Introduction to the Special Issue Syntax and Discourse at the Crossroads

1. This Special Issue

This Issue features eleven papers that explore significant aspects of the syntax–discourse interface using empirical data from different languages (English, German, Spanish, French, and Italian) and their variants. The majority of these papers adopt a formal generative approach and attempt to demarcate the object of study of formal linguistics in addressing the core question of the limits of syntax—that is, whether syntax is self-contained with respect to pragmatics or whether notions of discourse should be included in the syntactic derivation instead, as in the so-called syntactization of discourse. I will briefly discuss here three details of the relationship between syntax and pragmatics that have attracted scholarly attention in recent generative studies and then summarize the papers in this issue, highlighting their contributions to this ongoing debate. The final section outlines the potential paths they pave for further research in this field.

2. The Syntax–Discourse Interface

Given the importance of external interfaces with syntax in the Minimalist Program, the connection between syntax and pragmatics, which is programmatic in the functional theories of language, is studied extensively in present-day generative grammar. In his most recent work, Chomsky has argued that the only uniquely human component of the faculty of language (the so-called faculty of language in the narrow sense) essentially consists of a computational mechanism, Merge, that is capable of generating recursive structures. According to this view, the remaining properties of language originate from the interaction of this mechanism with other mental systems not exclusively devoted to language, such as the articulatory perceptual system (i.e., what we can hear and say, or see and sign, and our capacity to process sounds) and the conceptual–intentional system (i.e., the need to successfully convey meaning and pragmatic intentions); for further reading on this, see, among others, Hauser et al. (2002); Pinker and Jackendoff (2005); and Mendívil-Giró (2018) and the references therein. Consequently, we are spurred to ask which pragmatic information conditions sentence form and should therefore be syntactically encoded in terms of specific categories and discourse features.

On the basis of seminal work by Rizzi (1997), a general consensus has emerged that at least the different types of topics and foci must be syntactically represented, along with the types of information that they convey, essentially correlated with what is taken to be given and new information in the sentence. Researchers therefore seek to provide a fine-grained analysis of the left periphery of the clause, defining the sets of hierarchically ordered projections and features that articulate the different types of topicalized and/or focalized constituents displaced to the periphery of the sentence under particular communicative conditions (see Rizzi, 1997; Cinque, 1999; Frascarelli & Hinterhölzl, 2007; Cruschina, 2012, a.o.).

Another fruitful area of research has focused on the relationship between syntax and the speech act and how it affects the structure of the left periphery. This line of inquiry emphasizes the necessity of incorporating discourse participants such as the speaker and the addressee, along with concepts like commitment, evidentiality, evaluation, mirativity, point of view or sentience which are pertinent to them. These investigations have provided principled explanations for various phenomena whose grammatical properties are substantially influenced by the discursive status of the proposition (see Ross, 1970; Speas & Tenny, 2003; Haegeman, 2014; Miyagawa, 2022; Krifka, 2023, a.o.).

Finally, there is also a growing body of literature on the syntactic, semantic, and discursive properties of the subject. The subject is the DP (a verbal argument or an expletive) that marks the morphological agreement with the verb, and languages have traditionally been parametrized in terms of their SV/VS order and whether null subjects can be licensed in them or not. Indeed, many generative works have discussed the structural position of the subject and the licensing requirements for null categories in its position, together with the discursive import of this lexical/null DP in a sentence. This is crucially connected to the opposition between categorical statements, where something is asserted about an entity, and thetic statements, which are logically unstructured and recognize a certain state of affairs in a given place or at a given time (see, a.o., Kuroda, 1972 and Sasse, 1987); terminological pairs such as declarative/presentational or sentence-focus/predicate-focus sentences refer to this same dichotomy. In categorical statements, the subject is generally understood to be the subject of predication, that is, the aboutness topic of the sentence signaling what the sentence is about (the intentional base, in Ojea, 2017’s terms); in thetic statements, however, the subject is simply an entity involved in the eventuality expressed rather than the topic itself.

Significantly, the categorical/thetic dichotomy finds some structural correlation in Romance languages, with categorical sentences unmarkedly displaying an SV order and thetic sentences a VS order. In relation, some authors contend that the interpretative properties of preverbal subjects in categorical statements can only result from their placement in a non-argumental position within the CP-layer, while others argue that preverbal subjects should sit in Spec-TP, the canonical subject position crosslinguistically, irrespectively of the intentional reading of the sentence (see Lobo & Martins, 2017 for an overview of this controversy). Another point of contention is whether the propositional content of all clauses should be checked against a topic, which would imply that a topic–comment articulation applies to thetic statements too and that they are predicated of a stage topic, as suggested by Erteschik-Shir (1997).

The eleven papers in this volume, which have been arranged alphabetically, serve to delineate the intersection of syntax and discourse further in their presentation of innovative research on grammatical descriptions, the linguistic variation at both the micro and macro levels, code-switching scenarios, and diachronic evolution. These studies also lay the groundwork for future advances in the field since they address languages and constructions understudied so far and adopt methodological tools that facilitate the collection of large samples of naturalistic data.

3. Summary of the Contributions to This Special Issue

Delia Bentley and Francesco Maria Ciconte’s contribution to this Special Issue is situated within the framework of Role and Reference Grammar, a parallel architecture theory which represents discourse–pragmatics and semantics separately from syntax. Their paper focuses on presentational sentences, which, as argued above, are characterized by a verb–subject (VS) order in Romance languages. In certain dialects in Northern Italy, these constructions sometimes incorporate an etymologically locative presentational clitic. By employing data gathered through questionnaire-assisted interviews conducted in situ, Bentley and Ciconte investigate the patterns of the microvariation in these presentational clitics in Milanese and Turinese. Their findings indicate that in Milanese, the clitic ghe, which originated as a resumptive locative pronoun, has evolved into a subject agreement marker. However, in Turinese, alternative findings are uncovered regarding the clitic je, as this clitic does not preclude the number agreement between the verb and the postverbal noun phrase. These authors argue that je in Turinese is not the syntactic subject but rather can be interpreted as the subject of predication, serving as a marker of the aboutness topic of the utterance. The analysis in their paper therefore hinges on a critical distinction between the concepts of the subject of predication and the syntactic subject, which, as previously noted, need not align.

In their study of the behavior of the expletives a and chiru in Fornese and Cilentano, two understudied Romance varieties spoken in Northern and Southern Italy, respectively, Simone De Cia and Mariangela Cerullo also use corpus data to analyze linguistic variation. Utilizing novel data collected during several field trips, they show that a and chiru occur in the same syntactic contexts as subject expletives in non-pro-drop languages but argue that they actually are discourse-pragmatic expletives and as such, the manifestation a formal requirement at the syntax-pragmatics interface. This requirement forces Fornese and Cilentano to satisfy the discourse feature [aboutness] structurally, hence why the expletives a and chiru are externally merged as a last-resort strategy to indicate the lack of aboutness, i.e., the absence of an aboutness referent in the utterance, as applies in thetic statements. Nevertheless, the lexicalization of the discourse–pragmatic expletive does not target the same functional projection in these two languages but is subject to parametric variation: a in Fornese occupies the higher portion of the TP-domain, whereas chiru in Cilentano satisfies [aboutness] in the canonical syntactic position for overt aboutness-shift topics within the C-domain.

Mara Frascarelli and Giorgio Carella’s paper tackles the occurrence of null subjects in non-pro-drop languages, highlighting the intricate interplay between the structural, semantic, and discourse factors behind the omission of the subject. It analyzes data drawn from three transcripts from two online corpora of Modern Colloquial French, namely the CLAPI and the CFPP2000. Their statistical analysis of the results reveals that the overwhelming majority of null subjects in Colloquial French result from the omission of expletive subjects in predicational sentences. Thus, Frascarelli and Carella contend that the subject il in these predicational contexts is an expletive pronoun and susceptible to being dropped; in contrast, ce, the subject of non-predicational copular sentences in French, is not a true expletive but rather a referential pronoun, making its omission by speakers less likely. The authors also demonstrate that referential DPs may be omitted when they establish an Agree relationship with a referent that is strongly active in the current discourse context, i.e., an aboutness-shift topic. Lastly, in comparing French and English, they confirm that the omission of the subject is permissible in both languages but that referential subjects are omitted significantly more frequently in English than they are in French.

In his investigation of the positioning of subjects, Roland Hinterhölzl correlates the distribution of DPs in German with the phenomenon of complementizer agreement in West Germanic languages. He assumes that the context values become accessible in a specific functional head in the C-domain and proposes that in German, indefinite DPs and weak quantifiers are interpreted within the V-domain, while attributively used definite DPs and strong quantifiers (unless they are discourse-anaphoric) are interpreted in the T-domain. Conversely, referentially used definite DPs and anaphoric strong quantifiers (i.e., discourse-anaphoric DPs) must access the C-domain and enter into a licensing relationship with the head Fin⁰ to be fully licensed. Hinterhölzl then outlines the pattern of subject licensing in Cimbrian, a German dialect spoken in the village of Luserna, Trentino, to illustrate how this pattern is replicated in the systems of complementizer agreement (CA) in West Germanic dialects, a multifaceted and complex phenomenon that is more substantial than a formal peculiarity. He demonstrates that the data from Cimbrian and the CA in West Germanic show striking similarities and warrant a unified explanation in which CA is understood as serving the purpose of anchoring the subject into the context.

Ángel L. Jiménez-Fernández and Mercedes Tubino-Blanco discuss inferential interrogative sentences with qué in Spanish, which exhibit the shape of a wh-question but the interpretation of a polar question and convey a reading of evidentiality (e.g., ¿Qué vas, en coche? ‘Are you going by car (I infer)?’). They propose a multi-layer system with two contrasted levels in their internal composition—an utterance level, configured as an interrogative speech act, and a clausal level—and discuss the prominent roles of the speaker and the addressee in the derivation. Although the evidential element in inferential interrogative sentences has an unusual speaker-oriented interpretation, Jiménez-Fernández and Tubino-Blanco demonstrate that the Interrogative Flip—a noteworthy cross-linguistic feature of interrogatives with morphologically overt evidentials—is also activated in these constructions due to the way in which the speaker and addressee are anchored within the discourse: the speaker draws the inference, while the addressee is solicited for confirmation. Their formal analysis therefore accounts for this unconventional interpretation of the interrogative clause, its hybrid nature as both a wh- and yes/no question, and its distinctive intonation pattern of fall–rise.

The paper by Sergio López-Martínez examines Old English, which has traditionally been characterized as a V2 language. In particular, it analyzes a productive presentational construction from this period in which an adverb or a prepositional phrase indicating location occupies the leftmost position of the clause, resulting in inversion of the finite verb. For López-Martínez, this construction is equivalent to locative inversion in Present-Day English, which has often been viewed as a root phenomenon. However, as demonstrated in this article, it can occur quite productively in non-root subordinate clauses in Old English. Utilizing data from the York–Toronto–Helsinki Parsed Corpus of Old English Prose, López-Martínez qualitatively and quantitively analyzes various types of subordinate constructions that feature clause-initial prepositional phrases (PPs) with and without finite verb inversion. He documents a relatively balanced distribution of PP-S-V and PP-V-S word orders in subordinate sentences and investigates the factors that contribute to the occurrence of a PP-V-S order in these clauses in Old English. He concludes that while the type of verb is a critical determinant of the inversion in main clauses, in subordinate clauses, the length and type of the subject are more critical, with longer and heavier subjects triggering finite verb inversion more readily.

María Mare focuses on the so-called PRO_[PL] with-DP construction in examples such as “Con mi mujer nos casamos en abril”, “My wife and I got married in April”, which can only be fully understood by involving syntactic and informational considerations. The verb in this type of construction, as observed in some varieties of Spanish (such as Argentinian and Chilean Spanish), exhibits dual number information despite the fact that the reference of the argument, introduced by con, “with”, is a single entity. Additionally, both holistic and distributive interpretations of the predicate are permitted, even if a comitative item that generally voids a distributive reading is present. Mare reexamines this kind of construction and proposes that the with-DP phrase constitutes part of a complex subject DP containing an empty plural pronominal, which may be doubled by an overt pronoun being positioned at the left periphery of the sentence. She approaches the analysis from an information structure perspective, with two mechanisms determining the referential properties of the pronominal and the plural number agreement with the verb: one involving the features associated with person-related information, such as the combination of [−Participant]/[+Author], and the other pertaining to distinctions related to anaphoricity.

Research addressing the intricate relationship between code-switching—the integration of multiple languages into a single sentence or discourse—and the organizational principles that govern information structure is currently limited. However, Antje Muntendam and M. Carmen Parafita Couto’s paper aims to address this gap by providing a comprehensive overview of the current, albeit sparse, research concerning the intersection between code-switching and information packaging. It is widely understood that the information structure across languages can be conveyed through various means, including syntax (e.g., word order), morphology (e.g., the use of topic and focus markers), and prosody (e.g., intonation). Consequently, code-switching between languages that employ distinct methods for expressing information structure is expected. To illustrate this phenomenon, Muntendam and Parafita Couto examine specific instances of code-switching at the interfaces, demonstrating how the study of interfaces informs the study of code-switching (and vice versa) and allows hypotheses to be tested that cannot be examined using monolingual data alone. They also acknowledge the limitations of the existing research and discuss certain theoretical and methodological considerations that should guide future studies on code-switching, emphasizing the need to use naturalistic data to properly analyze the prosody, syntax, and morphology involved in the expression of the information structure within code-switched speech.

My own paper, “The Syntax of Speech Acts: Deictic Inversion as an Evidential Strategy in English”, offers a novel analysis of deictic inversion (DI) in English, a construction which is used by a speaker to gesture towards a proximal or distal location, directing the addressee’s attention to an entity related to said location (e.g., “Here comes the bus”). Superficially, DI has much in common with standard locative inversion (LI): In both cases, the sentence is understood as a non-predicative assertion of a state of affairs where the grammatical subject receives the informational focus. Furthermore, these two constructions are headed by the same types of verbs (the copula be or unaccusative predicates) and feature a locative constituent at the beginning. I explain that DI functions as an evidential strategy in English, a language which lacks standard evidential markers, and that this specific discursive status explains its structural properties and its main differences from locative inversion. My analysis of DI as an evidential strategy also accounts for its restricted distribution and the otherwise unexpected difference between DI and its non-inverted counterpart in readings of the present/past tense.

In his paper, Imanol Suárez-Palma explicitly engages in the debate about the informational status of preverbal subjects, a controversial issue in the current literature regarding the projection of pragmatic information within syntax. He investigates a type of middle-passive sentence in Spanish with a dative possessor in a preverbal position, either as the sole fronted constituent or used in conjunction with the theme DP, which functions as the syntactic subject, i.e., “A Ismael se le ven las arrugas fácilmente”, “A Ismael las arrugas se le ven fácilmente, or “Las arrugas a Ismael se le ven fácilmente”, with all of these sentences meaning “Ismael’s wrinkles are easy to see”. Suárez-Palma’s analysis posits that the possessor originates within the theme DP and rises to the position of the specifier of an applicative projection in order to be licensed with the dative case. To circumvent the minimality violation that would arise if the preverbal theme DP were to occupy the canonical Spec-TP position, he proposes that both the preverbal lexical dative and the theme DP function as left-dislocated constituents, co-referring with empty pronominals in argument positions. This proposal then supports the view that preverbal subjects and dative DPs ultimately occupy non-argumental positions in Spanish when they serve as the aboutness-shift topic.

Last but not least, Julio Villa-García’s paper investigates the syntax of the clausal left edge, parameterizing (and micro-parameterizing) the various realizations of the left-peripheral heads observed in English and Spanish. Specifically, Villa-García explores the lexicalization of the complementizers that and que in their respective linguistic varieties, demonstrating their presence in nearly all constructions traditionally associated with the CP discourse domain, including exclamative clauses, interrogative contexts, and subjunctive clauses. This study reveals the significant degree of variation in the lexicalization of these complementizers, identifying instances where the complementizer may remain silent (e.g., the high that in English), configurations where both the complementizer and the left-peripheral phrase are simultaneously realized (e.g., exclamatives that use vaya and que in certain varieties of Asturian Spanish), scenarios in which only the left-peripheral constituent is expressed (such as wh-interrogatives in Peninsular Spanish), and cases of recomplementation involving an element flanked by overt instances of the complementizer. Through a detailed analysis of these options, Villa-García concludes that the presence or absence of that or que may be indicative of a complex underlying structure and cannot be merely reduced to a pronunciation parameter. On the contrary, complementizer lexicalization seems to be processing-based, with discourse playing a crucial role in determining these various options across linguistic varieties.

4. Looking Ahead

The issues discussed in this volume represent just a small sample of the various aspects that formal theories must address in order to gain a full understanding of the syntax–pragmatics interface and thus pave the way for future research in this domain.

In studying the syntax–pragmatics interface, linguists aim to clarify the interaction between the computational system and the intentional system. This requires moving away from the strong conception of modularity that was foundational to the Chomskyan generative approach and towards the assumption that the computational system is not blind to discourse. However, to maintain the theory’s predictive power and explanatory adequacy, it is essential to avoid an unconstrained syntactic representation of pragmatic information. Thus, much more research is required to establish a sufficient yet restrictive repertoire of pragmatic features and functional categories that should be integrated into the syntactic structure. Additionally, defining all of the constructions in which these pragmatic features play a crucial role remains crucial, specifically constructions in which the constituents are organized based on their relative salience in discourse or their contribution to illocutionary values. Finally, to provide new insights into the interplay between grammatical and pragmatic competence, the range of languages examined must be broadened, and larger samples of data must be sourced.