpretation, and typological approaches to applicatives and external possession.

Tuesday, 21 February 2023 (12-1pm, UK time)

Paolo Acquaviva (University College Dublin) – Varieties of definiteness and constructionist interpretation

Abstract: Grammatical categories are not the same thing as logical ones, but logic is the necessary basis to model the meaning of morphemes expressing Boolean operators, (in-)definiteness, or universal/existential quantification. Yet these are distinctly linguistic elements, whose contents vary across languages and also across constructions. I would like to consider in some detail the variety of contents expressed by definite marking, focusing on the diverse but mutually comparable phenomena offered by the Romance family. Names, mass terms, and kind terms are particularly important as they highlight how definite articles can contribute to the interpretation of a complex structure without either being the source of definiteness (uniqueness, familiarity) nor being ‘expletive’.

Tuesday, 14 February 2023 (12-1pm, UK time)

Ares Llop Naya (Fitzwilliam College, University of Cambridge), Anna Paradís (University of Oxford), Eloi Puig-Mayenco (King’s College London) – An exploration of the development of Negative concord items in Ln Catalan: the case of L1 English, L1 Italian, L1 Portuguese and L1 Spanish speakers

Abstract: In this talk we are going to present a comprehensive study on the representation of negation in the multilingual grammars of non-native speakers of Romance languages throughout development. Our approach provides a fresh insight on how the negative system, specifically NCIs in non-veridical contexts, is represented in the grammar of learners of Catalan at different stages of development and how this non-native grammatical system interacts with both other already-acquired systems of negation and the implicit and explicit linguistic evidence about the way the negative system works in Catalan presented in Catalan Ln textbooks. Our study contributes to the further understanding of which intra and extralinguistic factors model L3/Ln developmental trajectories (e.g., Cabrelli Amaro et al., 2020; Cabrelli and Iverson, submitted; Puig-Mayenco, et al. 2022).

To do so, we examine Catalan Negative Concord Items (ningú ‘nobody’, res ‘nothing’, cap+DP ‘no+DP’ and mai ‘never’) in non-veridical contexts (questions and conditions) by 82 participants (L1 Portuguese, L1 Italian, L1 Spanish and L1 English speakers). The interpretation of Catalan NCIs patterns alike in these contexts with English NPIs and Italian NCIs and differently from Spanish and Portuguese NCIs. The choice of these L1s, contexts and items allows us to manipulate and explore the interaction between languages spoken, classroom evidence and availability in input and more importantly, how this might predict L3/Ln developmental patterns.

We devised a truth-value judgment task in Catalan that allowed us to tap into the interpretation of NCIs by the targeted learners. In the tasks participants read a context and saw a sentence with a target NCI and were asked to say whether the sentence was true or false. There were three main conditions (fillers, non-veridical condition and anti-veridical condition) with 32 items in each, each condition manipulated lexical item (ningú, res, cap, mai), directionality of truth-value (true vs false) and within the non-veridical condition, we further manipulated the operator (conditional and question). In total, there were 96 experimental items. In addition, participants responded a background questionnaire eliciting information regarding the language background, and exposure to Catalan and took a short proficiency test in Catalan.

The results of the study indicate that previously acquired knowledge plays an important role early on. Italian provides a facilitation effect from the very beginning, whereas Portuguese and Spanish provide non-facilitation. English does not provide facilitation where it could have done so. Regarding development, our results show that overcoming non-facilitation is not straightforward. L1 Portuguese and L1 Spanish speakers do not overcome such non-facilitation and L1 English speakers do so quite quickly. L1 Italian speakers do slightly worse as proficiency increases, which we argue it is potentially an effect from the actual teaching itself.

We discuss the results in in relation to current theories of morphosyntactic transfer (e.g., Fallah, et al., 2016; Rothman, 2015; Rothman et al., 2019; Slabakova, 2017; Westergaard, 2020) and recent proposals that explicate developmental patterns in L3 acquisition (Cabrelli and Iverson, Submitted) and we build links with language pedagogy research and how NCIs are taught in the Catalan Ln classroom. We will end the presentation by presenting an SLA-and-language-pedagogy-informed teaching sequence of the negative system of Catalan as a Foreign Language (CFL) for multilingual learners based on the results of the empirical study. The proposed sequence considers the evidence from our study and follows a spiral curriculum-based approach (Bruner, 1960) to the teaching of non-veridical contexts in Catalan, from ab initio (A1) to proficient (C2) levels. The learner is exposed to explicit input (with focus-on-form activities, Long 1991) and to implicit input throughout the learning path, and to multiple and various cues that allow them to subsequently access the properties of all NCIs (see encoding variability hypothesis, Edmonds et al. 2021). The proposal provides strategies to enhance the input and to increase the availability, saliency and the accessibility of the more difficult and opaque forms.

Tuesday, 7 February 2023 (12-1pm, UK time)

Julio Villa-García (University of Oviedo) – On the extrasententiality of left dislocations

Abstract: In this talk I argue for a bisentential, paratactic account of Hanging Topic Left Dislocations (HTLDs) in Romance and English, wherein the structurally unconnected hanging topic phrase (Cambridge)  is the remnant of an elliptical copulative sentence linearly juxtaposed to the second, host sentence (I’ve been there) à la Ott (2014, 2015, 2016, 2017). This proposal represents a natural extension of Ott’s system for Clitic Left Dislocations (ClLDs) and predicative non-restrictive nominal appositives. At the same time, the analysis constitutes a radical departure from integrated, monosentential approaches which analyze HTs as intrasentential, albeit left-peripheral, constituents in the left spine of the clause. 

The paratactic approach provides a principled account of various issues raised by monosentential analyses of HTs within cartography (e.g., anticonnectivity, agreement mismatches, coreference with the resumptive/epithetic correlate, insensitivity to locality constraints, islandhood, Case, potential presence of interjections between HT and host sentence, “comma intonation”/pause potential, different illocutionary force, amongst others). The account is also successful in capturing orphaned topics, which are not linked to any constituent in the sentence they occur with, alongside what I term hyperdetached HTs.

Tuesday, 31 January 2023 (12-1pm, UK time)

James Hawkey (University of Bristol) – Variation in Spanish past perfectives: Aoristic drift in a situation of complex language contact

Abstract: Romance varieties differ in their usage of preterit and present perfect verb tenses. Both are past perfectives, but whereas Portuguese uses the preterit in most contexts, spoken French prefers the present perfect. Peninsular Spanish lies between the two, though evidence indicates that the present perfect is becoming the default past perfective (Schwenter and Torres Cacoullos 2008) in a process of ‘aoristic drift’ (Squartini and Bertinetto 2000). How does speaker multilingualism affect this? Semi-structured interviews were conducted with second-generation members of the Portuguese diaspora in Andorra. We might expect native competence in Portuguese to inhibit aoristic drift in Spanish, since contact has been shown to affect past perfective verb tense in other Romance varieties (Gili Gaya 1993, Hawkey 2020). Contrary to expectations, participants demonstrated aoristic drift. Dense and multiplex migrant networks are, however, shown to favour the maintenance of vernacular norms (Milroy 1980), including generalising the function of the present perfect.

Tuesday, 24 January 2023 (12-1pm, UK time)

Gabriel Martínez Vera (University of Newcastle) – ‘Dizque’ and ‘como que’: a comparative analysis

Abstract: Evidentiality encodes information about the speaker’s relationship to the source of a statement (v. Aikhenvald 2004). Although evidential marking occupies a fixed position in the clause in many languages, two evidential markers, dizque and como que, in Colombian Spanish (from Medellin) can appear in different positions (v. (1)-(2); these are based on Grajales 2017). We provide a unified account of the clausal- and constituent-scope of these markers in an approach that likens them to focus sensitive elements, such as even.

(1)    a.  Dizque va a llover esta noche.                                   
‘Allegedly, it is going to rain tonight.’

b.  Como que va a llover esta noche.
   ‘It looks like it is going to rain tonight.’                                                                  

(2)    a.  Juan trajo dizque un compás marino, que resultó ser un visor de fotos.
  ‘Juan brought an alleged sea compass, that ended up being a photo viewer.’

b.  Juan trajo como que un compás marino, que resultó ser un visor de fotos.
 ‘Juan brought a kind of a sea compass, that ended up being a photo viewer.’

References

Aikhenvald, A. (2004). Evidentiality. Oxford: Oxford University Press.
Grajales, R. 2017. La estrategia evidencial dizque en el español de Medellín, Colombia. Onomázein 37. 244–278.

**Thursday**, 19 January 2023 (**12.30-1.30pm**, UK time)

Afra Pujol i Campeny (University of Oxford) – XVS in old Catalan

Abstract: This talk explores the distribution, syntax, and information structure of XVS clauses in the narrative text and the reported speech of a 13th century Old Catalan chronicle, the Llibre dels Fets. It is shown that XVS occurs mainly within reported speech and in embedded clauses, responding to the archaising nature of these syntactic domains (the former reproducing syntactic structures echoing epic literature, the latter being inherently archaising syntactically), while XVS less frequent in narrative text, more innovative and closer to spoken language. The data presented demonstrates that by the 13th century, XVS constructions were mainly used to express verum focus within the scope of nonveridical operators and not to a structural V2 requirement. This conclusion is supported by 11th data from legal sources and data from a 14th century chronicle.

Tuesday, 10 January 2023 (**12-1pm**, UK time)

Fernanda Pratas (University of Lisbon) – Time notions in human language: empirical clues from Caboverdean

Abstract: This talk will focus on the expression of some temporal meanings in Caboverdean, a Portuguese-related language, under the assumption that a comprehensive analysis of these phenomena crosslinguistically must consider their relation to a few fundamental time notions, such as: (a) certain temporal locations are perceived/presented as somehow detached from the speaker (cf. Binnick 2010; Bybee et al 1994); in Pratas (2021) it is proposed that a low accessibility is indeed the value conveyed by the suffix -ba, which appears with verbs in some past meanings and also with verbs – in contexts which might otherwise be analysed as modal – that correspond to infinitives in other languages; and (b) the nature of time may be perceived/presented by the speaker as involving some kind of motion, which is expressed by the use of the verbs ben ‘come’ and bá ‘go’ with a specific temporal import (Pratas, forthcoming). 

Michaelmas Term 2022

Tuesday, 11 October 2022 (1 – 2.15pm, UK time)

Ian Roberts (University of Cambridge) & Dalina Kallulli (University of Vienna) – On Voice

Abstract: In this talk, we focus on the morphosyntactic and semantic properties of a range of constructions involving non-active voice and/or reflexive morphology across Romance and beyond. We show that if we treat reflexives as involving Voice, they fit into a consistent, independently attested pattern of cross-linguistic variation in properties of functional heads, allowing us to derive the syntactically different constructions (reflexive, anticausative, middle, passive) exhibiting such morphology in a unified manner. Specifically, we will argue for the general cross-linguistic schema in (1), whereby (i) Voice licenses the Internal Argument in all cases but in differing ways; (ii) in the case of reflexives, there is a Voice-feature (refl) which marks the predicate as reflexive (along the general lines of Reinhart & Reuland 1993), with the dimensions of variation involving (a) the exponence of Voice (clitic, affix, auxiliary, zero …); and (b) the nature of the IA (pro, weak pronoun e.g. sich, special/body-part pronoun e.g. X-self, …), which in turn are part of a more general pattern of variation.

            (1)          [VoiceP Voice  … [VP   V   IA ]]

We will conclude with some general reflections on the nature of the external argument of passives.

Tuesday, 18 October 2022 (1 – 2.15pm, UK time)

Prudence de Pontbriand (University of Göttingen) – Null Objects in Old French: Scoping out syntactic and semantic restrictions

Abstract: The aim of the talk is to provide a preliminary profile of null objects in Old French (with a particular emphasis on prose texts from the 13th century). The talk will focus on potential restrictions, either semantic or syntactic, that can be found with regards to null objects. In this research, null objects are taken to be phonologically null elements base-generated in the complement position of the verb (Roberge and Cummins, 2005). They correspond most frequently to a salient pronominal element with an antecedent mentioned in the preceding context.

(1) Et quant il voit venir Galaad si (-) vet a l’ encontre

and when he sees come.INF Galahad thus goes to the=meeting

‘And when he sees Galahad coming, he goes to meet (-)’ (Queste del Saint Graal, II, 43)

The present talk will show that, while Old French does not display absolute restrictions with regards to null objects, we can observe strong tendencies in their semantic and syntactic properties. This observation is especially true concerning their semantic properties, where animacy and specificity appear to play a very important role. With regards to their syntactic properties, null objects allow more variation but they still show a preference for a specific bundle of properties (namely 3rd person singular direct objects).

Tuesday, 25 October 2022 (1 – 2.15pm, UK time)

Annamaria Chilà (University of Messina) – La classe flessiva dei plurali in -a in calabrese meridionale e siciliano

Abstract: Le varietà calabresi meridionali e quelle siciliane presentano, accanto ai plurali in -i e, più raramente, in -ora, numerose forme di plurale in -a. Il talk si propone di offrire un confronto tra i dialetti calabresi e quelli siciliani, considerando in particolare i due seguenti aspetti:

1) Sul versante morfologico, i diversi meccanismi di espansione dei plurali in -a nel lessico. Nel solo siciliano, questi plurali scavalcano la barriera dell’inanimatezza, catturando i derivati [+ animati] e [+ umani] suffissati in -turi (piscatura ‘i pescatori’), -uni (patruna ‘i padroni’) ed -aru (putiara ‘i bottegai’), e sembrerebbero dunque più produttivi e vitali che non in calabrese. La produttività, tuttavia, non si misura solo in termini di type frequency sulla base di un’espansione interna al lessico, ma anche dalla capacità di innescare processi morfologici. Da questo punto di vista, il mantenimento della flessione in -a nelle forme alterate del calabrese (brazzicedda ‘braccini’, jiritedda ‘ditini’) diagnostica la non trascurabile vitalità di tale classe anche in queste varietà.

2) Sul versante sintattico, la presenza vs. assenza di un accordo dedicato in -a. Il calabrese meridionale mostra(va), seppur con tratti di forte recessività, una desinenza dedicata -a anche nei target dell’accordo. Ce ne sono esempi, per il calabrese di età medievale, nel ricettario medico di Luca Geracitano da Stilo (un poco de linni arsa ‘un po’ di legni bruciati’; falli bevere la celebrella de lo gallo ‘fagli bere le cervella del gallo’); per quello moderno, nelle cc. AIS 919-920 (i ligna cotta ndumanu boni ‘la legna secca brucia bene’). Questo accordo non è documentato in siciliano, dove, come conferma il corpus Artesia già per la fase medievale, tanto i nomi in -a quanto quelli in -ora presentano accordo plurale in -i.

Tuesday, 1 November 2022 (1 – 2.15pm, UK time)

Sonia Cyrino (State University of Campinas ‘UNICAMP’) – More on negation in Brazilian Portuguese

Abstract: It has been shown in several works (Schwenter 2005, Cavalcante 2007, 2012, Biberauer & Cyrino 2009, Cyrino & Biberauer 2009, Teixeira de Sousa 2015) that Brazilian Portuguese (BP) has three positions for the negative marker não in sentential negation (não-V, não-VP-não, and V-não), and they are related to different pragmatic/discourse functions. However, a fourth and reportedly recent position for the negative marker has been noted in the literature about BP (Cyrino 2010, 2013, De Paula 2014): não is possible between an auxiliary and an uninflected verb, that is, in the slot in the structure AUX ___ V_[-finite]. In this talk, I advance an analysis for this new position of não in BP. Although syncretic with sentential negation, I show that this negative marker is an instance of constituent negation. Following the nanosyntax approach in DeClercq (2013), I entertain the hypothesis that this fourth não is a Foc^neg item in contemporary BP, as the result of a diachronic reanalysis. I propose that Foc^não has its origin in prefixal não-, with which it is syncretic. It has been detected (Campos 2002, 2009; Pereira 2012) that prefixal não- was rarely used in the 16^th century, becoming more frequent in the 20^th century. Since contemporary BP has been shown (Cyrino 2010, 2013) to allow certain elements such as clitics and emphatic subjects to occur in the AUX ___ V_[-finite] structure, the diachronic reanalysis of the prefixal não- into Foc^neg might have been facilitated.

Tuesday, 8 November 2022 (1 – 2.15pm, UK time)

Eva-Maria Remberger (University of Vienna) & Natascha Pomino (University of Wuppertal) – Romance theme vowels: Not just ornamental morphemes, but not syntactic elements either

Abstract: Theme vowels (ThV) are considered by many linguists (e.g. Oltra-Massuet 1999) to be mere ornamental elements without any effect on syntax and semantics (see however Fábregas 2017). We agree with the idea that ThVs are not to be confused with the realizations of the verbalizer v°, but we would like to show that they are not as ornamental as one may think. Work on the systematicity of stem allomorphy has shown that irregularity of inflected elements is related to a reduced number of affixes of these forms (cf. Vanden Wyngaerd 2018). In Romance, many irregular verbs lack a ThV, i.e. they are athematic. Thus, it seems as if ThVs have an impact on the (ir)regularity of verbal forms (Calabrese 2015). One central question is, however, how athematicity is implemented in the theoretical framework, in this case Distributed Morphology (DM, see Halle & Marantz 1993): Is athematicity explainable by zero exponence, by fusion or by cumulative exponence? The aim of this talk is to investigate more in detail the connection between irregularity and athematicity and to argue for an approach based on cumulative exponence.

Most Romance languages have reduced the Latin conjugation classes (CCs) maintaining only three main classes which are marked as general rule by different theme vowels (ThVs). The development in French, however, is different and has led to less transparent verbal forms and to a CC system that is not describable, at first glance, in terms of ThVs. In our talk, we will show that the link between athematicity and irregularity is attested also in other verbal forms in Romance and, more important, it holds also for French, a language for which the assumption of ThV is contradictorily discussed in the literature. We will discuss the CC system of French and propose that it has two thematic and several athematic CCs. Certain types of root allomorphy, e.g. the (non)appearance of root final consonants, depend, in essence, on whether or not there is a ThV-position available (cf. also Schane 1966). As will be shown, the athematic CCs are exactly the ones with the (most) irregular verbs, in French as in other verbal forms in Romance.

Following the DM-based Vocabulary Insertion-Only model proposed by Haugen & Siddiqi (2016), we will propose, for selected Romance verbal forms, an analysis for the mentioned link between athematicity and irregularity based on Spanning. Since the exact conditions and rules for Spanning are still to be investigated, we contribute to the discussion of this issue by the analyses of verbal forms in Romance. The core idea of the analysis proposed is that the spanning size of the vocabulary items (VIs) realizing the roots depends on the respective CC-features and is motivated by the (a)thematicity of the roots. What is more, our analysis is predominantly based on the general process of Vocabulary Insertion to correctly derive the Romance forms. There is no need for Fusion, Pruning nor Impoverishment; root allomorphy is instead explained in essence via Vocabulary Insertion only.

References

Calabrese, Andrea. 2015. Irregular Morphology and Athematic verbs in Italo-Romance. Isogloss, Special Issue on Italo-Romance morphosyntax: 69–102.

Oltra Massuet, I. 1999. On the constituent structure of Catalan verbs. MIT Working Papers in Linguistics 33. 279–322.

Fábregas, A. 2017. Theme vowels are verbs. In Caha / DeClercq & Vanden Wyngaerd (eds.), The unpublished manuscript. A collection of Lingbuzz papers to celebrate Michal Starke’s 50th birthday. 51–62.

Halle, Morris, and Alec Marantz. 1993. Distributed Morphology and the pieces of inflection. En Kenneth Hale, and Samuel Keyser (eds.), The view of building 20: Essays in honor of Sylvain Bromberger, 111–176. Cambridge, Mass.: MIT Press.

Haugen, Jason D. & Daniel Siddiqi. 2016. “Towards a Restricted Realization Theory. Multimorphemic monolistemicity, portmanteaux, and post-linearization spanning.” In Morphological Metatheory. Edited by Daniel Siddiqi and Heidi Harley. Amsterdam and Philadelphia: Benjamins, 343–86.

Vanden Wyngaerd, Guido. 2018. Suppletion and affix reduction. Accessed October 29, 2021. https://lingbuzz.net/lingbuzz/004095.

Schane, Sanford A. 1966. The morphophonemics of the French verb. Language 42/4: 746–758.

Tuesday, 15 November 2022 (1 – 2.15pm, UK time)

Maj-Britt Mosegaard Hansen (University of Manchester) – Linear and non-linear pragmaticalization (and why it’s not just grammaticalization all the way down)

Abstract: In this paper I address the diachronic evolution of pragmatic markers, with a focus on what I call linear vs non-linear forms of pragmaticalization. The rise of pragmatic markers has been an increasingly popular research topic for more than three decades. The most frequently attested pathway involves linguistic items or constructions that originally have truth-conditional meaning and belong to « core » grammar, but which more or less gradually evolve non-truth-conditional, more (inter)subjective, uses that lie outside « core » grammar. This form of evolution is widely assumed to be regular, unidirectional, and thus fundamentally linear in nature. Saliently, the study of such cases been used to argue for a redefinition of the notion of grammaticalization.

The literature has, however, reported examples of markers that appear to have taken more complex, non-linear, paths at the semantic-pragmatic and/or the syntactic level. These types of cases have been/are the focus of two international research networks of which I was/am the PI : the British Academy-funded « Cyclicity in Semantic-Pragmatic Change : from Latin to Romance » (2017-18) and the AHRC-funded « The Role of Pragmatics in Cyclic Language Change » (2021-23). In my talk, I argue that because pragmatic markers may evolve along a variety of non-linear – including but not limited to cyclical – paths, it is not helpful to subsume the rise of pragmatic markers under the concept of grammaticalization. Instead, in order to arrive at a descriptively adequate account, it is more useful to draw on a distinction between grammaticalization, pragmaticalization and lexicalization. I identify four non-linear forms of pragmaticalization, based on attested patterns of evolution prominently involving interaction between, on the one hand, two levels of meaning (the Content Level and the Context Level), and on the other hand, two levels of grammar (Micro-Syntax and Macro-Syntax). My examples will be adduced from several Romance languages.

Tuesday, 22 November 2022 (1 – 2.15pm, UK time)

María Arche (University of Greenwich) – Modal and temporal issues of evaluative adjectival copular clauses in Spanish

Abstract: In this talk I will present work carried out with Tim Stowell on temporal and modal issues of copular clauses headed by dispositional evaluative adjectives. As discussed in Stowell 1991, Arche 2006, Martin 2011, Fabregas et al 2013, Arche & Stowell 2019, Arche et al 2021, a.o., these adjectives present straddling properties between dispositions and events. Illustrative examples are John was smart to sell his shares last year, John is cruel to scold little children. Drawing from data from Spanish, where tense and aspect are overt in the inflection of the copula, I will discuss the following facts: the adjective expresses the evaluation of a judge at the time of the attitude holder while the tense and aspect overt in the copula seem interpreted in the infinitival clause; when actuality entailments are associated to the event expressed by the infinitive, these depend on the aspectual value (perfective, imperfective) expressed in the copula; the copula choice seems mostly restricted to ser in Spanish, which we argue can be accounted for by the predication as individual-level between the infinitive and the adjective.

Tuesday, 6 December 2022 (***2 – 3.15pm***, UK time)

Judy Bernstein (William Paterson University) – On the morpho-syntax of Romance vocative expressions

Abstract: In this talk I resurrect an old idea, namely that vocative expressions involve a structure ‘smaller’ than true arguments and do not project up to a DP-external functional projection such as VocP (see Stavrou 2013, Hill 2014, Espinal 2013, Corr 2022, Slocum 2016, among others). I hope to show that at least some of the expressions taken to be vocatives do not actually meet the relevant criteria, mainly because they do not encode a required 2nd person feature (Szabolcsi 1994, Bernstein 2008). Several Romance languages (e.g., French, Romanian, Sardinian) display definite articles in vocative expressions, a fact that is problematic under some analyses of vocatives. These definite article facts are accommodated in the articulated DP structure developed in recent work by Bernstein, Ordóñez, and Roca (2018, 2019, 2021).

Tuesday, 13 December 2022 (1 – 2.15pm, UK time)

Oana Uță Bărbulescu (University of Oxford) – Analytic marking of obliques in Romanian: the case of lu(i) and le

Abstract: What is traditionally labelled in Daco-Romance varieties as analytic marking of obliques by lui, lu, li, ii,  le, etc. is – if the grammars and monographs are anything to go by – a kind of Cinderella, a minor, marginal aspect of the morphosyntax of these varieties. The grammars and monographs of the Daco-Romance varieties offer an over-simplified picture and, most significantly, they ignore the diachronic aspect of analytic marking of obliques. In this talk, I aim to present the diachronic development of the analytic marking of obliques by lui, lu, li etc., taking into account the classes of nouns in which it first appeared, its syntactic configurations, and its extension over time, including its sensitivity to register. I also (re)assess the relationship between the feature ‘animate’ and this analytic marking. On the basis of new data, I discuss the status of lui / lu, le etc. (are they (pre)proprial articles, i.e. articles with a distribution limited to proper nouns? oblique case markers?) and their grammaticalization. 

(October) 2021 – (June) 2022

Michaelmas Term 2021

Tuesday, 5 October 2021

Jacqueline Almeida Toribio (University of Texas) – On the uniqueness of Dominican Spanish

______

Tuesday, 12 October 2021

Jelena Zivojinovic (University of Verona) – The use of gerund in Rhaeto-Romance: a diachronic perspective

______

Monday, 18 October 2021

Francisco Ordoñez (Stony Brook University) – Inalienable Location in Catalan and Spanish (joint work with Olga Fernández Soriano)

______

Tuesday, 26 October 2021

Christine Meklenborg-Nilsen (University of Oslo) – Topicalisation strategies in Old French

______

Tuesday, 2 November 2021

Stefano Freyr Castiglione (University College London) – Italian Right Dislocation is Biclausal: An Argument from Binding

______

Tuesday, 9 November 2021

Giuliana Giusti (University Ca’ Foscari of Venice) – Protocol Linguistics: from theory to empirical advancements back to theory again

______

Tuesday, 16 November 2021

Federica Breimaier (University of Zurich) – When the field moves online: a three-step approach to collect and analyze crowdsourced dialect data

______

Tuesday, 23 November 2021

Valentina Colasanti (Trinity College Dublin) – Spatial and discourse deixis and the interactional structure of nominals

______

Tuesday, 30 November 2021

Adina Dragomirescu (‘Iorgu Iordan – Alexandru Rosetti’ Institute of Linguistics/University of Bucharest) – Word order in Istro-Romanian. New data

______

Tuesday, 7 December 2021

Martin Maiden (University of Oxford) – Putting standard Romance languages in their place: two examples of misleading “standard language bias” in the historical morphology of Romanian

Lent Term 2022

Tuesday, 18 January 2022

Diego Pescarini (Université Côte d’Azur and Centre National de la Recherche Scientifique) – Microvariation and statistical hypothesis testing: negation and negative concord in northern Italo-Romance

______

Tuesday, 25 January 2022

Virginia Hill (University of New Brunswick) – AUX cliticisation and loss of SVO in Romanian

______

Tuesday, 8 February 2022

Javier Ormazabal Zamakona (University of the Basque Country) – All that glitters is not syntaX: agreement and the notion of default value (joint work with Juan Romero)

______

Tuesday, 15 February 2022

Michele Loporcaro (University of Zurich) – Nouns as agreeement targets in Romance

______

Tuesday, 22 February 2022

Ian MacKenzie (University of Newcastle) – True versus accidental V2 in Old Spanish

______

Tuesday, 1 March 2022

Richard Kayne (New York University) – Hypercomplex Inversion and the Status of Expletive Pronouns

______

Tuesday, 8 March 2022

Ana Lúcia Santos (University of Lisbon) – Control and inflected infinitives

______

Tuesday, 15 March 2022

Heloisa Salles (University of Brasilia) & Maria Aparecida Torres Morais (University of São Paulo) – The pronominal realization of indirect objects in ditransitive constructions: the case of Dialectal and Standard Brazilian Portuguese

______

Tuesday, 22 March 2022

John Charles Smith (Oxford University) – Defining Romance periphrases: paradigms and grammaticalization

Easter Term 2022

Tuesday, 26 April 2022

José Ignacio Hualde (Illinois University) and Céline Mounole (Université de Pau et des Pays de l’ Adour) – Romance contact in the restructuring of the Basque verbal system: Convergent and surprisingly divergent phenomena

______

Tuesday, 3 May 2022

Paulo Ângelo Araújo Adriano (Universidade Estadual de Campinas/University of Cambridge) – From synthetic to analytic: The present tense in Brazilian Portuguese

______

Tuesday, 10 May 2022

Carolina Fraga (City University of New York) – Completive todo in Rioplatense Spanish

______

Tuesday, 17 May 2022

Giusy Truncellito (Università per Stranieri di Siena) – Reduplication in northeastern Calabrese

______

Tuesday, 24 May 2022

Carlos Yebra López (New York University) – Revitalising Ladino by digital means: Towards a critical approach

______

Tuesday, 31 May 2022

Carmelina Toscano (Università degli Studi di Firenze) – Su alcuni parametri di variazione: la pseudo-coordinazione in Calabria

______

Tuesday, 7 June 2022

Janice Carruthers (Queen’s University Belfast) and Marianne Vergez-Couret (Université de Poitiers) – Temporality in Occitan and French oral narrative

______

Tuesday, 14 June 2022

Lori Repetti (Stonybrook University) – Semantically vacuous insertions in Romance and beyond

______

Tuesday, 21 June 2022

Chris Pountain (Queen Mary University of London) – Learnèd morphology? The grafting of cultured borrowing onto the morphology of Spanish

______

Tuesday, 28 June 2022

Sara Cardullo (University of Cambridge) – Pseudocoordination in Eolian

***

(October) 2020 – (June) 2021

Michaelmas Term 2020

Friday, 9 October 2020

Ionuț Geană (Arizona State University & ‘Iorgu Iordan – Alexandru Rosetti’ Institute of Linguistics) – The Morphosyntax of Istro-Romanian DPs

______

Thursday, 15 October 2020

Larisa Nicolaie (University of Bucharest) – An acoustic analysis of the /ij/ diphthong in Standard Daco-Romanian

______

Thursday, 22 October 2020

Francisco Javier Calvo del Olmo (Universidade Federal do Paraná) – Intercomprehension between Romance Languages: a key for plurilingual education

______

Thursday, 29 October 2020

Giuseppina Silvestri (UCLA) – Lost voices: sociolinguistic observations on the epistemic future in two varieties of the same dialect

______

Thursday, 5 November 2020

Nicola Swinburne (University of Oxford) – The grammaticalization of ‘do’-support in the northern Italian Camuno dialect

______

Thursday, 12 November 2020

Alice Corr (University of Birmingham) – Romance vocatives and the topological mapping of deviance

______

Thursday, 19 November 2020

Sam Wolfe (University of Oxford) – A Fresh Look at Old Italo-Romance ‘Si’

______

Thursday, 3 December 2020

Silvia Terenghi (Utrecht University) – Demonstrative systems in Portuguese-based creoles: Some markedness considerations

Lent Term 2021 (joint seminars with RoLO)

Tuesday, 2 February 2021

Laura Migliori (Universiteit Leiden) – Purpose Prepositional Infinitives in Romance: properties and emergence

______

Tuesday, 9 February 2021

Alberto Frasson (Utrecht University) – Interface properties of subject clitics: a study on antecedent selection

______

Tuesday, 16 February 2021

Paul O’Neill (University of Sheffield) – The boundary between morphology and phonology and the concept of morphologization

______

Tuesday, 23 February 2021

Jan Casalicchio (Università di Palermo) & Peter Herbeck (Universität Wien) – The structure of Pseudo-relatives in Spanish. A preliminary account

______

Tuesday, 2 March 2021

Jan Casalicchio (Università di Palermo) & Michelle Sheehan (Anglia Ruskin University) – Long passives of causatives and perception verbs in Italian: implications for phase theory

______

Tuesday, 9 March 2021

Andrés Saab (IIF-SADAF-CONICET & Universidad de Buenos Aires) – From free pronouns to probes. A theory for a subset of Spanish clitics

______

Tuesday, 16 March 2021

Caterina Bonan (University of Cambridge) – Low focus movement in Romance

Easter Term 2021

Tuesday, 4 May 2021

Clara Cuonzo (University of Maryland) – Expressive adjectives in Romance: a comparative study

______

Tuesday, 11 May 2021

Silvio Cruschina (University of Helsinki) – Ignorance and competence in polar questions: Discourse particles in Sicilian

______

Tuesday, 18 May 2021

Luana Sorgini (Universiteit Utrecht) – A tripartite analysis of DOM. Evidence from Friulian varieties

______

Tuesday, 25 May 2021

Ioana Aminian Jazi (Austrian Academy of Sciences, Vienna) – The Boyash in Hungary. A sociolinguistic outline of two (Romanian) vernaculars in decline

______

Tuesday, 1 June 2021

Alessandro De Angelis (Università di Messina) – Sulla genesi del vocalismo “siciliano”

______

Tuesday, 8 June 2021

Nigel Vincent (University of Manchester) – Suppletion is not silence

______

Tuesday, 15 June 2021

Delia Bentley (University of Manchester) – Two types of external causation and the causative alternation in Romance

______

Tuesday, 22 June 2021

Onkar Singh & Stefania Costea (University of Cambridge) – V2 reconsidered: a comparative analysis of verb-participle inversion in old Daco- and (old) Italo-Romance