Verbless Constructions

We present here the first data collected on the AN.ANA.S. Multilingual Treebank (AN.ANA.S._MT), consisting of Italian, English and Spanish task-oriented dialogues and spontaneous speech for a total of about 21300 words. AN.ANA.S. (Annotazione e Analisi Sintattica) is a system freely downloadable at the portal www.parlaritaliano. it, developed within a treebank project for the syntactic annotation of both spoken and written texts (Voghera and Cutugno in press). The system is constituent-based and allows the organization of syntactic units within a hierarchical structure, according to XML standards. Constituent relations are coded directly using elements nesting XML properties. Our research focuses on the treatment of ‘non-canonical’ syntactic structures, such as verbless sentences, which are usually marginal in the annotation schemes of treebanks. In this paper we will take into account both methodological and quantitative aspects. Firstly, we will compare how verbless sentences are ...

Polar verbless clauses and gapping seem to differ in their capacity to be embedded. While polar verbless clauses can be easily subordinated, gapping constructions are traditionally considered main clause phenomena restricted to root contexts. Nevertheless, some languages, like Farsi, Rumanian and Spanish seem to allow gapping embedding with some particular predicates. This paper provides corpus data which show the extent of the capacity of subordination of Spanish gapping constructions, and their differences to the less restricted polar verbless clauses: on the one hand, gapping, like other fragments, can be embedded by verbal and non-verbal epistemic predicates. On the other hand, polar verbless clauses can be subordinated to these predicates, but are not restricted to them. They are much more frequently embedded, as can be seen by their distribution in the different genres of the CORLEC corpus.

Along with the verbal sentence (where a subject accomplishes an action), Lucien Tesnière argues that there is an attributive sentence—where a subject receives an attribute. Two attributive relations exist: (1) sentences with ‘to be’, and (2) verbless sentences. Focusing on the verbless sentence of Biblical Hebrew, this paper explains that the absence of the verb has a linguistic consequence: following Harald Weinrich, the verbless sentence is dependent on its ‘linguistic context’. That is, its syntactic value is provided by its syntactic VSO or SVO ‘head’.
This paper begins with a methodologic discussion drawing on Tesnière and Weinrich, and proceeds with a grammatical and a text-linguistic analysis of the verbless sentence as found in the indirect speech of 1Samuel. With regard to the claim of Zewi 1994 (‘the subject and predicate of the nominal sentence are indistinguishable from … “theme” and “rheme”’), this paper explains that her position is methodologically unconvincing and invalidated by several examples.

The paper presents an overview of different syntactic gaps in modern colloquial Russian and focuses on so-called ‘syntactic emptiness’, i.e. the zero copula, the referential zero lexemes Øljudi and Østixii, the generic ‘you’ Øty2 and omission of verbs of action [ØV], taking into account the pragmatical impact of these phenomena (text economy, explicitness, politeness). In addition, the study provides an historical review of colloquial empty slots in Russian (on the example of Avvakum’s ‘Autobiography’) and approaches it from a typological perspective, specifically, by comparing it to West Slavic languages (Polish, Czech) as well as to other languages which corroborates the idea of the “Eurasian orientation” of colloquial Russian.

Keywords: syntactic emptiness; zero sign; ellipsis; colloquial Russian; typology

Una serie di linee guida per l'annotazione manuale degli enunciati nominali (o frasi senza verbo) in un corpus di testi dialogici informali scaricati dal web.
Queste linee guida tengono conto della varietà di forme degli enunciati nominali e della natura substandard dei testi annotati.
Queste istruzioni sono state applicate durante l'annotazione del corpus COSMIANU (liberamente scaricabile a questo link: https://hlt-nlp.fbk.eu/technologies/cosmianu1.0), presentato a CLiC-it 2018.

The paper discusses nominal clause patterns in the Dead Sea Scrolls. Nominal clauses were collected from several Dead Sea Scrolls (DSS) that represent an independent language, and are not late versions of biblical texts. These are Pesher Habakkuk, the Damascus Document, the War Scroll, The Temple Scroll, and Serekh Ha-Yah . ad. The DSS nominal clause patterns revealed in these sources can be divided into two major groups: (1) simple nominal clauses in the order of predicate-subject or of subject-predicate, and (2) extended nominal clauses involving extraposition. Although both word orders of the simple nominal clauses found in the DSS exist in Biblical Hebrew, those of the DSS display more restrictions in their preference for a certain word order with certain types of predicates. The DSS simple nominal clauses in the order of subjectpredicate also contain a pattern, unknown in Biblical Hebrew and limited in Mishnaic Hebrew, in which the predicate is a content clause. This pattern appears only in Pesher Habakkuk, and its content clause is introduced there by the subordinate, more commonly relative particle, asher. Extended patterns appear to be very rare in the DSS, and they mostly include clauses also labeled as the exegetical formula, which show similarity to a certain type of appositional clauses in Biblical Hebrew.

In the article I examine utterances without finite verbs in different varieties of contemporary Polish, notably in new genres of internet communication (blogs and forums), colloquial usage and narrative texts. This subject is motivated by the behavior of colloquial Russian which allows the omission of finite verbs far more freely than colloquial Polish does. The study attempts to seize some crucial factors either favoring or blocking the emergence of verbless sentences in the abovementioned Polish varieties. Anaphoric ellipsis, missing copulae and grammatical vagueness are not taken into account. The article is mainly based on data collected through Google searches and my own observation of colloquial usage. Its principal results may be summarized in the following way: 1. The omission of finite verbs is much more restricted in Polish spontaneous speech than in its Russian counterpart; we observe however the rise of new verbless structures strongly reminiscent of colloquial Russian in Polish internet communication. 2. The degree of explicitness/ vaguess or ambiguity does not seem to have an impact on the acceptability of a given construction with missing verb. 3. The narrative register proves to be more favorable for the omission of verbs of movement and communication than the discourse register.

This paper proposes a methodology for investigating populism by analyzing proto-slogans, nominal utterances (NUs) typical of a political community on social media. We extracted more than 700.000 comments from the public Facebook pages of two Italian populist parties' leaders (Matteo Salvini and Luigi Di Maio) during the week preceding the 2019 European elections (i.e., from May 20 to May 26, 2019). These comments have been automatically clustered and manually annotated to find protoslogans coming from the parties' supporters (bottom-up). We applied four layers of analysis: Nominal Utterances (NUs), a syntactic device widely used for slogans; Slogans for NUs with a slogan function; Top-down/Bottom-up, to recognize the slogans produced by the politicians and those produced by supporters; Protoslogans, for NUs devoid of specific political content while expressing partisanship and support for the leaders.

The properties and use of fragments (or elliptical clauses) are recently being object of attention in different works (Fernandez 2002, Merchant 2004, Schlangen 2003). There is no agreement, however, concerning their nature and classification. Firstly, some authors treat them as pure syntactic units: the remnants of verbless clauses which have undergone ellipsis (Merchant 2004). Secondly, others classify them as pragmatic objects, different from non-elliptical clauses (Schlangen 2003), by their function in discourse. Thirdly, other works stress their independence from non-elliptical clauses and classify them with a combination of syntactic and pragmatic criteria (Fernandez 2002).

The aim of this paper is to show the extent to which fragments and polar verbless clauses (“yes”, “no”) can be analysed as syntactic or discourse units, as well as to determine a typology based in their syntactic and pragmatic properties and to present their distribution in the different genres of a corpus. In order to achieve this goal, we have retrieved the totality of fragments in the corpus of contemporary oral Spanish (CORLEC) (Marcos Marín 1992), composed by more than 63 000 utterances and we have classified them according to their syntactic and pragmatic properties. Finally, we have counted the frequencies of each type in the different genres.

The results of this analysis indicate that fragments containing a segment with a counterpart in their source have a predictable discursive relationship with it: they perform a particular speech act (answer, agreement, correction, check question, etc.) that is determined by the syntactic and semantic properties of the source and the target clauses. This combination of properties is detailed in the following table, with reference to constructed examples of the various speech acts.

Dans le discours oral on trouve une grande variété de structures dotées d’un contenu sémantique phrastique dont la syntaxe n’exprime qu’une partie de ce contenu. En effet, ces structures, qu’on appelle fragments, récupèrent une partie de leur contenu des structures qui le précèdent, comme les exemples en gras de (1) et (2).

(1)  A: -¿Qué quieres? B: -Un perro.    A: -‘Qu’est-ce que tu veux ?’ B: -’Un chien.’
(2)  A: -Mira mi vestido. B: -¡Magnífico!        A: -’Regarde ma robe.’ B: -’Magnifique !’

Ces unités, communément analysées comme des phrases elliptiques, présentent une grande diversité de propriétés syntaxiques, sémantiques et discursives qui, bien qu’ayant été objet de quelques études récents (Lefeuvre 1999 et suivants ; Fernández, R. and J. Ginzburg (2002) et suivants ; Abeillé & Delaveau 2016), présente de nombreux aspects qui restent encore à être clarifiés, comme par exemple, leur typologie. Ainsi, l’analyse de ces propriétés nous permet de distinguer des fragments qui correspondent à des phrases qui ont (au moins) une tête verbale elliptique (1) de ceux qui correspondent à des phrases averbales (2) dont l’argument sélectionné par la tête averbale est elliptique. De même, le fragment, comme construction à propriétés syntaxiques, sémantiques et illocutoires précises, se distingue du concept d’énoncé, comme on peut noter par l’accumulation de plusieurs fragments au sein d’un même énoncé (3,4). En conséquence, il est nécessaire de distinguer les fragments composés par plusieurs syntagmes (5) de l’accumulation de plusieurs fragments dans le même énoncé (3,4).

(3) A: -Nos vamos a la playa. B: -Vale, perfecto. ¿Cuándo?
    A: -’Nous allons à la plage.’ B: -’Ok, parfait. Quand ?’
(4) ¿Ha venido alguien? Claro. María. A: -Quelqu’un est venu ? B: -’Bien sûr. Marie.’
(5) A: -¿Quién va a venir? B:-Seguramente María.
  A: -’Qui viendra ?’      B: -’Probablement Marie.’

Enfin, les fragments semblent se distinguer des marqueurs d'énonciation par une propriété précise : les fragments ont un contenu sémantique phrastique et réalisent un acte illocutoire lors de leur énonciation. Les marqueurs, en contraste, réalisent un acte illocutoire sans avoir de contenu phrastique. Cette propriété permet de les différencier, même si en occasions un même mot peut être employé comme fragment et comme marqueur, comme on peut voir en (6,7), où l’on distingue le fragment, qui apporte un contenu évaluatif (6) (équivalent à “ton article me semble bon”) et réalise l’acte de réponse, du marqueur sans contenu phrastique (7), qui réalise l’acte illocutoire d’indiquer suivie de la conversation et acquiescement.

(6) ¿Qué te parece mi artículo? - Bueno.
  A: -’Comment te semble-t-il mon article?’ B: -’Bon.’
(7) Nos vamos -Bueno.      A: -’On y va.’ B: -’Bon.’ (=OK.)
Nous proposons donc une typologie qui permet de distinguer, d’un côté, les fragments issus des phrases averbales avec ellipse de l’argument sélectionné par la tête prédicative (2), et de l’autre, les fragments qui correspondent à des phrases verbales elliptiques (6). Parmi les premiers on trouve des fragments avec des contenus sémantiques évaluatifs (2) et épistémiques (5), alors que parmi les secondes on trouve des fragments avec coréférence, comme María en (4), et sans coréférence, comme ¿Cuándo? ‘Quand ?’ en (3).

Enfin, on trouve aussi des fragments qui complètent leur contenu grâce à un contenu illocutoire : un pour chacun des cinq types majeurs d’acte de parole de Searle (1969) : représentatifs (-Encantado. ‘Enchanté’), expressifs (-Buenas tardes ‘Bonsoir’), promissifs (-Una cerveza? ‘Une bière ?’), directifs (¡A comer!. ‘A table !’) et déclaratifs ou performatifs (¡Castigado! ‘Puni !’). De même, la typologie de fragments avec coréférence peut être enrichie si l’on distingue l’acte de parole résultant de la relation entre la source et la cible de la coréférence : réponse (4), acquiescement (vale ‘d’accord’ en (3)), suivi, demande de confirmation et correction.

Cette classification a été réalisé grâce au recueil et à l’analyse des propriétés de l’ensemble d’énoncés sans verbe du corpus CORLEC (Marin, 1992) de l’espagnol oral contemporain. Ce corpus est composé d’environ 1 100 000 mots et plus de 63 000 énoncés, distribués en 17 genres différents, qui assurent une quantité et une diversité langagière. Ce travail sur corpus fournit une profusion et une variété d’exemples qui permet de rendre compte des possibilités d’articulation syntaxique des fragments, tels que leur structure et type syntaxique, leur valeur illocutoire et la partie de discours de la tête. Ce travail permet également de noter la distribution de chaque type de fragment dans les différents genres qui composent le corpus, regroupés en tant que dialogiques ou monologiques.

The paper presents a corpus-based contrastive analysis of the predication involved in the translation of verbless sentences from Russian to English based on a pilot parallel corpus consisting of Dostoyevsky’s dialogue-based Russian Brat’ja Karamazovy (1880) and the Pevear and Volokhonsky English translation The Brothers Karamazov (1990). In contrast to English, known for its dependency on the finite verb phrase, Russian permits the use of verbless sentences more productively than any other Indo-European language (McShane, 2000; Kopotev, 2007). Combining the parallel-text approach to contrastive linguistics developed by Guillemin-Flescher (2003) with a new method of automatic verbless sentence extraction, the present study examines reoccurring patterns regarding the way that predication is gained or lost in translation. Following automatic segmentation, morphosyntactic annotation and extraction, verbless sentences and their translation correspondences are manually annotated for verbal and non-verbal predication in accordance with Hengeveld’s (1992) definitions. The results present a typology of a phenomenon we call ‘predication transformation’, in which translation correspondences are transformed in terms of predication type. Quantitative results reveal the rate at which verbs are gained in translation of the verbless sentences from Russian to English, as well as the predication verbalization rate. We argue that a verb-centric notion of semantic predication is not cross-linguistically stable.

https://www.researchgate.net/publication/333076161_Predication_Transformation_A_parallel_corpus-based_study_of_Russian_verbless_sentences_and_their_English_translations

English. Nominal utterances are very frequent, especially in social media texts, and play a crucial role as they are very dense from a semantic point of view. In spite of this, their automatic identification has received little to no attention. We have thus developed a framework for the annotation of nominal utterances and created the manually annotated corpus COSMIANU (Corpus Of Social Media Italian Annotated with Nominal Utterances), which could be used to train automatic systems. Italiano. Gli enunciati nominali sono un fenomento linguistico molto frequente, specialmente nello scritto dei social media, e di cruciale importanza, data la loro alta densità semantica. Tuttavia, ben poca attenzione è stata dedicata al loro riconoscimento automatico. In quest’ottica, questo lavoro illustra le guidelines per l’annotazione manuale degli enunciati nominali da noi sviluppate e presenta il corpus dell’italiano dei social media da noi annotato con gli enunciati nominali (COSMIANU), utilizzab...

Entre los marcadores discursivos recogidos por Zorraquino y Portolés (1999), aquellos que expresan modalidad epistémica (por supuesto, seguro) y deóntica (bien, genial) muestran un comportamiento particular que los distingue del resto, como puede verse en los ejemplos extraídos del corpus CORLEC (Marcos Marín, 1992). A diferencia del resto de marcadores, éstos seleccionan un contenido semántico, tienen un núcleo predicativo, pueden expresar el contenido seleccionado como argumento frástico y pueden aparecer en subordinación. Estas propiedades se encuentran también en las frases averbales predicativas, lo que sugiere que estos marcadores no deberían clasificarse como tales, sino como un caso particular de frases averbales. Sintaxis, enunciados sin verbo, marcadores discursivos, frases averbales, modalidad 1. Introducción Los marcadores discursivos, también llamados enlaces extraoracionales (Gili Gaya, 1961), marcadores discursivos, conectores argumentativos o procesadores textuales, han sido descritos y clasificados por Zorraquino y Portolés (1999), que los definen de la siguiente manera: « Son unidades lingüísticas invariables, no ejercen una función sintáctica en el marco de la predicación oracional-son, pues elementos marginales-y poseen un cometido coincidente en el discurso: el de guiar, de acuerdo con sus distintas propiedades morfosintácticas, semánticas y pragmáticas, las inferencias que se realizan en la comunicación. » Zorraquino y Portolés (1999) clasifican los marcadores del discurso en cinco grupos según sus «funciones discursivas»: primero, los estructuradores de la información (pues bien, por otro lado, Por cierto), que comentan y ordenan la información contenida en un discurso; segundo, los conectores (incluso, en cambio, sin embargo), que vinculan semántica y pragmáticamente un miembro del discurso con otro anterior; tercero, los reformuladores (o sea, de todos modos, en conclusión), que introducen una nueva formulación de un miembro anterior; cuarto, los operadores argumentativos (de hecho, por ejemplo), que delimitan las posibilidades

The difficulty of automatic processing has forced most analyses of verbless sentences to rely on fragmented data. This paper overcomes fixed annotation and verb-centric syntactic modelling to present a new method of automatic retrieval of verbless sentences. Parallel-text corpora in English and Russian are used with the aim of revealing the semantic factors associated with the absence of the verb. Relying on automatic and manual annotation, the analysis is both quantitative, through the use of descriptive and inferential statistics, and qualitative, through the means of extracting, annotating, and interpreting the results. We propose a novel approach to identifying the differences between English and Russian in terms of their use of verbless sentences. In addition to showing the statistical significance of anticipated verbless sentence frequency differences between the two languages, the results reveal a surprising over-representation of antecedent-based ellipsis in English, and provide several prospects for further quantitative and qualitative corpus-based research of verbless sentences.

https://doi.org/10.1080/00393274.2019.1616221

Les études sur l'ellipse se concentrent généralement dans des phénomènes d'ellipse à l'intérieur de la phrase et souvent se limitent à la coordination, comme le gapping ou le sluicing. En revanche, l'ellipse extra-phrastique a reçu une certaine attention récente, notamment dans l'étude des énoncés comme la réponse courte. Cependant, il y a peu de consensus quant à l'analyse de ces unités 1 . Ainsi, le domaine de l'ellipse extra-phrastique, constitué des phrases elliptiques ou fragments, est encore aujourd'hui peu connu. De même, la relation que ces unités entretiennent avec la phrase averbale, dont la tête n'est pas elliptique mais réalisée par une partie du discours différente du verbe, reste à déterminer. L'objectif de cet article est ainsi de présenter une typologie basée sur les propriétés des structures averbales de l'espagnol, qui inclut aussi bien les fragments que les phrases averbales, mais aussi d'exposer les relations qui s'établissent entre les deux types de structures. Pour cela, nous avons extrait l'ensemble de structures averbales du corpus CORLEC Corpus Oral de Referencia de la Lengua Española Contemporánea) (Marcos-Marín, 1992).

CLARIN Workshop on Translation memories, corpora, termbases: Bridges between translation studies and research infrastructures,
Austrian Academy of Sciences

https://digital-humanities.at/en/dha/s-news/call-participation-clarin-workshop-translation-memories-corpora-termbases-bridges-between

Polar verbless clauses and gapping seem to differ in their capacity to be embedded. While polar verbless clauses can be easily subordinated, gapping constructions are traditionally considered main clause phenomena restricted to root contexts. Nevertheless, some languages, like Farsi, Rumanian and Spanish seem to allow gapping embedding with some particular predicates. This paper provides corpus data which show the extent of the capacity of subordination of Spanish gapping constructions, and their differences to the less restricted polar verbless clauses: on the one hand, gapping, like other fragments, can be embedded by verbal and non-verbal epistemic predicates. On the other hand, polar verbless clauses can be subordinated to these predicates, but are not restricted to them. They are much more frequently embedded, as can be seen by their distribution in the different genres of the CORLEC corpus.

Spanish verbless utterances in the CORDE corpus have been classified in a taxonomy and annotated for distribution frequency and syntactic properties (part of speech of the head, structure and syntactic type). This work has allowed to note that Spanish verbless utterances are a non-negligible part of oral utterances: they amount to around 19% of the 63,000 utterances from the corpus, both in root and subordinate contexts. Among these verbless utterances, fragments are significantly more frequent as roots than verbless clauses, but they are both equally rare in subordination.

The variety and profusion of verbless utterances drawn from the oral corpus CORLEC has allowed us to present a detailed classification of them. Verbless utterances with clausal content can be classified as clausal (non-elliptic) or non-clausal (elliptic) structures, also called fragments. Among the former, predicative verbless clauses are composed by a predicative head and the argument they select. This paper exposes the properties and structures of both verbless clauses and fragments found in the corpus, and analyses the syntactic and informational status of the NP selected by the predicative head of predicative verbless clauses. 1 Introduction Verbless utterances constitute a highly significant part of the total of utterances in a spoken corpus of Spanish, as it is shown by Cresti and Moneglia (2005), who counts 37% in the corpus CORAL -ROM. This proportion may vary according to the genre of the corpus, as it can be seen in the different distribution of frequencies in the CORLEC corpus (Marcos-Marín, 1992), where monologic genres contain less verbless utterances than dialogic genres. This proportion is nevertheless smaller if we leave aside verbless utterances with no propositional meaning. Interjections without arguments (1), calls (2), fillers (3) and discourse markers (4) do not express clausal meaning, and are thus different from verbless clauses, which have a non-verbal head (5), and fragments, which have an elliptical head (6) or an elliptical argument (7). (1) ¡Oh! (2) ¡Maria! (3) A:-Vamos a… B:-¿casa? A:-'Let's go…' B:-'Home?' (4) El tren llega a las tres, ¿verdad? 'The train arrives at three, truth?' (i.e.: 'isn't it?') (5) ¡Qué poca inteligencia la tuya! 'What little intelligence yours!' (i.e.: 'from you') (6) A:-Y mira, y así nos vemos. B:-Pues entonces ¿el miércoles a las seis? A:-'Look, this way we will see each other'. B:-'So then, Wednesday at six?'

BONDARENKO (Antonina) et CELLE (Agnès), « Traduire l’absence. Les questions à prédicat zéro dans un corpus parallèle russe et anglais », Des mots aux actes, n° 7, 2018, Sémantique(s), sémiotique(s) et traduction, p. 341-363.

What are the semantic and pragmatic factors associated with the absence of a predicate? By focusing on zero-predicate questions and their Russian and English translations in direct speech sequences, this study adopts an approach that combines the methods of both corpus and contrastive linguistics. Recurrent translation patterns highlight linguistic differences that would otherwise remain hidden. The zero-predicate questions tend to be rhetorical questions and are more frequent in English-Russian than in Russian-English translations.

Quels sont les facteurs sémantiques et pragmatiques associés à l’absence de prédicat ? En se focalisant sur les questions à prédicat zéro dans des séquences de dialogue en russe et en anglais et en examinant leurs correspondances en traduction, cette étude adopte une approche qui combine les méthodes de la linguistique de corpus avec celles de la linguistique contrastive : les modèles récurrents de traduction mettent en évidence les différences linguistiques qui autrement resteraient cachées. On montre d’une part que les questions à prédicat zéro ont tendance à être des questions rhétoriques, d’autre part que les questions à prédicat zéro apparaissent beaucoup plus dans le sens anglais – russe que dans le sens russe – anglais.

https://www.researchgate.net/publication/331700883_Traduire_l'absence_les_questions_a_predicat_zero_dans_un_corpus_parallele_russe_et_anglais

ENGLISH
Nominal utterances are very frequent, especially in social media texts, and play a crucial role as they are very dense from a semantic point of view. In spite of this, their automatic identification has received little to no attention. We have thus developed a framework for the annotation
of nominal utterances and created the manually annotated corpus COSMIANU (Corpus Of Social Media Italian Annotated with Nominal Utterances), which could be used to train automatic systems.

ITALIANO
Gli enunciati nominali sono un fenomento linguistico molto frequente, specialmente nello scritto dei social media, e di cruciale importanza, data la loro alta densit`a semantica. Tuttavia, ben poca attenzione `e stata dedicata al loro riconoscimento automatico. In quest’ottica, questo lavoro illustra le guidelines per l’annotazione manuale degli enunciati nominali da noi sviluppate e presenta il corpus dell’italiano dei social media da noi annotato con gli enunciati nominali (COSMIANU), utilizzabile per addestrare sistemi automatici.

The Hate Speech Detection (HaSpeeDe 2) task is the second edition of a shared task on the detection of hateful content in Italian Twitter messages. HaSpeeDe 2 is composed of a Main task (hate speech detection) and two Pilot tasks, (stereotype and nominal utterance detection). Systems were challenged along two dimensions: (i) time, with test data coming from a different time period than the training data, and (ii) domain, with test data coming from the news domain (i.e., news headlines). Overall, 14 teams participated in the Main task, the best systems achieved a macro F1-score of 0.8088 and 0.7744 on the in-domain in the out-of-domain test sets, respectively; 6 teams submitted their results for Pilot task 1 (stereotype detection), the best systems achieved a macro F1-score of 0.7719 and 0.7203 on in-domain and out-of-domain test sets. We did not receive any submission for Pilot task 2.

Illocutive verbless utterances (like Good luck with your exams!) have been analysed as elliptical structures. Nevertheless, this analysis presents some problems, since these units do not take their semantic content from the syntactic context, but from situational context, as, for instance, the type of speech act they realize. This article aims to determine whether these verbless utterances that express speech acts are clausal structures, having the syntactic variety that characterizes clauses (syntactic type, part-of-speech of the head, and syntactic structure) or lexicalized units. For this purpose, we have extracted the 1 578 occurrences of this kind of utterances found in the corpus CORLEC and we have classified them by the speech act they perform and by their syntactic properties. This analysis has allowed us to determine the place these units occupy within the classification of clause types.