GDGauthier Delaby

Gauthier Delaby

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Gauthier Delaby has a total of 7 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Constructional contamination through horizontal links: evidence from a structural priming experiment on Dutch passives

    by Gauthier Delaby, Timothy Colleman, Robert J. Hartsuiker

    Pijpops and Van de Velde (Constructional contamination: How does it work and how do we measure it? Folia Linguistica 50(2). 543–581) observe how an alternation within a construction can be influenced by neighbouring constructions with a different meaning but a similar form via a process they call constructional contamination. Previous studies found that if an instance of the target construction uses a lexeme that is strongly associated with a contaminating construction or if the instance is very similar in meaning to the contaminating construction, the instance of the target construction is more likely to appear with the alternant that is typically associated with the contaminating construction. While this was originally explained as the consequence of the storage of unanalyzed exemplar chunks produced by a contaminating construction, which are later recycled in a target construction (Pijpops, Dirk & Freek Van de Velde. 2016. Constructional contamination: How does it work and how do we measure it? Folia Linguistica 50(2). 543–581), it has since been proposed that such contamination effects may also arise because of the so-called horizontal similarity links that are assumed to exist between distinct but formally similar constructions in construction grammar (Delaby, Gauthier & Timothy Colleman. 2025a. Constructional contamination between two constructions with krijgen ‘to get’ in Dutch: National variation, bidirectionality and ambiguity avoidance. Constructions and Frames . Advance online publication). In this paper, we test this hypothesis by means of a structural priming experiment. More specifically, we test whether the choice for the preposition van ‘from/by’ or door ‘through/by’ in the agent PP of the Dutch so-called “receptive” construction with krijgen ‘to get’ (Colleman, Timothy. 2015. Constructionalization and post-constructionalization: The constructional semantics of the Dutch krijgen-passive from a diachronic perspective. In Jóhanna Barðdal, Elena Smirnova, Lotte Sommerer & Spike Gildea (eds.), Diachronic construction grammar , 213–256. Amsterdam: John Benjamins Publishing Company; Hoekstra, Teun. 1984. Krijgen [To get]. In Hans Bennis & Willem van Lessen Kloeke (eds.), Linguistics in the Netherlands , 65–72. Dordrecht: Foris; Landsbergen, Frank. 2009. Cultural evolutionary modeling of patterns in language change: Exercises in evolutionary linguistics . Utrecht: LOT publications) is influenced by prior activation of the regular Dutch passive construction, in which the agent PP can only appear with the preposition door . We indeed find such an effect, and, crucially, this effect is found both when the prime sentence contains a passive construction with an agent PP as well as when the prime is a passive construction without an agent PP, indicating that we are dealing with priming on the constructional level, and not with the mere repetition of a lexical item or an abstract syntactic configuration. Our results provide further evidence for the existence of horizontal links and constructional contamination.

  • Open Access

    Constructions are influenced in their realization by the degree of activation of horizontally linked constructions: using BERT sentence embeddings to test contamination between constructions

    by Gauthier Delaby, Timothy Colleman

    Pijpops and Van de Velde (2016. Constructional contamination: How does it work and how do we measure it? Folia Linguistica 50(2). 543–581) demonstrated that an alternation within a construction may be influenced by a different construction that is coincidentally formally similar through the process of constructional contamination. This mechanism has been explained as a consequence of the storage of unanalyzed chunks produced by a contaminating construction, which are later recycled in a target construction. Consequently, constructional contamination has hitherto been measured by calculating the frequency of individual lexemes (which are assumed to be part of exemplar chunks eligible for recycling) in the contaminating construction. In this paper, we propose a different measurement for constructional contamination. Using the large language model RobBERT, we define “prototype embeddings” for both the target construction and the contaminating construction involved in constructional contamination. For each instance of the target construction, we calculate its similarity to these prototypes. We then show that these similarity variables are significantly correlated with the alternation within the target construction: the more an instance of the target construction is semantically similar to the contaminating construction, the more likely it is to appear with the alternant that is exclusively used in that contaminating construction. Our results are in line with a scenario in which horizontal links, which hold between formally similar constructions, can directly trigger constructional contamination. In addition, we find that the more semantically similar an instance is to the prototype of the target construction itself, the more likely it is to appear with the alternant that is not found in the contaminating construction, supporting a view of the constructional network in which constructions are themselves conceptualized as links rather than nodes.


  • Open Access

    Constructional contamination between two constructions with krijgen ‘to get’ in Dutch

    by Gauthier Delaby, Timothy Colleman

    Two structurally unrelated constructions can affect each other’s realization through a process that has been called constructional contamination. According to this effect, lexemes participating in a grammatical alternation will deviate in their stochastic preference for an alternant if they appear frequently in a contaminating construction that is formally similar to the alternant in question. In the present article, we evaluate whether such contamination effects can also be found between two constructions in Dutch which share the same form, including the verb krijgen ‘to get’ and a past participle, but have distinct meanings (a “receptive” vs. a “resultative” meaning), and, if so, whether these effects occur in both directions. We zoom in on word order differences: if krijgen and the participle appear in a verb cluster, both word orders are possible in the receptive krijgen -construction, while, for the resultative krijgen -construction, the order with krijgen preceding the participle is reported to be ungrammatical in the grammatical literature but is not altogether absent from real-language corpora. Logistic regression analyses, based on data culled from the SoNaR-corpus, show that the word order in both constructions is indeed affected by constructional contamination, thus showing that this phenomenon can be bidirectional. Additionally, we demonstrate that these contamination effects differ between two national varieties of Dutch, viz. Belgian vs. Netherlandic Dutch, and we argue that contamination can sometimes also result in a disambiguating reflex. These results suggest that subtle differences in the organization of the constructional network can result in (partly) different contamination effects.


  • Open Access

    Constructional contamination blocking full-fledged alternation? (The lack of) word order variation in Dutch verb clusters with resultative krijgen ‘to get’

    by Gauthier Delaby, Timothy Colleman, Marithé Buysse

    In Dutch verb clusters consisting of an auxiliary and a past participle, both orders of participle and auxiliary are fine. This distinguishes auxiliaries from copular verbs combined with an adjectival participle, for which only the [Part + V]-order is straightforwardly possible. Reference grammars claim that for the resultative krijgen -construction, the order krijgen ‘to get’ + participle is ungrammatical and that, thus, the construction is adjectival. In this paper we first show that the [Aux + Part]-order is actually used in resultative krijgen -clusters in real-language corpora, albeit infrequently so, and that, in grammaticality judgements, language users find it quite acceptable. More crucially, we present evidence that the order in resultative krijgen -clusters is affected by constructional contamination (Pijpops, Dirk & Freek Van de Velde. 2016. Constructional contamination: How does it work and how do we measure it? Folia Linguistica 50(2). 543–581). On the one hand, other auxiliary constructions with krijgen contaminate the resultative construction, resulting in sporadic occurrences of [Aux + Part]-orders. On the other hand, we show that resultative infinitive clusters, which are historically very strongly attracted to the [Part + Aux]-order, occur significantly more often than with other auxiliary constructions and hence promote the [Part + Aux]-order in non-infinitive clusters, thus preventing a full-fledged alternation. In this way, usage tendencies explain the observed word order facts better than postulated grammatical properties such as the alleged “adjectival” status of resultative krijgen . Additionally, we show that constructional contamination does not only affect language production, but also acceptability judgements.


  • How to Do Things with Corpora: Introduction

    by Jóhanna Barðdal, Gauthier Delaby, Anaïs Vajnovszki, Torsten Leuschner

  • Open Access

    ‘’t is ça va dan he’

    by Gauthier Delaby, Ulrike Vogl

    This article discusses the adjectivization of the Dutch interjections ça va and ok. Lemmens (2013) argues that this adjectivization process was driven by analogical relations based on broad syntactic categories. A corpus investigation and acceptability judgements show (i) that the adjectival use of ça va and ok has become more frequent since Lemmens’s research, but that their adjectivization is not (yet) completed, (ii) that not only abstract generalizations, but also lower-level generalizations and specific constructions play a crucial role in the adjectivization of both interjections, and, (iii) that ça va and ok already as interjections showed limited grammatical behaviour, making them more susceptible to adjectivization. We argue – contra Lemmens (2013) – that not (only) reinterpretation has led to the adjectivization of both interjections, buth rather their lexical and horizontal links in the constructional network.