There are two general views on the connection between language acquisition and language change. Functional usage-based theories usually argue against children as possible agents of phonological and morphosyntactic changes [1]. By contrast, formal rule-based accounts generally assume that incrementation of new variants in linguistic community can be kicked off by the acquisition process [2, 3]. A controversial point is whether or not children converge on the parental grammar after a period of overgeneralisation. While usage-based researchers claim that early innovations are fully adjusted to the adult model before children reach adolescence [4, 5, 6], acquisition-based studies argue that though children do retract after overgeneralisation, this retraction still leaves them ahead of the input, and that it is the uncorrected input-divergences that socially motivated incrementation later draws on [7, 8, 9]. In this study, based on probabilistic language modelling, we argue that language change can be triggered by language learning (which does not preclude the possibility of life-long linguistic innovation).

Using extensive corpus data of spontaneous child speech (CHILDES [10]; 34 children; 6,252 utterances) and a Bayesian model of collostructional analysis, we investigated how children learn construction-verb relations during the acquisition of two English future-tense constructions: GOING.TO and WILL. The model estimated, for an utterance U encountered in child speech at time T, two probabilities that a lexical item N will be used in either of these constructions: one conditional on the current state of the child grammar C and another conditional on the current state of the parental grammar P (inferred from the observed outputs these grammars produced):

(1) P(utterance UT = construction A | word = N, grammar = C{UChild,1,..,UChild,T-1}),
(2) P(utterance UT = construction A | word = N, grammar = P{UParents,1,..,UParents,T-1}).

To account for the process of choosing one between these options, we conceptualised the choice of grammar on which to model the respective utterance as an action and assigned utility values to the outcomes. The logic is that at the earlier stages of language acquisition, children may want to reproduce their parents’ linguistic habits, but in the process of establishing their own collostructional preferences they eventually reach the point where the parental grammar becomes less important. Both probabilities and utilities were estimated from data.

The results reveal that the acquisition of English future tense includes the following developmental stages: (i) children learn which words their parents preferably use in certain constructions and reproduce these patterns, (ii) children introduce some innovations into the acquired system, and (iii) once their own preferences are established, children become less sensitive to the input from child-directed speech; at the same time, as the new system grows more stable and self-contained, further innovations become less likely, albeit not impossible. Crucially, we observe that after the period of overregularization and retraction, the next-generation grammars still remain not fully adjusted to the grammars of the previous generation. The model predicts that as children reach adolescence, there is, on average, a 10 % chance they will select a different alternating construction than their caregivers.

A core tenet of usage-based linguistics is its commitment to analyzing concrete instances of language use. Novel technologies, including NLP tools and large language models, can assist linguists with this (Torrent et al., 2023). While the Universal Dependencies project provides a robust cross-linguistic framework for syntactic/semantic annotation (DeMarneffe et al., 2021), annotated spoken data, particularly child-directed and child-produced speech, is still scarce. Recent work has begun to address this gap: Yang et al. (2025) release UD annotations for a large portion of English CHILDES data. We present a novel tagger for utterance-level constructions based on a novel, self-trained dependency parser for CHILDES data, trained with Stanza (Qi et al., 2020) and based on RoBERTa embeddings (Liu et al., 2019). Our goal is to automatically classify child and child-directed utterances into a holistically applicable construction taxonomy, following the classification schemata developed in previous studies on language development (Cameron-Faulkner et al., 2003; Cameron-Faulkner and Hickey, 2011;Cameron-Faulkner and Noble, 2013; Noble et al., 2018; Bunzeck and Diessel, 2025). In line with these earlier studies, we adopt a taxonomy of nine utterance-level construction categories: fixed formulaic expressions, fragments without finite verbs, wh-questions, yes/no questions, copula constructions, imperatives, subject-predicate intransitive, subject-predicate transitive, and complex multi-clausal sentences. Our tagger uses the POS-tags and UD-annotations (provided by the CHILDES-based UD parser) as the base of a decision tree-like classification procedure. We evaluate our tagger on manually annotated test data from the Helen dataset of the MPI-EVA-Manchester corpus (Lieven et al., 2009). Using our CHILDES-trained model yields the highest accuracy at 91.71%, substantially outperforming both a POS-tag-only baseline (87.54%) modeled after the parser by Bunzeck and Diessel (2025), and a sentence embedding model (Reimers and Gurevych, 2019) with an MLP classifier (81.50%). These results demonstrate that dependency information is an extremely valuable source of information for distinguishing utterance-level constructions. In light of these improvements in classification accuracy, our tagger also enables analyses that previously remained out of scope. For example, the studies that our construction taxonomy was lifted from(Cameron-Faulkner et al., 2003; Bunzeck and Diessel, 2025) exclusively focus on child-directed speech. As our tagger also performs well on child speech, it can be used to map out early development more holistically. Whereas CDS results from our analysis align with the studies mentioned earlier, the results for child speech deviate from them quite tremendously, showing considerably more fragments, but less questions and complex utterances. Although the results remain a proof-of-concept, they clearly show that taggers based on UD-annotations specifically created with CHILDES data in mind outperform other automatic classification approaches and offer a novel, robust way of analysing large amounts of developmental data.

Mental verbs are generally understood as verbs referring to perception, emotion, and cognition. Nonetheless, the definition of mental verbs, the labels used for their categories, and the identification of their constituent elements can vary considerably across the international literature (cf. Croft 1993; Biber, Conrad & Leech 2002; Israel 2008) and the Hungarian literature (cf. Tolcsvai Nagy 2017; Kugler 2024). In Hungarian cognitive semantic research, mental verbs are commonly discussed under various labels, whereas Croft’s (1993) tripartite classification is rarely applied in full. Although text-based analyses of mental verbs are available (cf. Pavlenko & Suetina 2024; Heller 2024), there is still a need for a corpus-based study addressing the specific semantic properties of these verbs.
This presentation focuses on a segment of my ongoing doctoral research on Hungarian mental verbs, conducted on the Corpus of Hungarian Lyrical Poetry (Horváth et al. 2025). Although the corpus is relatively small, comprising approximately 170, 000 tokens, the inclusion of fictional apostrophe (cf. Tátrai 2024) ensures a significant representation of mental verbs. Using MAXQDA software, the study aims to (i) evaluate the efficiency of Croft’s tripartite classification model; (ii) explore the possible semantic network of mental verbs in relation to other verb categories; and (iii) compile items that exemplify inchoative mental verbs (cf. Croft 1993).
For this study, the annotation of mental verbs was carried out on the Corpus of Hungarian Lyrical Poetry by taking into account Croft’s categories, the contextual meaning of each verb, and, in ambiguous cases, definitions from the Comprehensive Dictionary of Hungarian and the Hungarian Dictionary. The annotation process involved two annotators. The annotation system was based on preliminary annotations and used Croft’s tripartite categories, the inchoative category, and an additional “other” category for ambiguous items. Approximately, counting with duplums, 5900 mental verbs were annotated and around 6500 counting the perception, cognition, and emotion.
The findings indicate that while the prototypical categories of perception, emotion, and cognition include easily identifiable elements, peripheral items are not readily associated with the primary categories. Instead, they tend to form networks with other semantic verb categories, such as verb.motion and verb.communication (cf. WordNet, Fellbaum 1998). This observation is further supported by the annotators’ comments on additional categories, which reflect a centre-periphery dynamic within the pragmatic structure of these verbs.
By examining the semantic properties of Hungarian mental verbs and gathering verbs relevant to this category, this work contributes to a re-evaluation of how mental verbs are classified and conceptualized in Hungarian scholarship.

Language contact has been studied from a wide variety of angles, but there has been little contact between researchers working on synchronic speech phenomena versus diachronic contact-induced change, and lexical versus grammatical effects. In this colloquium we hope to show how different approaches can be brought together under a usage-based approach (Backus, 2020; Langacker, 1986). In this paper I will first review recent insights into usage-based approaches to language contact. Then I hope to show how a study of switches within multiword units, in particular mixed compounds, can provide new insights into the interface between syntax and the lexicon.
A key question for researchers interested in compounding is what determines the headedness of a mixed compound: is it a purely syntactic phenomenon, in that the language of the head of the compound determines its internal structure, or do other variables play a role too? Central to the analyses proposed are corpus linguistic approaches, which look at the strength of the association between the two parts of a multiword unit, and the language of the clause in which the compound appears. Furthermore, I hope to show that entrenchment and the variability in the speech community, as well as the degree of language contact between the languages plays an important role.

Backus, A. (2020). Usage-based approaches. In Adamou, E. & Matras, Y. (eds.), The Routledge Handbook of Language Contact. London/New York: Routledge. 110-126.

Langacker, R. W. (1987). Foundations of cognitive grammar: Theoretical prerequisites. Stanford, CA: Stanford University Press.

Treffers-Daller, J., Yap, N. & Majid, S. (2025). Pizza ice cream with salami sprinkles: an analysis of the creativity of nominal compounding in multilinguals. To appear in the International Journal of Bilingualism. https://doi.org/10.1177/13670069251400978

In this paper, we investigate the emergence of bilingual patterns in contact between structurally similar languages, namely, Italian and Peruvian Spanish.
Following a DCxG perspective (Boas & Höder 2018, 2021, 2025), code-mixing data can be portrayed as a combination of language-specific idioconstructions (involving lexicon and grammar from one single linguistic repertoire), and diaconstructions, that is, language-unspecific abstract patterns that rely on constructional similarity and may be filled with linguistic material from multiple lexical repertoires.

no sì sardegna è bello #HE $sent-IDO che questo mare è es- (.) bellissimo
no yes Sardinia is beautiful I HAVE heard that this sea is beautiful

o altro #LUGARES, no. #_NO HEMOS PODIDO:: (.) visitare.
or other PLACES, no WE COULD NOT visit

The past tense and modal constructions (1)-(2) are a case in point for a diaconstructional interpretation, as they present a pattern shared by the two languages which is filled by lexical entries from both.
In this study, we aim to quantitatively assess to what extent such constructions are emergent vs entrenched in cognition and conventionalised in usage, and whether there are differences based on the speakers’ variable exposure to Italian (in terms of years spent in Italy).
We rely on data from the StraParla-TO corpus, involving 16h 53’ of recordings from 41 speakers who have been living in Italy for a variable amount of time (a few months to >20 years) or are Italian-born (6/41).
A frequency-based approach to code-mixing data will allow us to explore how similarity between languages favours the emergence in usage of hybrid constructions (Barking et al. 2022) and whether this may lead to the conventionalisation and entrenchment of multilingual diasystematic patterns at a more abstract level. At the same time, accounting for inter-speaker variation will enable us to locate these diaconstructions on a continuum between the individual’s and the community-shared mental constructicon.

References

Barking, M., Backus, A., & Mos, M. (2022). Similarity in Language Transfer – Investigating Transfer of Light Verb Constructions From Dutch to German. Journal of Language Contact, 15(1), 198–239. https://doi.org/10.1163/19552629-15010005
Boas, H. C., & Höder, S. (Eds). (2018). Constructions in Contact: Constructional perspectives on contact phenomena in Germanic languages. John Benjamins. https://doi.org/10.1075/cal.24
Boas, H. C., & Höder, S. (Eds). (2021). Constructions in Contact 2: Language change, multilingual practices, and additional language acquisition. John Benjamins. https://doi.org/10.1075/cal.30
Boas, H. C., & Höder, S. (Eds). (2025). Constructions in Contact 3: Constructional schemas and patterns in language contact. John Benjamins. https://doi.org/10.1075/cal.40

Linguistic creativity has received considerable attention in constructionist research (see, e.g., Hartmann and Ungerer 2023; Hoffmann 2024; Norde & Trousdale 2024; Turner 2020). Now, it is well established that individuals differ greatly with respect to their creative ability (Kandler et al. 2016). Yet, so far, constructionist studies ignore the creativity of individuals (an exception being Van den Stock et al. 2024), while in psychology, most verbal creativity tasks only test single words. In the Divergent Association Task (DAT) (Olson et al. 2021), e.g., subjects are asked to name 10 nouns “that are as different from each other as possible, in all meanings and uses”.

To also test the constructional creativity of individuals, we developed a constructional version of the DAT, the Constructional Divergent Association Task (CXN-DAT). In the CXN-DAT, subjects are given the following four constructional contexts:

(1) Yesterday, the man ______ the woman. (Transitive cxn)
(2) Yesterday, the woman _____________ in her room. (Intransitive cxn)
(3) Yesterday, the woman _____________ the hell out of the man. (V the hell out of cxn)
(4) Yesterday, the man ______ his way to the top.(Way cxn)

In the test, the stimuli are counterbalanced so that the order of (1)–(4) is varied and that man and woman appear as the subject (and object) slot of all constructions (to minimise gender stereotype effects). We tested the CXN-DAT in a pilot study by recruiting 55 native speakers of British English (18–60 years | 37 females, 18 males). Subjects first completed a short BIG5 personality questionnaire (Lang et al. 2011) and then DAT and CXN-DAT. As in the DAT, we used GLoVe (Penningston et al. 2014) to get a semantic vector for each verb response. DAT scores were calculated for each subject and construction by averaging all pairwise cosine distances and multiplying the result by 100 (Olson et al. 2021). A mixed effects linear model with PARTICIPANT as a random factor revealed only that CXN-DAT scores are significantly lower than the DAT scores (see Figure 1 below): [see pdf version of abstract] Since constructions constrain the search space for subjects, this effect can easily be explained. More interesting is the fact that we observe evidence for coverage effects (Goldberg 2019). For the V the hell out of cxn, e.g., all speakers drew on the two most frequent verbs (beat and scare) as well as other verbs from their respective semantic frames (e.g. hit, slap, kick / annoy, frighten, worry; Perek 2016; Hoffmann 2021). Finally, using the AI-based scoring tool OCSAI (Organisciak et al. 2023) we obtained originality ratings for all responses. A linear regression model regressing originality against the order in which items were produced in a series, separately applied for each participant and construction, revealed a linear increase of originality, t(54) = 7.45, p < .001. Across participants, the mean slope of this increase correlated significantly with the openness score of the BIG5 personality inventory, r = .280, p = .041, after controlling for outliers. This suggests that the CXN-DAT successfully detects inter-individual creativity differences.

Extravagance is a pragmatic phenomenon related to Keller's (2014 [1990]) maxims of action, which Haspelmath (1999: 1055) formulated as 'talk in such a way that you are noticed'. Given the increasing popularity of pragmatics within usage-based, constructionist approaches, it is unsurprising that interest in extravagance has also grown in recent studies (e.g. Baumann & Mühlenbernd 2022; Hartmann & Ungerer 2023; Norde & Trousdale 2025; Ungerer & Hartmann 2024), particularly in studies of language change from a Diachronic Construction Grammar perspective (e.g. Eitelmann & Haumann 2022; Norde & Trousdale 2023; Trousdale & Norde 2025). Nevertheless, the term 'extravagance' is open to interpretation and strikingly underoperationalised (Ungerer & Hartmann 2020). Whilst current research focuses on diachronic aspects, the present project aims to rethink extravagance as a synchronic phenomenon influenced by structural, cognitive and communicative factors. This reconceptualisation requires us to consider the complexity of the mechanisms studied and address the empirical implications by employing a combination of methods focusing on these structural, cognitive and communicative features. This approach enables us to capture the synchronic range of extravagance and may also assist in the comprehension of diachronic processes. The talk presents the current state of the project by proposing a usage-based construction grammar model of synchronic extravagance which aims to identify more feasible empirical approaches to extravagance.

One major phenomenon in construction grammar is coercion, an event in which constructional meaning overrides denotational meaning (cf. Taylor 2012, Hilpert 2014, Goldberg 2019). This phenomenon can inform other branches of linguistics, such as im/politeness studies, where there has been debate regarding the notion of impoliteness being inherent in linguistic forms. Most arguments seem to oppose the view that impoliteness can be inherent in linguistic features, citing the importance of contextual factors (e.g., Fraser & Nolen 1981, Locher & Watts 2005). However, there is evidence that some expressions have a stable association with impoliteness effects – i.e., they are entrenched – and may cause negative evaluation in the hearer, irrespective of a particular social context (e.g., Culpeper’s 2011 conventionalised impoliteness formulae). We propose that particular constructions, specifically the construction embedded in the example You are [such a [noun]], and its equivalents in German and Dutch, may carry relatively stable associations with negative evaluative meaning. We thus apply a construction grammar approach to an im/politeness research question.

When we encounter an example of You are such a [X], the noun X is very likely an epithet (see, e.g., Bolinger 1972, Taylor 2012). This is supported by corpus evidence, for instance, in the Corpus of Contemporary American English (cf. Queisser 2024). Taylor (2012) identifies distributional peculiarities of epithets and points out that the sentence You are such an idiot is acceptable, while You are such an architect is marked. As constructions may cause coercion effects, the neutral noun architect may even acquire a negative connotation due to its occurrence in this particular construction. We argue that this is due to the particle such occurring in the form, and propose an epithet construction with the formal representation of [such + indefinite article + nominal], which occurs predominantly, but not exclusively, in predicative contexts in which the subject is a personal pronoun: [PRON BE such ART N].

To test our theory, we conducted a series of rating surveys in English and German, investigating coercion effects for neutral nouns like architect occurring in the construction (cf. Queisser & Pleyer 2025). We found statistically highly significant evidence in support of our theory. In all studies, informants rated example sentences overall more negatively when they contained such (so in the German condition). Interestingly, an increase in positive ratings of individual neutral nouns was also observed, suggesting an interaction between lexical and constructional meaning as well as a broader evaluative effect. The studies were conducted with first, second, and third person pronouns in the subject slot, and ratings also varied according to the subject pronoun used. For this talk, we will extend our focus and conduct additional studies in Dutch. The construction has already been described (see Ghesquière & Van de Velde 2011). We hypothesize that Dutch zulk/zo’n will show coercion effects similar to those found for English such and German so. As the observed effect was much stronger for German than for English, further exploring the phenomenon across language boundaries is a logical next step.

In this introductory talk, the theme session organizers will offer a theoretical and methodological outline of the Columbia School (CS) framework (Diver 1995, 2012; Huffman 2001, 2006; Stern 2019), as well as highlight key points of contact and divergence between core CS tenets and the related approach of ethnopragmatics (García 1995; Martínez 2004, 2009; Martínez y Speranza 2009). CS is a functionally-oriented school of linguistics that was initiated by William Diver, Erica García and their students in the 1960s. It is based on two fundamental understandings:
● that language is an instance of human behavior, so it can be “expected to share characteristics that are found in other aspects of human activity and cognition” (Stern 2019: 3);
● that language is an instrument inherently designed for communication, so “it must be the case that its fundamental structural components be signals and meanings; that is, that language share in the characteristic that is common to all instruments of communication” (Diver 2012: 47).
Originally born as a reaction to the empirical failure of formalist approaches, the aim of CS is to discover how language is structured based on facts of actual usage. It adopts an inductive, corpus-driven approach with the objective of explaining what features of the linguistic system lead speakers/writers to deploy language in the ways they do.
Closely connected to CS, García’s later innovation, ethnopragmatics, emerged from an interest to “explain the cognitive processes that guide speakers in the solution of their communicative needs” (Martínez, 2009: 261, our translation). This approach thus goes beyond establishing the identity of linguistic signals, seeking to uncover the cognitive profilings that guide their variable use in different contexts, with a special interest in “cultural aspects as motivators of communicative needs” (Martínez 2004: 1, our translation). The following distinctive features are common to CS and ethnopragmatics:
● an “aposterioristic” (Huffman 2006: 44) approach to linguistic analysis: no categories are assumed, but rather hypotheses are proposed and tested based on empirical data;
● an understanding that linguistic structure is not autonomous, but invariably motivated by the communicative needs and the physical, cognitive, and cultural traits of human beings;
● an instrumental view of meaning: it is recognized that language is not compositional; rather, linguistic inputs serve as mere hints from which messages can be contextually inferred;
● a reliance on naturally occurring discourse as a basis for linguistic analysis;
● a qualitative-quantitative methodology: close textual analysis is combined with testing of predictions regarding “co-occurrence favorings” (Sabar 2018: 95, 132) of the item(s) under study in specific contexts.
The papers in this theme session will illustrate the current vitality of CS, and its potential to shed light on different aspects of human language and cognition. These works offer valuable contributions to a variety of fields including grammar, lexical semantics, phonology, (ethno)pragmatics, variation, dialectal studies, language contact, translation, discourse analysis, and endangered languages.

Van Soeren (2023) analyses the distribution of consonants in lexical corpora of Spanish CVCVCV forms (e.g. pájaro ‘bird’), and English, and Dutch CVCVC forms. Figure 1 shows the results for the Spanish corpus (but highly similar results were found for English and Dutch); labial consonants are favoured in position C1 (consonant 1) and apicals are favoured in C2 and C3. Moreover, the proportion of labials, but also palatals and velars, diminishes from the beginning to the end of the word. The reverse pattern is observed for apicals: the smallest proportion in C1, a larger proportion in C2 and the largest proportion in C3.

Van Soeren (2023) explains the observation with communicative load (Diver 1979/2012; Tobin 1997); as a word is uttered, every sound/segment eliminates more word candidates, and therefore the beginning has a higher communicative load than the end, which is more redundant for word identification (cf. Grosjean 1980; King & Wedel 2020). At the beginning of a form, there is thus a high burden of distinctiveness. Labials have a perceptual advantage over non-labials; they are visible as well as audible, which explains their favouring in a position with high communicative load. At the more redundant end, the need for optimal distinctiveness is lessened, and articulatory ease is favoured. This is where Van Soeren (2023) observes a favouring of apicals, which is explained with their articulatory efficiency.

It is not surprising that similar results were found for Spanish, English and Dutch, as two of the three languages are from the same language family, and all three are Indo-European. Therefore, I have compared these results to the distribution of labials, apicals, palatals and velars in Guaraní, a polysynthetic Tupi language. As this language has few monosyllabic CVCVCV words, a lexical corpus of CVCV forms was collected. In general, I expected to find similar results; one can argue that an explanation in terms of communicative load applies to any language, as all languages have forms with a beginning and an end, where the end will always be more redundant for its identification.

Figure 2 shows that in Guaraní CVCV forms, we indeed observe the same favouring of labials in C1, and a smaller proportion of labials observed in C2. The reverse pattern is again observed for apicals; we observe a small proportion of apicals in C1, and a clear favouring in C2. Again, the proportion of palatals and velars diminishes from C1 to C2 position.

A curious detail is that Guaraní /ʋ/ does not pattern with the other labials (although it is included in the “labial” category in Figure 2). Further research will explore possible explanations as to why /ʋ/ should be favoured in C2, rather than C1.

The talk discusses the meaning of a form called L-participle in the second edition of the Polish novel Król Obojga Sycylii by A. Kuśniewicz (Kuśniewicz, 1972), published in 1972. The traditional definition of L-participle is the following: a common Slavic verbal adjective, formed from a main stem of the verb and a suffix -l- that tends to appear in the contexts that denote past (Lunt, 2001, p.109). The contextual senses of L-participle in Slavic languages in general have been a subject of discussion in the last decades (for the comprehensive overview, see (Bláha, 2020), especially its usage in the context not related to the past (Benić, 2024), for the comprehensive overview see (Mirić, 2020). However, even among the works that made attempts at semantic analysis of L-participle (Proeme, 1991), there were none that entertained monosemic hypotheses, the ones that would explain the distribution of the L-participle across all of the contexts, and not only subset of them. This is especially relevant for L-participle in Polish, which is used in a variety of situations, including past, present and future contexts. The talk is an attempt at explaining this, presenting a monosemic account of an L-participle in A. Kuśniewicz’s Król Obojga Sycylii.
The research brings evidence in favor of hypothesis that L-participle, at least in KOS2, has a single meaning PAST, whether it is used for characterisation of the events in the past or in the future from the perspective of the speaker. The key methodology used is a sign-based aposterioristic analysis, performed in the spirit of the Columbia School of Linguistics (CSL), shaped by W. Diver and É. García (García & Otheguy, 1983; Huffman, 2002; Diver, 2012). The research analyses all of the contexts where L-participle is present. It pays special attention to the segments of the text that contain the lowest and the highest number of L-participles. The careful study of these fragments shows that its meaning is indeed PAST. This has significant consequences, for instance, for the messages that describe past events. When L-participle is almost absent (and, as a contrast, utterances with traditional present tense forms or without any verbs are on the rise), the narrative tends to be the most immersive, while when L-participle is heavily present, the narrative becomes report on the facts that happened (see examples in appendix A). The talk showcases three episodes of the novel, two descriptions of battles and a single letter (all being a narrative within a narrative), to bring the arguments in favour of the hypothesis.
Overall, the presented study provides additional evidence to the earlier suggestions of the Polish verb system being an intricate interlock of various grammatical systems, used with a high degree of variety by different speakers (Proeme, 1991). The research also serves as an initial point in the semantic exploration of L-participle in the various Slavic varieties.

This paper adopts Diachronic Construction Grammar (Traugott & Trousdale, 2013; Barðdal & Gildea, 2015; Smirnova and Sommerer, 2020) and Cognitive Historical Sociolinguistics (Swiggers, 2022; Kristiansen et al. 2021; Claes, 2018) to analyze the Scottish be-perfect as a linguistically and cognitively motivated construction, moving beyond traditional grammaticalization accounts (Mustanoja, 1960; Denison, 1993). Using corpus data from the Scottish Pauper Petition Corpus (ScotPP) (Gordon & Groot, 2022-), this paper demonstrates how constructional competition, language usage and social stratification may jointly shape diachronic change in temporal periphrastic structures.

It will be argued that the be-perfect and have-perfect are two distinct, diachronically competing constructions (Goldberg, 1995; 2006; 2019) that require independent representation in the constructicon, under an overarching [AUX]-perfect construction. The be-perfect construes perfectivity through a CHANGE OF STATE image schema, linking it cognitively to motion and intransitive verbs (Kytö 1992, 1997). This semantic-constructional motivation explains the observed verb-class restrictions: speakers recruit the be-perfect (1) for mutative predicates where the endpoint state is profiled, while the have-perfect possessive-achievement frame suits transitive, agentive contexts. (2)

(1) …but now weakneʃs and infermity with old age are arriaved... [Perthshire, 1837]

(2) That your Petitioner would {therefore} humbly{im}plore to your Honours an offer of what fishes I have caugth ſince the ſaid first day of may last… [Aberdeenshire, 1767]

ScotPP uniquely preserves authentic, unmediated language from lower-class populations, a demographic usually excluded from diachronic linguistic study due to manuscript scarcity (Groot, 2024). Social stratification in the corpus reveals differential usage patterns. By using a newly-created corpus for the first time, the paper will integrate cognitive sociolinguistic principles into the historical analysis. Low-literacy speakers show significantly higher be-perfect frequency, reflecting conservative preservation of this construction despite standardization pressures favoring have-perfect (Traugott, 1992; Tagliamonte, 2000). Rather than treating social differences as surface variations, I argue they reflect genuine differences in how speakers cognitively organized the same construction-semantic space through exposure to different exemplar distributions. This usage-based, cognitive sociolinguistic analysis, grounded in authentic low-class correspondence, suggests a methodology for accessing historical speakers’ mental grammars.

By bridging construction-based and cognitive sociolinguistic perspectives to bear on diachronic corpus data from historically marginalized and underrepresented speakers, this paper illustrates 1) how, in historical change, constructional competition can be resolved through usage frequencies, 2) social stratification reflects differential constructional usage across communities and 3) how constructions evolve as cognitively grounded patterns shaped by speakers’ embodied experience in social contexts.

This paper investigates competition among quantifier constructions in Modern English. Specifically, it will trace the diachronic trajectory of the long-established monomorphemic quantifiers (many and much) in contrast to some of their younger grammaticalized competitors, namely the complex binominal expressions a lot of/lot(s) of and a bunch of/bunch(es) of (Langacker 1991, 2008; Huddleston & Pullum 2002; Raden & Dirven 2007; Biber et al. 2021; Brems 2011). Examples (1-3) illustrate the transition from literal to quantificational/grammaticalized usage of one of the binominal quantifiers, while example (4) depicts one of the default monomorphemic expressions:
(1) At the highest grows a bunch of flowers, clustering together like grapes. (EEBO 1578, Text_ID: A20579)
(2) With great care they do pluck out, only keep a bunch of hair on the crown of their heads. (EEBO 1598, Text_ID: A05569)
(3) That said Desires were Dropsies, bid the Mind take a Bunch of Reason, that grows in a well-temper 'd Brain. (EEBO 1671, Text_ID: A53048)
(4) So suddenly, that I had no time to read many books written on American subjects. (ARCHER 1787, Text_ID: 1787mark_f4a)
Using data from Early English Books Online (EEBO), A Representative Corpus of Historical English Registers (ARCHER), and the British National Corpus (BNC), the corpus study will examine the four quantifiers’ distribution of functional labor within the quantificational domain and reconfiguration of the noun phrase (NP), covering Early to Late Modern English and a wide range of genres. The analysis combines descriptive statistics with quantitative modeling, including distributional analyses and logistic regression, to trace frequency developments, genre sensitivity, and semantic domain constraints underlying quantifier choice (Hilpert 2013; Gries & Stefanowitsch 2004; Bohmann & Sommerer 2025). Accordingly, I will be addressing the following interrelated questions:
(i) how the relative frequencies of monomorphemic and binominal quantifiers change over time;
(ii) which linguistic factors (e.g., genre, period, and properties of the NP) best predict quantifier choice; and
(iii) whether competition results in specialized profiles rather than functional replacement.
Preliminary results indicate that binominal quantifiers develop distinct distributional profiles, showing increased preference for informal and narrative registers and greater flexibility in NP-internal contexts, while monomorphemic quantifiers retain more conservative usage patterns.
To capture and model this competition, my paper draws insights from Diachronic Construction Grammar (DCxG) and approaches quantified NPs as an evolving network of quantificational constructions, in which newly grammaticalized binominal expressions (e.g., a lot of) enter into competition with canonical monomorphemic quantifiers. It models change as the constructionalization of new semi-specific templates and shifts in productivity (Goldberg 2006; Bybee 2010; Traugott & Trousdale 2013; Hilpert 2013), while remaining compatible with broader linguistic approaches to grammaticalization/lexicalization and variation. Ultimately, I will postulate that quantifier competition reflects a reconfiguration of the constructional network of English quantifiers and newly developing quantificational coding strategies within the NP, driven by frequency effects, analogical extension, and contextual sensitivity.
The study shows how competition among alternative quantifying expressions can motivate the development of distinct functional niches, suggesting gradual restructuring within a grammatical domain.

This paper investigates the diachronic trajectory of the English prepositions over and under and the constructions (PPs) headed by these prepositions throughout Early Modern English (1450 – 1700).
Some diachronic studies have looked at prepositions viewing them as relatively stable grammatical items of minimal change. In cognitive linguistics, however, it is well known that they are highly polysemous form-meaning pairings which get recruited into many different constructional schemas (Tyler & Evans 2003; Guarddon-Anelo 2006; Lindstromberg 2010), with usage preferences, constructional complexity and semantics often changing historically (Kovbasko 2022) making them a premier target for diachronic studies. In this paper, particular attention will be given to their metaphorization trajectories, changes in frequency (Traugott & Trousdale 2013; Hilpert 2013; Perek 2016) and the formal complexity of the PPs. Diachronic Construction Grammar (DCxG) principles and Cognitive Linguistics concepts, such as image schemas, polysemy, and the metaphorical extension of SPACE to conceptualise non-spatial relations (Lakoff & Johnson 1980) will be employed as a baseline theoretical analytical framework.
Over and under were chosen because they share similar (PATH/SOURCE-PATH-GOAL, UP-DOWN) and complementary (ENABLEMENT, RESTRAINT, DIVERSION) image schemas (Lindstromberg 2010; Hampe 2005), while also being part of a contrast set (Tyler & Evans 2003), with similar usage in figurative (ex. 1, 3) and literal (ex. 2, 4) constructions:

(1) Witness the Emperor Caligula, who so demeaned himself in his Empire and tyranny over others… (EEBOv3: 1657, A47013)

(2) … Mustaffa commands all his host to pass over the River Canac (EEBOv3: 1595, A07559)

(3) The last sort of Fanatioism among them, resisting authority under pretence of Religion (EEBOv3: 1671, A61540)

(4) …and that the Army was encamped under the Walls of the Town… (EEBOv3: 1698, A55335)

Note that the study will look solely at prepositional phrases headed by over or under and with NPs headed by proper or common nouns (e.g., over the river).
The paper will focus on the following research questions:

RQ1: How do the variables TIME (HALF-CENTURY) and GENRE (NONFICTION, LETTERS, LAW TEXTS, RELIGIOUS TEXTS, etc.) affect the complexity of the PP in the investigated period?
RQ2: How do the prepositions’ semantic profiles differ and which other factors (e.g., lexical biases, and structure of the NP) increase the complexity of the PP?

Empirically, the paper will use data from the Early English Books Online corpus (EEBOv3) and the ARCHER corpus (cqpweb.lancs.ac.uk, Hardie 2012).
COMPLEXITY is defined as the number of words in the PP but more variables (e.g., presence or absence of determination and modification) will also be taken into account.
The collected data will be analysed by running a logistic regression model using TIME, GENRE and COMPLEXITY as variables (Bohmann 2016; Levshina 2020; Bohmann & Sommerer 2025). Among other things, it will be tested if a) semantics and function (i.e., semantic roles of the PP; noun classes in the NP) influence the PP’s complexity but also b) how formal complexity, modification and determination of the NP influence the semantic meaning of the PP.

metaphors, euphemisms, dysphemisms, staged controversy, political ciscourse, Trump, Selensky, Weidel

Aims of the study

As a core concept in Cognitive Linguistics, construal refers to the way a speaker conceptualizes and describes a situation (Croft & Cruse, 2004). This study focuses on the conceptualization of events (or more precisely, eventualities, an umbrella term for events, processes and states proposed by Bach [1981]), which can be construed as agentive or non-agentive, whereby agents are understood as volitional participants and/or causes (cf. Dowty, 1991). Moreover, the agents can be explicitly expressed or omitted due to demotion strategies or clause deranking. Previous research provides examples of how speakers make systematic use of such different coding strategies: for instance, Henley, Miller & Beazley (1995) show that passive language in news media is significantly more frequent with verbs describing sexual violence compared to control verbs. Importantly, the cognitive processing of events can be affected by such linguistic choices (Faucey & Boroditsky, 2010). The use of different construals is particularly relevant in political discourse, where political actors’ conceptualizations of events play a central role in persuasion (Hart, 2014). Our study aims to compare agentivity construals and expression of agents in political X (formerly Twitter) posts by the five major political parties in the German federal parliament (Bundestag).

Data and method

Our data covers the period from January 2025 to July 2025 which includes the months before and after the 2025 German federal election and therefore marks an eventful political time for the country. We collected 70 posts per party. Events and agents are complex categories that have a prototype structure (Lakoff, 1977). Using detailed coding guidelines (Author, 2025), which account for both core and more peripheral cases, we identified all events in the posts, which were then coded for whether they were agentive and whether the agent was explicitly mentioned. In addition, we also annotated the syntactic constructions expressing the events, as well as the tense, agent’s person, number, and other properties. To answer the research questions, we fitted mixed-effects logistic regression models with parties as fixed effects and tweets as random intercepts. Agentivity and the agent’s expression were treated as the response variables. Other contextual variables were controlled for. To capture event semantics of single-word events, we employed fastText word embeddings (Grave et al., 2018) and analysed their association with agentivity by using the embeddings as predictors in a conditional inference tree. The distributional information was also used to cluster events to facilitate the interpretation of the results.

Preliminary results

Our preliminary analyses show significant differences between parties. In particular, tweets by the Christian Democrats (CDU) contain more agentive events than those of the Green Party (die Grünen), followed by the Left (die Linke). At the same time, the agentive events in the posts by the Alternative for Germany (AfD) have the highest ratio of explicit agents compared to the other parties. To interpret these results, we examine the specific constructions used in the tweets, and compare the semantic classes of agentive and non-agentive events and their rhetorical function in the political discourse of each party.

Abstract: An der Schnittstelle von Konstruktionsgrammatik und Phraseologie untersucht dieser Beitrag geflügelte Worte aus der Politik bzw. weit verbreitete politische Slogans, die sich zu produktiven Patterns für die Produktion neuer Instanzen entwickelt haben.
Im Bereich der Politolinguistik wurde das Thema politischer Slogans bereits mehrfach behandelt (Elspaß 2007). Dabei herrscht kein Konsens darüber, ob Slogans als formulaic text patterns, Phrasemkonstruktionen, geflügelte Worte oder – im Sinne der kognitiven Linguistik – als Frames aufzufassen sind (Behrens 2001; Lüger 1999; Klein 1998; Steyer 2014). Ausgehend von der theoretischen Unterscheidung zwischen Mustern, deren lexikalische Leerstellen regelmäßig und variabel neu besetzt werden, und Phrasemen, die lediglich okkasionell modifiziert werden, soll die Produktivität politischer Slogans anhand der Slotfiller korpusgestützt (SketchEngine) und kontrastiv (Englisch, Deutsch, Spanisch, Italienisch) analysiert werden. Die zentrale Frage aus theoretischer Perspektive ist, welche Faktoren (z.B. Gebrauchsfrequenz, Salienz, Variabilität, Extravaganz) den Übergang von festen Phrasemen zu abstrakten, produktiven Konstruktionen begünstigen, während der Fokus der vergleichenden Analyse auf der Veränderung der semantischen Bedingungen sowie auf den unterschiedlichen Gebrauchskontexten liegt.
Als Fallstudie dient die Konstruktion [make X great again]. Die Korpusanalyse zeigt, dass die Konstruktion hochgradig produktiv ist, eine Vielzahl von Slotfillern aufweist und die Bildung strukturell sprachspezifischer Patterns getriggert hat (z. B. dt. [mach X wieder Y] vs. it. [rendiamo di nuovo grande X] und sp. [que X vuelva a ser Y]). Zudem lassen sich basierend auf dem englischen Schema auch kreativere Instanziierungen wie make America lasagna-addicted again, sowie die Entstehung weiterer Mikrokonstruktionen feststellen, etwa in Form von Substantivierungen wie z. B. [V el make great again], favorece el make great again, [S del make great again], nel celebre discorso inaugurale che sancì la politica del Make America Great Again; [Das Make America Great Again V], Das „Make America Great Again“ trifft einen Nerv.
Schlagwörter: Konstruktionsgrammatik, Slogans, Produktivität, Modifikation, Phraseologie.

Bibliografie
Behrens, Gerold & Esch, Franz-Rudolf & Leischner, Erika & Neumaier, Maria (Hg.). 2001. Gabler Lexikon Werbung. Wiesbaden: Gabler.
Bergs, Alexander. 2019. "What, If Anything, Is Linguistic Creativity?". Gestalt Theory 41 (2), 173–183.
Elspaß, Stephan. 2007. "Phrasemes in political speech". In: Burger, Harald & Dobrovol’skij, Dmitrij & Kühn, Peter & Norrick, Neal R. (Hg.). Phraseologie. Ein internationales Handbuch der zeitgenössischen Forschung. Berlin: De Gruyter, 284-292.
Hartmann, Stefan & Ungerer, Tobias. 2024. "Attack of the snowclones: A corpus-based analysis of extravagant formulaic patterns". Journal of Linguistics 60 (3), 599–634.
Klein, Josef. 1998. "Politische Kommunikation als Sprachstrategie". In: Jarren, Otfried et al. (Hg.). Politische Kommunikation in der demokratischen Gesellschaft. Ein Handbuch mit Lexikonteil. Wiesbaden: VS-Verlag, 376-395.
Lasch, Alexander & Alexander Ziem. 2014. Grammatik als Netzwerk von Konstruktionen: Sprachwissen im Fokus der Konstruktionsgrammatik. Berlin: De Gruyter, 3-13.
Lüger, Heinz-Helmut. 1999. Satzwertige Phraseologismen. Eine pragmalinguistische Untersuchung. Wien: Praesens.
Steyer, Kathrin. 2014. "Zitierkarrieren: Vom Vergehen und Überleben markanter Sätze". In: Grandl, Christian & McKenna, Kevin J. (Hg.). Bis dat, qui cito dat. Gegengabe in Paremiology, Folklore, Language, and Literature. Honoring Wolfgang Mieder on his seventieth birthday. Frankfurt am Main: Lang, 585-592.

Maria Fritzsche & Elias Schmitt & Jonas Trochemowitz:
Perspectives at the Interface of Cognitive Linguistics and the Linguistics of Religion. New Spiritualities, Expressions and Methods
As religion and spirituality encapsulate various cognitive and cultural aspects, research by the Linguistics of Religion can provide a better understanding of the intersections of Cognitive and Cultural Linguistics. In order to outline this perspective, the keynote aims to deliver a conceptual introduction for the following contributions and to present the main focus of the panel. First, a working definition of ‘Religion’ from a Cognitive Linguistic perspective is provided that encompasses various religious and spiritual forms from rigid ritual practices to individual forms of bricolage involving diverse concepts. In order to explore how these religious and spiritual systems are formed, expressed and shared, we distinguish three varieties of orientations towards transcendent concepts (cf. Schmitt 2023). In this context we want to bridge the discussion in the German research of the Linguistics of Religion (cf. Lasch/Liebert 2025) to the broader international discourse (cf. Richardson et al. 2021).
Building on this methodological basis, the keynote will address three methodological strategies that can be used to study transcendent concepts and their verbal as well as multimodal expressions.
1. The first part investigates how cognitive linguistic analysis can make use of tools and concepts provided by Linguistic Ethnography. Building on the methodological reflections of Lytra (cf. 2020), this part aims to illustrate how ethnographic fieldwork can give insights into the cognitive scripts (cf. Richardson et al. 2021: 19–20) that underlie communication within religious rituals. One advantage of this perspective is that religious cognitive concepts are not viewed as merely abstract and dislocated, but rather as situated within the concrete sociocultural practices of lived religion in everyday life.
2. The second part combines digital corpus building (cf. Weisser 2018) with a usage-based Construction Grammar approach (cf. Ziem/Lasch 2013; cf. Diessel 2019) to communication in contemporary spiritual offerings. It compiles an exploratory web corpus of websites advertising workshops, retreats etc., and prepares it through text extraction, lemmatization, and part-of-speech tagging. The analysis focuses on recurrent form-meaning pairings that construe agency, including impersonal process constructions, non-human subject constructions, passive guidance patterns, and invitation formulas, such as healing happens, transformation unfolds, energy flows, the body knows, and the universe guides.
3. Especially in digital contexts, we observe individualised and often commercialised forms of spirituality that combine concepts from various religious traditions, indicating an individual need for spiritual orientation within so-called secularised Western societies. Methods to explore this growing phenomenon can be taken from Ethnolinguistics of Mediatised Communication and Digital Pragmatics (cf. Kessler 2026; cf. Meier-Vieracker et al. 2023). We will discuss how data from influencers and online communities present on social media platforms such as TikTok, YouTube, Instagram and Reddit can be extracted and analysed with regard to the multimodal and multivocal construction of transcendent concepts.

The concept of HEAVINESS is grounded in embodied experience and lexicalized by highly frequent and polysemous adjectives across languages. While the basic meanings of these adjectives relate to physical weight and the effort required to lift an object, they display extensive figurative extensions motivated by metaphorical and metonymic processes. A number of contrastive linguistic studies have highlighted differences in the semantic behavior and usage of these adjectives (cf. Yu et al. 2017; Yu & Huang 2019; Ryzhova et al. 2019; Galac & Zayniev 2023: 472) – however, thorough corpus-based analyses incorporating frequency data are still lacking.
The present study seeks to address this gap by examining the semantic structure of the central adjectives encoding the meaning ‘heavy’ in English, German and Hungarian (English heavy, weighty, German schwer, Hungarian nehéz, súlyos), alongside some related items (e.g., German schwierig, gewichtig, Hungarian nehézkes). Preliminary observations suggest that although the adjectives of the three languages share a wide range of figurative uses – essentially rooted in the primary metaphors IMPORTANCE IS WEIGHT (Yu et al. 2017) and DIFFICULTY IS WEIGHT (Yu & Huang 2019) –, there are considerable differences in the frequency and typicality of these uses. English appears to favor object-oriented profiles focusing on the weight of an object as a source domain and resulting in the extended meaning ‘intense’ or ‘of great quantity’ (e.g., a heavy smoker, heavy rain, heavy silence, a heavy accent). On the other hand, German and Hungarian draw more typically on the experiencer-oriented source meaning ‘difficult to lift’, leading to the extended meaning ‘difficult for someone’ (e.g., German eine schwere Krise ‘a serious crisis’, Hungarian nehéz kérdés ‘a difficult question’, nehéz terep ‘a difficult terrain’, nehéz álom ‘troubled, restless sleep’). In addition, the typical figurative uses of German schwer seem to be somewhere between those of English heavy (typically meaning ‘intense’) and of Hungarian nehéz (typically meaning ‘difficult’), since they often mean ‘serious and bad’ (e.g., schwere Schäden ‘serious damage’, ein schwerer Unfall ‘a serious accident’, schwerkrank ‘seriously ill’), thus combining the conceptual metaphors IMPORTANCE IS WEIGHT and DIFFICULTY IS WEIGHT.
The analysis examines the phenomenon on the basis of empirical data drawn from the TenTen corpus family: a collection of large, multipurpose, and comparable corpora consisting of written texts created mostly in digital contexts. The methodology primarily consists of collocational analyses and the manual annotation of randomly generated concordance samples. The semantic differences outlined above are hypothesized to have measurable consequences for the collocation frequencies and the KWIC profiles of the examined keywords. The derivational possibilities of the keywords are also considered (e.g., English usage-heavy, German jemandem schwerfallen ‘to be difficult for someone’, sich schwertun mit etwas ‘to struggle with something’, Hungarian nehézfejű ‘thick-headed, slow on the uptake’, neheztel valakire ‘to bear a grudge against someone’), providing further data on the behavior of adjectives meaning ‘heavy’ in the languages studied.

Many conceptual metaphors originate directly or indirectly from bodily experience (see Sullivan 2017, Gibbs 2017). However, even though having a body is a shared human experience everywhere, metaphors involving the human body are not universal across all languages (see, e.g., Yu 2007, Ibarretxe-Antuñano 2013, Gibbs 2017). Rather, research on Basque, Mandarin Chinese and Australian Indigenous languages indicates that body metaphors are often culturally specific to some extent (see Ibarretxe-Antuñano 2013, Yu 2007, Evans & Wilkins 2000). This shows the importance both of including underrepresented languages in cognitive linguistic research and of integrating their respective cultural context into the analyses (see also Rice 2017 for an overview of the contributions of Indigenous language research to the field of cognitive linguistics). Conceptual metaphors involving body part vocabulary are also found in Taruma, a highly endangered and as of yet undescribed Indigenous language of Guyana (see Eberhard et al. 2025): examples include doya ada ‘house head’, i.e., ‘roof’, or achirahĩ ‘think’ being derived from achi ‘eye’. Following Conceptual Metaphor Theory (Lakoff & Johnson 1980), these can be analyzed as A HOUSE IS A BODY and as VISUAL PERCEPTION IS COGNITION, respectively. The study proposed here is a qualitative exploration of conceptual metaphors in Taruma compounds, idioms and lexical items derived from body part terms. The aim is to categorize the semantic domains that feature in body part metaphors, and to provide some socio-cultural context for these mappings. The data for this research stems from my fieldwork with the Taruma, which included word list elicitation as well as recordings of culturally significant topics. The vocabulary elicitation was based on the Intercontinental Dictionary Series, which includes 22 semantic domains (Key & Comrie 2023). The lexical items were then entered into a FLEx database (SIL International 2025), which currently comprises approximately 950 entries, in preparation for a Taruma dictionary. Despite the comparatively low number of lexical items, the variety of semantic domains included in the data nevertheless makes it possible to show tendencies regarding conceptual metaphors in body part vocabulary: in compounds, the domains appearing most frequently relate to the natural world, such as plants or landscape features, and to items from everyday life, such as tools. In idioms and derived words, the body serves as source for emotions and cognition. This shows that in Taruma, the human body constitutes a central source domain used to make sense of the environment, both natural and man-made, while also being used to form words for cognitive processes like thinking. While this reflects tendencies common in many languages (see, e.g., Tjuka 2024 for a systematic analysis of body-object colexifications in 396 language varieties), the exact mappings in Taruma are also influenced by sociolinguistic and historical factors, such as the speakers’ Taruma-Wapichan bilingualism, as some of the compounds may be calqued from Wapichan.

How languages encode spatial location offers a privileged window into the relationship between linguistic structure and human cognition. While influential work in cognitive semantics and typology has proposed universal primitives such as in, on, and under as foundational spatial concepts (e.g. Jackendoff 1983; Landau & Jackendoff 1993), cross-linguistic research increasingly suggests that languages differ in how spatial meaning is distributed across grammatical components (Talmy 2000; Levinson & Wilkins 2006). This paper advances this debate by examining locative constructions in Malayalam, a Dravidian language whose spatial system challenges strong universalist assumption about spatial conceptualisation. The study draws on data from a Topological Relations Picture Series (TRPS) survey, (Bowerman & Pederson 1992), supplemented by native-speaker judgments, to analyse how Malayalam encodes simple locative scenarios and to identify its Basic Locative Construction (BLC). The analysis is grounded in a cognitive-semantic model of locative meaning that extends Talmy’s macro-event framework to states of location, distinguishing an inner frame (Figure, Ground, fact-of-location, core spatial schema) and an outer frame (posture, manner, placement, and causal information). The findings show that Malayalam overwhelmingly employs a construction of the form Figure + Ground + LOC + copula as its BLC, with the locative case marker -il functioning as a highly underspecified General Spatial Term. A single locative case marker -il is used in contexts of co-incidence, contiguity, containment, support lexicalised by the prepositions IN, ON, AT etc in English. The locative is used mostly for denoting canonical/underspecified/spatially vague relationship between the figure and the ground as long as the Figure and Ground are co-planar. 1. pusthakam mesa-il uNDə book table-LOC COP ‘the book is on the table’ 2. meen veLLath-il uNDə fish water-LOC COP ‘the fish is in the water’. A more specialized post-position like uLLil (inside) etc. is used only when the figure-ground interaction is non-canonical. Eg.- a stamp inside an envelope instead of on it, which is the canonical location for a stamp. Rather than encoding fine-grained topological distinctions in spatial relators, Malayalam systematically offloads semantic detail to posture verbs and converb constructions. Posture verbs such as irikkuka ‘sit’, nikkuka ‘stand’, and kiDakkuka ‘lie’ encode cognitively salient distinctions relating to prototypicality, functional relevance, and intentionality. Eg. 3. pusthakam mesa-il irikk-unnu book table- LOC sit-PRES ‘the book is on the table’ (properly placed) vs. 4. pusthakam mesa-il kiDakk-unnu book table- LOC lie-PRES ‘the book is on the floor’ (haphazardly placed). More complex Figure–Ground configurations are expressed through manner–placement constructions, as in eg.5 which construe location as the result of an implicit agentive event. 5. noolə mezhukuthiri-il keTT-i -vechə- iTTə-uNDə Ribbon candle-LOC tie-CVB-keep-CVB- PRF-COP ‘The ribbon is (tied) on the candle’ These patterns indicate that Malayalam spatial cognition privileges canonicality, prototypicality, functional interaction, and event structure over purely geometric distinctions. The paper argues that spatial concepts such as containment and support are not universally encoded as discrete primitives, but may instead emerge from language-specific strategies that integrate grammar, pragmatics, and embodied experience. By foregrounding how Malayalam distributes spatial meaning across case, verbal predication, and inference, this study contributes

Research on bilingual speech has shown that majority-language words appear with varying frequency in minority-language contexts, prompting the question of whether highly frequent items have become integrated into speakers’ repertoires rather than functioning as genuine code-switches. The frequency of these items in bilingual utterances has therefore been used to determine their status as borrowings (Poplack, Sankoff & Miller 1989; Myers-Scotton 1993; Hakimov 2021). Frequently occurring items are treated as established borrowings, while less frequent ones are classified as nonce borrowings (Sankoff, Poplack & Vanniarajan 1990; Poplack 2018), mixed constituents (Myers-Scotton 1993; Hakimov 2021), or code-switches (Backus 2015). However, because individual studies define the frequency threshold for an “established” item arbitrarily, the psychological plausibility of this distinction remains uncertain.
The aim of the present study is therefore to experimentally investigate the processing of lexical insertions that vary in frequency in bilingual discourse by utilising behavioural and neural measures. The stimuli consist of Russian sentences containing embedded single-word German items, whose frequencies range from one to ten occurrences in a corpus of 25 hours of Russian–German bilingual speech (Hakimov 2021), as well as Russian sentences containing their Russian equivalents.
The study tests adult Russian–German bilinguals residing in Germany. Behavioural responses are collected through a word-monitoring task (Adamou & Shen 2019), while electrical brain activity is recorded via EEG (Luck 2014). Following the experiment, participants rate the target lexemes in terms of their perceived usage frequency within their bilingual community.
We expect to show that the processing of frequent German insertions, i.e., established borrowings, is not associated with high cognitive costs, whereas the processing of infrequent insertions and of Russian equivalents of frequent German insertions does incur such costs. Additionally, the findings will clarify how corpus-based and speaker-perceived frequency measures interact in the cognitive processing of lexical insertions.

References
Adamou, Evangelia & Xingjia Rachel Shen. 2019. There are no language switching costs when codeswitching is frequent. International Journal of Bilingualism 23(1): 53–70.
Backus, Ad. 2015. A usage-based approach to code-switching: The need for reconciling structure and function. In Gerald Stell & Kofi Yakpo (eds), Code-switching between structural and sociolinguistic perspectives [Linguae & Litterae 43], 19–38. Berlin: De Gruyter.
Hakimov, Nikolay. 2021. Explaining Russian–German code-mixing: A usage-based approach. [Contact and Multilingualism]. Berlin: Language Science Press.
Luck, Steven J. 2014. An introduction to the event-related potential technique. 2nd ed. Cambridge, Mass.: MIT Press.
Myers-Scotton, Carol. 1993. Duelling languages: Grammatical structure in codeswitching. Oxford: Clarendon Press; Oxford University Press.
Poplack, Shana. 2018. Borrowing: Loanwords in the speech community and in the grammar. New York: Oxford University Press.
Poplack, Shana, David Sankoff, & Christopher Miller. 1988. The social correlates and linguistic processes of lexical borrowing and assimilation. Linguistics 26(1): 47–104.
Sankoff, David, Shana Poplack, & Swathi Vanniarajan. 1990. The case of the nonce loan in Tamil. Language Variation and Change 2(01): 71–101.

In this paper, we apply a usage-based approach to the acquisition of figurativity in German-English bilingual children focussing on the verbs see and sehen to investigate how far input functions as a proxy for production, and how the complexity of the syntactic objects correlates with the level of figurativity of the verb see/sehen. We consider figurativity to reflect a continuum from literal (“I can see the elephant”) to metaphorical (“I see what you mean”), including in-between cases in which both, a literal and a metaphorical meaning, applies (“Let’s see what’s inside the box”).

We address our questions using data from two highly dense corpora of German-English bilingual children aged 2 to 4 (Fion and Silvie from the MPI-EVA-Leipzig corpora, Quick et al., 2018). We retrieved every instance of the lemmas see (N = 4621) and sehen (N = 1176) in the input and the output, and coded these for metaphoricity (as a three level factor) as well as for the constructional pattern they occurred in.

We found that literal and figurative uses prefer different syntactic patterns (Fig. 1). This correlation is particularly strong in the input. The association of figurative meaning with a distinct grammatical form (in line with Construction Grammar, Goldberg, 1995) may help children navigate the different senses of see/sehen. With respect to their developing figurativity, we found that the children’s output reflects their input (Fig. 2), i.e. the more figurative input, the more figurative output, and that in our context of child-caregiver interaction, figurative uses primarily occur in cases where “both”, a literal and a metaphorical meaning applies (cf. Grady, 1997), often functioning as a kind of verbal pointing gesture.

References

Goldberg, A. (1995). Constructions. A Construction Grammar Approach to Argument Structure. The University of Chicago Press.

Grady, J. E. (1997). Foundations of Meaning: Primary Metaphors and Primary Scenes [Dissertation, University of California, Berkeley]. http://escholarship.org/uc/item/3g9427m2

Quick, Antje Endesfelder, Elena Lieven, Ad Backus & Michael Tomasello. 2018. Constructively combining languages: The use of code-mixing in German-English bilingual child language acquisition. Linguistic Approaches to Bilingualism 8(3). 393–409. https://doi.org/10.1075/lab.17008.qui.

Mixing languages is one of the more salient phenomena that result from bilingualism (e.g., Gardner-Chloros 2009). From a usage-based perspective, children’s mixed utterances are not random but are grounded in productive lexico-grammatical constructions, where open slots in frame-and-slot patterns are filled with lexical material from the other language (Quick et al. 2021). In many cases, these patterns predominantly come from one of the languages, creating the asymmetry often found in adult code-mixing known as the ‘matrix language effect’ (Myers-Scotton 1997, see Wasserscheidt 2020 for a critical discussion). Building on this perspective, the present study investigates morphological code-mixing, defined as the combination of morphemes from two languages within a single word, in the spontaneous speech of young German–English bilingual children, with a focus on how such patterns emerge from usage and input.
We propose that children acquire morphology in a bottom-up fashion, beginning with unanalyzed forms and gradually abstracting recurring patterns such as German VERB-en infinitives or English NOUN-s plurals. With increasing exposure, these patterns become entrenched and productive, allowing children to extend them analogically across languages. This process is reflected in morphological mixes such as jumpen (‘jump’ + German infinitive marker), wrinkelig (‘wrinkle’ + German adjective marker), and feete (‘feet’ + German plural marker), which illustrate how children integrate cross-linguistic material into emergent schemas.
The study draws on longitudinal naturalistic data from three simultaneous bilingual children, Fion, Silvie, and Lily (ages 2–3), comprising 109,815 utterances. All children were raised in Germany under a one-parent–one-language approach, but their input conditions differed in the relative proportion and distribution of German and English. A total of 10,040 code-mixed utterances were extracted and systematically analyzed for instances of morphological mixing. Across the full dataset, 696 instances of morphological code-mixing were identified and coded for type of mixing, part of speech, language of stems and affixes, and grammatical categories involved. Morphological mixing was relatively rare (fewer than 2% of utterances) but present in all children’s speech. Results show that nouns, verbs, and adjectives function as templates for mixed morphology (Figure 1), and inflectional mixing emerged as the most frequent type (Figure 2). Fion and Silvie, both German-dominant, predominantly produced English lexical stems combined with German inflectional morphology, particularly in verbs (e.g., fighten ‘fight’ + German infinitive marker). Lily, whose input and proficiency are more balanced, also produced the reverse pattern, combining German stems with English morphology (e.g., schalted on ‘turned on’). In the nominal domain, all children produced mixed compounds in both German–English and English–German configurations, such as apfel juice (‘apple juice’) and birthtag (‘birthday’).
Overall, the findings demonstrate that morphological code-mixing is systematic, input-sensitive, and closely related to children’s developing proficiencies in their two languages, showing that the observed asymmetries can be accounted for by the relative entrenchment of frame-and-slot patterns

This study examines the use of English swearwords in Italian WhatsApp chats among Italian young adults. The data consists of 89 dyadic and group conversations among 194 users (mostly university students), with around 80,000 text messages and 280 voice messages, collected between 2015 and 2024 at the University of Pavia (Fiorentini et al. 2024, Fiorentini 2025).

Drawing from Digital Conversation Analysis, Interactional Linguistics, and the notion of Affordances, the analysis shows how language contact enables speakers to mobilize and extend their repertoire of profanity, contributing, i.e., to “playfulness” in gamers’ settings (Calabria et al. 2025), but also alignment and intersubjectivity in dyadic chat interactions (Sciubba & Fiorentini 2024), similarly to what happens in offline interactions (Calabria & Sciubba 2022, Sciubba & Calabria 2023, Calabria & Sciubba 2025).

From the participants’ perspective, English swearwords are not treated as problematic in any of the chats analyzed, as they are not repaired in subsequent turns, and they, in some cases, are further creatively morphologically adapted following Italian (phono-)morphological rules (Palermo 2015), as seen in the excerpt below, l.2408.

Excerpt
2407 29/02/20, 15:08 - EV01: Ho voglia di una sessione lunga as fuck
I feel like a session long as fuck
2408 29/02/20, 15:09 - OV01: Dipende da chi fucka
It depends on who’s fucking
2409 Se parli di me è corta
If you’re talking about me it’s short

Considering that Italian has already a wide variety of swearwords, both in the so-called standard language and in the other Italo-Romance varieties present in Italy, some of the English swearwords used are
- either not available in Italian: “lunga (long) as fuck (l.2407 above) /compromettenti (compromising) as fuck”, there is no modifier/intensifier that could replace as fuck in that same position in the clause and in turn-final position; wtf as an exclamation/interjection, or
- are used as a crystallized construction “spicy af/suspicious af” taken as is from the available lexical/syntactic repertoire available to the participants.

In addition, English is a language widely available to (young) Italians for three main reasons: it is learned as a foreign language in schools for 13 years; it has been a primary site of language contact on internet in the last few decades (Coco 2003, Gualdo 2003); and, last but not least, English is available as a vehicular language in online videogaming, normalizing the use of English profanities (Calabria et al. 2025).

The data shows that for Italian chat participants the English swearwords are just another addition to their available linguistic repertoire, blurring the boundaries of what can be placed on which continuum (loanwords/semantic calques, lexicon/syntax).

Word-by-word repetitions of a prior utterance may be used to express a recognisable and often complex attitude, feeling, or situation occurring in a different context, yet involving a similar attitude, feeling, or situation. For example, the common phrase ‘Eureka feeling’, referring to a sudden moment of insight, originates from Archimedes’ legendary exclamation of discovery “Εύρηκα!” (‘I have found it!’, Pascual 2014). Such verbatim non-actual quotes have mainly been studied as adaptation strategy of communication in language pathology (Bánréti 2010; Versluis & Kleppa 2016), and marginally as rhetorical tactic in legal argumentation (Pascual 2014) or satire (Fonseca et al. 2020). However, to our knowledge, the non-quotative use of literal repetitions eliciting new meanings has never been systematically examined in online discourse. We aim to fill this gap by exploring the use of such enunciations in social media. We understand such intertextuality as involving fictive interaction i.e., the use of the Conversation Frame to structure thought, discourse, and language (Pascual 2002, 2006, 2014, Pascual & Sandler eds. 2016).
We collected forty verbatim fictive enunciations on Weibo (China’s equivalent of Twitter, now X). We also designed a questionnaire to understand netizens’ interpretations of high-frequency instances, also requesting respondents to provide 2-5 extra similar examples. These creative word-by-word repetitions are mainly from memorable communicative events in viral video clips, popular TV shows or movies. An example is the sarcastic use of “真香” zhēnxiāng (‘It’s so appetizing!’) to comment on somebody’s drastic change in behaviour or attitude after a video clip went viral of an urban child in a famous Chinese reality show exclaiming “It’s so appetizing!” upon tasting the rural food that he had previously shown contempt for. We categorized such instances across dimensions including: (i) syntactic structure, (ii) target meaning, (iii) pragma-rhetorical function(s), and (iv) creativity level.
We found that: (i) Chinese verbatim fictive enunciations may appear at various morphosyntactic levels, from the sentence to the lexical item; and (ii) they involve various cognitive mechanisms, such as frame metonymy (i.e., using a single utterance to evoke the whole situation it originally occurred in, then one prominent aspect of it); fictivity (i.e., both the enunciator and audience are fictive, meeting in a writer-reader blend, Herman 1999; Fauconnier & Turner 2002); and viewpoint mixing (cf. Dancygier et al. eds. 2016, e.g., in its fictive usage, ‘It’s so appetizing!’ is both anchored in and ascribed to the original utterer and ascribed to someone who had undergone a similar attitude change). We argue that: (i) the creative re-use of viral enunciations plays a vital role in community-building and shaping a sociolinguistic and generational subculture, since they are mostly based on trending online content shared by a generation of Chinese netizens; (ii) they enhance the expressive power of written language just as emojis and memes do; and (iii) they show the need for distinguishing between reported speech (as in ordinary quotes) and speech (re)presentation (in fictive speech, whether verbatim or novel).

Even though linguistic creativity is ubiquitous and diverse (Hoffmann 2019), prior experimental work has focused on investigating selected types of creative language, most notably metaphors (Holyoak & Stamenković 2018). It therefore remains unclear whether findings from metaphor research generalize to other types of linguistic creativity. To address this, we compared the perception of novel metaphors with novel “verbing,” that is, the formation of new verbs from nouns (e.g., to porch a newspaper; Clark & Clark 1979).

Specifically, we tested the generalizability of two key findings from metaphor research. First (H1), we predicted that novel verbing and metaphor would be perceived as unfamiliar but nevertheless meaningful, especially if embedded in a supportive linguistic context (Bambini, Resta & Grimaldi 2014). Second (H2), we predicted that verbing and metaphor, in virtue of being both novel (and thus challenging) and effective (i.e., interpretable), would be perceived as aesthetically pleasing (Errington, Thye & Mirman 2022).

We created 81 novel instances of verbing, such as (1a), and 81 novel metaphors, such as (2a). Each creative expression was matched with a prototypical control (illustrating canonical/literal use), as in (1b) and (2b), and an anomalous (uninterpretable) control, as in (1c) and (2c).

(1) a. He used to skillet the onions
b. He used a skillet to fry the onions
c. He used else skillet the onions

(2) a. Her smile clattered to the floor
b. Her keys clattered to the floor
c. Her beach clattered to the floor

In the experiment (N = 135), each participant rated one version of each triplet (such as (1) and (2)) on a 7-point scale along one of three dimensions: familiarity, meaningfulness, and aesthetic pleasure. Half the participants read the expressions in isolation (no context), whereas the other half saw them embedded in a short poem (with context).

A linear mixed-effects analysis yielded main effects of condition on familiarity and meaningfulness, indicating that both verbing and metaphors were rated as less familiar (medians ≤ 2 out of 7) than prototypical controls but more meaningful (medians ≥ 4 out of 7) than anomalous controls (all ps < .001). For verbing, we also found a main effect of context on meaningfulness (p = .046), showing that verbing was rated as more meaningful when embedded in context. These results support H1.

Aesthetic pleasure showed an interaction between condition and context for metaphors, indicating that they were more pleasurable than prototypical controls when embedded in context (p < .001). In contrast, a main effect of condition emerged for verbing, indicating that it was perceived as less pleasurable than prototypical controls (p < .001), with no effect of context. Thus, H2 was only supported for metaphors but not for verbing.

These results demonstrate that creative verbing, similar to novel metaphors, is perceived as meaningful, especially in context. However, contrary to metaphors, verbing is not perceived as aesthetically pleasurable. This suggests that linguistic creativity is not a uniform construct, and that conclusions drawn from metaphor research should not be hastily generalized to other creative devices.

Understanding and documenting the mental inventory of constructions, which form a complex network called constructicon, is an important part of Construction Grammar (CxG) theory.
In recent years, the field has seen the emergence of reference constructicons: they are computerized databases (mostly accessible online) of the constructions of a single language (Lyngfelt 2018). The term "reference" distinguishes these digital databases from the constructicons in the mind of speakers, because reference constructicons describe a version of constructions idealized at the level of the community of speakers, rather than a cognitive reality that could vary from one speaker to another (Herbst 2019). One example for German is the FrameNet-Konstruktikon des Deutschen (https://framenet-constructicon.hhu.de/), which has been gradually built up since 2018 by a team at Heinrich-Heine-Universität Düsseldorf.

In our talk, we will present an ongoing constructicographic project called the "Research Constructicon". It is a reference constructicon with a particular focus on ongoing research, notably that conducted since 2022 by doctoral students at the Research Training Group Dimensions of Constructional Space at Friedrich-Alexander-Universität Erlangen-Nürnberg (https://www.cxg.phil.fau.eu/).
The CxG community is diverse, with researchers working across many subfields. Early-career scholars in our group explore phenomena across time, languages, or modalities (e.g., speech-gesture interaction), employing varying degrees of formalism. Our project aims to establish common ground within this group, as a microcosm of the broader community.

This project went through several steps since 2022:

The idea to form a database based on ongoing research while adopting a constructionist approach led us to adopt the tools of Semantic Web technologies, which can adapt much more flexibly to evolving requirements than relational databases (Patel et al. 2022).

After agreeing on a database standard, the first step of the project was to define a vocabulary for publishing construction entries. We thus defined different ontologies, based on the current research and practices within the CxG community (Winckel et al. 2025, Winckel et al. 2026). When possible, we based the ontology on already existing ones, and refer especially to OLiA (Ontologies of Linguistic Annotation; Chiarcos 2015) for linguistic terminologies such as semantic roles or verb tense. We would welcome other constructicographic projects to adopt and expand these ontologies, so that we achieve better interoperability between our databases.

In a second phase, we collected construction descriptions and implemented them in the database. Each description includes information on the elements of the construction, their collo-profiles (i.e., the respective frequencies of possible fillers), but also how constructions are linked to each other (within a single language or across languages). They were all provided by the doctoral researchers themselves, as experts on their constructions.

Finally, we designed a web application providing online access to the Research Constructicon at https://bdlweb.phil.uni-erlangen.de/RCxn/.

In this presentation, we will explore the project’s development, the challenges we faced, the solutions we implemented, as well as remaining difficulties (e.g., integrating syntactic dependencies in the linearization of the construction elements). Our goal is not only to share our process but to invite dialogue on how such initiatives can evolve in the future.

For his 1985 paper “Popular views of the proverb”, Wolfgang Mieder asked 55 English-speaking informants for a definition of the word ‘proverb’. As the shortest common denominator, Mieder (p. 119) proposes “A proverb is a short sentence of wisdom.” At the same time, he also notes that the informants’ views do not necessarily correspond to how proverbs are usually defined in linguistics. For instance, not all informants mention that a proverb has to be a full sentence (or sentence-like unit), while several informants assume that a proverb by definition has to be figurative. This suggests that ‘proverb’ is a prototype category (cf. also Norrick 2014). In this talk, we will present two studies aiming to pinpoint this prototype.

The first part of the talk focuses on extended replication of Mieder’s study, for which we had 120 German-speaking informants with different degrees of education (30 without college/university education, 30 with college/university education without language as a core area, 30 with college/university education with language as a core area, and 30 professional linguists) fill out a questionnaire in which they had to give a spontaneous definition of the word ‘proverb’. We also asked to give five examples of proverbs and judge for six features whether they are necessary for an expression to count as a proverb (consists of an entire sentence, mentions worldly wisdom, contains rhyme, is figurative, etc.). Our results are in line with Mieder’s observations in that both the definitions and the examples show that the prototype of the category ‘proverb’ is centered around three features: proverbs are full sentences with a figurative reading and expressing worldly wisdom. There is, however, a striking difference between the definition attempts and the examples in that consisting of an entire sentence is the least central of the three features in the definitions, but more central than the figurative reading in the examples. We will discuss a few possible explanations, including that there seems to be some confusion among the informants as to what counts as a full sentence.

Obviously, a restriction of this first study is that we can only analyze the examples our informants came up with spontaneously; we cannot exclude that other expressions would be considered as even better representants of the category but that they did not think of them spontaneously. To this end, we will also present a second study for which we asked informants from the same four groups to rate expressions with regard to whether they are typical proverbs or not. The expressions were chosen in such a way as to be able to judge the role of the three central features and rhyme (not a central feature, but still repeatedly mentioned in the definitions in the first study) for the typicality of a proverb. In this way, this second study allows us to verify and refine the prototype as defined in the first study.

Montevideo occupies a distinctive position within the Río de la Plata linguistic area due to the stable coexistence of two informal second-person singular address signals: tú and vos. Although voseo (the use of vos) predominates in contemporary usage, tuteo (the use of tú) remains productive and appears systematically across discourse contexts. Previous scholarship has documented this alternation extensively, primarily through sociolinguistic analyses focused on age, education, formality, and interpersonal relations. While these studies describe important distributional tendencies, the communicative rationale underlying the coexistence of tú and vos remains insufficiently explained.
This paper examines the Montevideo address system through the lens of Columbia School linguistics, following the sign-based approach associated with William Diver. Within this framework, linguistic forms are treated as meaningful signals whose distribution reflects invariant semantic contrasts. Rather than assuming tú and vos to be equivalent variants conditioned exclusively by social factors, the study explores the possibility that the two signals encode distinct communicative meanings.
The analysis draws on written reported discourse from El País, El Observador, and Montevideo Portal, representing contemporary Montevidean Spanish. Through qualitative and quantitative analysis, the study investigates the discourse environments in which tú and vos occur, with particular attention to cases of intra-speaker alternation and interactions involving a constant interlocutor. The central hypothesis proposes that tú signals interlocutor individuation, construing the addressee as a distinct discourse target, while vos signals interlocutor access through interactional integration, construing the interlocutor as relationally embedded and interactionally accessible. Preliminary examples suggest that tú tends to occur in contexts involving explicit attribution, rhetorical targeting, or interlocutor singling out, whereas vos predominates in contexts characterised by relational continuity and interactional immediacy. This distinction becomes particularly visible in cases of alternation such as 'a vos te llama un presidente… y tú lo agarrás', where the same interlocutor is construed differently within a single discourse sequence.
Preliminary findings suggest that the alternation between tú and vos is neither random nor fully reducible to traditional sociolinguistic variables alone. Rather than functioning as fixed social labels, the two signals appear to guide contextual interpretation through distinct discourse construals.
By examining address forms from a Columbia School perspective, the paper contributes to the study of linguistic variation as meaning-driven behaviour and offers a functionally motivated account of pronominal choice in Montevideo Spanish. More broadly, it illustrates how Columbia School methodology can illuminate semantic organisation within variable systems.

In this presentation I will share partial results of an ongoing investigation that aims to account for some grammatical features of Mapudungun, an indigenous language of Chile and Argentina. I will focus on the semantic contribution of certain forms traditionally associated with the notion of grammatical number, with the aim of explaining their motivated distribution in a corpus of authentic texts.
Specifically, I will analyze the so-called third-person verbal inflections of the realis mood, ­-y, ­yngu and -­yngün. These forms are typically classified as singular, dual and plural, respectively (Augusta, 1903, pp. 26-29; Hernández Sallés et al., 2006, p. 80; Raguileo Lincopil, n.d., pp. 15-16; cf. Salas, 1992, p. 111; Smeets, 2007, pp. 151­152). However, this classification does not satisfactorily explain their actual range of use. For instance, frequently “the verb in the singular is used, even when the subject is plural or dual” (Augusta, 1903, p. 26, the translation is mine). In addition, the “dual” ­-yngu is sometimes used with a plural referent, as in example 1, while the “plural” ­-yngün is sometimes used with a dual referent, as in example 2.

1) Pu wenteche kurekefiŋu re kizu ñi mapu ñi che müten / The mountains people generally married women from their own land only (Guevara 1913, p. 53, the highlighting and translation are mine)
2) Ñi epu kure müten konfuiŋun / Only two of his wives could go in (Guevara 1913, p. 17, the highlighting and translation are mine)

My analysis is grounded in the Columbia School of Linguistics (Diver, 1995, 2012[1975]; Huffman, 2001, 2006; Davis, 2004; Stern, 2019), a functional approach that aims to find empirically-based explanations for facts of actual usage, rather than relying on aprioristic assumptions about how languages are structured. Prior work done within this framework attests to the value of such an inductive approach to the study of indigenous languages (e.g., Contini-Morava, 1987, 2002, 2012; Contini-Morava & Danziger, 2018, 2022; Lai, 2024; Ventayol-Boada, 2024). My study is further guided by the ethnopragmatic approach (e.g., García, 1995; García y Otheguy, 1983; Martínez, 2004, 2009, 2012, 2015; Zanfardini, 2020, 2022), a perspective that has shown that skewings in the use of comparable linguistic signals correlate with varying communicative needs.
The corpus is made up of the historical accounts (nütram) compiled in Guevara (1913). A mixed methods approach is adopted, combining qualitative and quantitative techniques.
My proposal is that occurrences of ­-y, ­-yngu and ­-yngün in the corpus can be explained by positing that, while the three forms point to what in traditional terms is referred to as a third-person subject −an entity that is different from speaker/writer and hearer(s)/reader(s)−, they differ in that ­-y makes no further specification with respect to that subject. In contrast, -­yngün indicates that the subject should be understood as comprising two or more constituent parts, while ­-yngu groups the subject together with another, external entity (e.g., an object) that is also fundamentally involved in the occurrence of the event.
Preliminary results reveal some previously unexplored facts about how Mapudungun works.

This study examines consonantal phonotactics in the poetry of Vasilisk Gnedov, one of the leading figures of Russian Ego-Futurism, by comparing his early poems with a reference corpus compiled from poems by Akhmatova, Bunin, and Gumilev and taken to represent a relatively normative variety of early twentieth-century literary Russian. Gnedov is a particularly relevant test case because his poetry contains neologistic and phonologically marked forms, such as proklenushek, zlatokopytok, kruzho vypenil, strepet, and galoche stanyval, which raise the question of whether his experimental diction departs from Russian phonotactic organization or systematically redistributes its resources. The analysis is conducted within the framework of Phonology as Human Behavior (PHB), introduced by Diver (1979) and further developed by Davis (1987), Tobin (1997), and others. Within PHB, phonology is understood not as an autonomous formal level but as part of general human behavior shaped by a constant compromise between communicative distinctiveness and articulatory economy.
The study is based on two comparable corpora: 870 words in Gnedov’s corpus and 854 words in the reference corpus. Since vowel parameters are nearly identical in the two corpora, the analysis focuses exclusively on consonants and consonant clusters: the frequency and density of two-member and three-member clusters, the positional distribution of clusters, especially in word-initial position, and selected PHB-relevant oppositions, including word-initial /r, l/ clusters involving the same active articulator as opposed to clusters involving a shift between active articulators, maximal constriction as opposed to first-degree stricture, and voiceless as opposed to voiced consonants in initial position.
The results suggest that Gnedov’s poetic language is best understood not simply as a departure from Russian phonotactic norms or as an arbitrary invention of exotic, difficult-to-articulate clusters, but as a form of phonological behavior that systematically redistributes available phonotactic resources within the Russian phonological system. Compared with the reference corpus, Gnedov’s poems show a tendency toward more frequent two-member clusters (54.6 vs. 50.5 per 100 words), especially in salient word-initial position (25.2 vs. 20.3 per 100 words), alongside a slightly lower frequency of three-member clusters (3.2 vs. 3.5 per 100 words). In the subset of word-initial clusters involving /r/ and /l/, both PHB-favored combinations, involving the same active articulator, and less-favored combinations, involving a shift between active articulators, are more frequent in Gnedov’s corpus: the former occur at 8.6 vs. 6.4 per 100 words, and the latter at 3.8 vs. 2.2 per 100 words. This pattern points to selective strengthening of certain cluster configurations rather than to generalized articulatory complexity.
Gnedov’s experimental Ego-Futurist poetics thus emerge not as a rupture with linguistic norm, but as a motivated and internally structured shift within the range of phonological behavior available in Russian.

In the conclusive talk, the panel organizers will offer a summary of the theoretical and methodological iinsights gained, and questions raised by the talks, presented at the panel. They are also going to discuss the contribution of the presented research to the further development of the Columbia School (CS) framework, highlighting the diversity of represented approaches and languages. After that, the floor is going to be opened for the exchange of opinions, as well as the suggestions for the next venue for the panel, alongside the possible journal; to publish the panel proceedings.

Cognitive Linguistics and Paradox: We examine how paradox can be approached through source-domain reversal (Richardson & Mueller, 2019), where situated explanations draw on a particular source-domain mapping that is subsequently reconceptualized through an opposing mapping. Together with examples involving conceptual blending, this supports the claim that paradox often serves a situated discourse function (Mueller & Richardson, 2025). We then consider dualistic experiences of divine relationship and non-dualistic experiences of non-conceptuality as clusters of idealized cognitive models, with the former underpinned by a LINKED_GOAL image schema and the latter by schema attenuation. Finally, we examine one possible framework for integrating these seemingly opposing strands.
Cognitive Linguistics and Multimodality: We discuss how multimodal channels in religious ritual can be analyzed through conceptual blending (Fauconnier & Turner, 2002). As an example, we present an analysis of the Korean Buddhist Ritual of the Four Instruments (Richardson, Mueller, & Pihlaja, 2021). The blend combines inputs from “ritual space” involving sounds and body movements, “dharma space” involving symbolic correspondences between instruments and beliefs, and “mythic space” involving metonymic associations. We reflect on the distinctive interpretive affordances of ritual, where complex networks of integrated meanings may be only partially accessed by believers at any given moment, yet remain available through repeated ritual encounter. Some meanings may be accessible to new believers, while others depend on specialized religious knowledge (Mueller, 2023).
Cognitive Linguistics and Discourse: Finally, we discuss how discourse can be approached through a cognitive linguistics lens. Focusing on everyday interaction, we show how categories of belief and religious identity are built through abstraction in discourse, with metaphor and metonymy playing key roles. Drawing on research in superdiverse contexts (Pihlaja, 2023), where discussions of religious belief involve comparing one’s own beliefs and actions with those of others, we illustrate how people understand themselves and others not through the strict application of theological labels, but through emergent models grounded in experience. We argue that interaction plays an important role in shaping how religious belief and practice come to be conceptualized as discrete categories, and that cognitive linguistics offers resources for addressing practical problems of speaking across difference in the contemporary world.
References
Fauconnier, G., & Turner, M. (2002). The way we think: Conceptual blending and the mind's hidden complexities. Basic Books.

Mueller, C. M. (2023). Symbols and icons in Buddhist worship. In S. Pihlaja & H. Ringrow (Eds.), The Routledge handbook of language and religion (pp. 96-108). Routledge.

Mueller, C. M., & Richardson, P. (2025). Modern Buddhist and Christian mysticism: Construal and transformation. Routledge.

Pihlaja, S. (2023) Narrative and Religion in the Superdiverse City. Cambridge University Press.

Richardson, P., & Mueller, C. M. (2019). Moving and being still: Exploring source domain reversal and force in explanations of enlightenment. Language and Cognition, 11(2), 310—339. https://doi.org/10.1017/langcog.2019.19

Richardson, P., Mueller, C. M., & Pihlaja, S. (2021). Cognitive linguistics and religious language: An introduction. Routledge. https://doi.org/10.4324/9781003041139

This study investigates how constructed languages, particularly Esperanto, cognitively structure transcendent concepts and contribute to the formation of alternative religious and spiritual worldviews. The research focuses on the relationship between language planning, religious conceptualization, and the cognitive organization of symbolic systems, examining whether planned languages facilitate the emergence of hybridized forms of spirituality comparable to those observed in minority religions and translingual spiritual practices. The theoretical framework combines approaches from Cognitive Linguistics, Conceptual Metaphor Theory, Translation Studies, and the Cognitive study of Religion. In particular, the study draws on theories of embodied cognition and religious framing to analyze how transcendent concepts are linguistically reorganized when sacred and spiritual discourses are translated into planned languages. Special attention is devoted to the role of L. L. Zamenhof’s ethical and spiritual project of Homaranism, understood as a transnational and universalist attempt to reshape religious and moral communication through linguistic neutrality. Methodologically, the research adopts an interdisciplinary and corpus-based approach. It combines Qualitative Discourse Analysis with comparative linguistic analysis of sacred texts translated into Esperanto and other constructed languages. The corpus includes translations of biblical, spiritual texts and community discourse. The analysis focuses on conceptual metaphors, semantic restructuring, religious frames, and the cognitive negotiation of transcendent meanings across linguistic systems.
The study aims to demonstrate that planned languages function not only as communicative tools but also as cognitive environments capable of reshaping religious imagination, symbolic structures, and spiritual categories. More broadly, the research contributes to the understanding of how linguistic systems influence the cognitive construction of transcendence and the emergence of alternative religious and spiritual identities.

The purpose of this paper is to introduce a new approach focused on linguistic and religious complexes, or, jointly, linguo-religious complexes (LRCs), with data from Inner Asia. Inspired by the study of complex systems, a new approach integrates comparisons of interacting languages and religions with an analysis of Inner Asian religious and ritual texts. Special attention is devoted to an interdisciplinary case study on how changes in the natural environment affect tributary relations, interrelated with religious and linguistic shifts.
On the theoretical side, the LRCs approach integrates a) the nexus of language and religion with b) the recognized interconnection of religion and traditional political institutions. As contemporary humanities prefer triangulation, the LRCs' approach does not prioritize the interconnection of linguistics, archaeology, and genetics (LAG). In search of a more precise framework for attested stages of written languages, the LRC approach focuses on languages, religions, and scripts. Because prehistoric rock art (Fitzhugh & Kortum 2024) and early script use (Schmand-Besserat 1992) already show a connection with religious institutions.
Regarding the LRCs, the paper explains interactions of languages and religions. While it can be proven that language shifts are spontaneous, the spread of religious systems and changes of scripts are intentional and influenced by geopolitical shifts: cf. the replacement of Mongolian by the Clear script among the Oirats (Taupier 2015) or the replacement of Written Mongolian by Cyrillic after religious change in Mongolia in the 20th century (Teleki 2022). Wider ethno-religious interactions with a changing natural environment (Di Cosmo 2023) were further mediated through tributary relations, as evidenced by the spread of Buddhism among the Mongols and the gradual replacement of Sanskrit influence by Tibetan in Inner Asian Buddhism.
As complex analysis and East Asian macro-history show, the degree of multi-ethnic, multi-lingual, and multi-religious variation increases from the steppe zone to the equator. These contrasts and long-distance interactions between the steppe and South/Southeast Asia provide strong evidence for the LRCs as an appropriate framework for the study of the interface between cognitive linguistics and the linguistics of religion.
Note: the LAG remains useful for cultural terminology (Wörter und Sachen approach) and for the study of the deeper past.

Summery of abstract:
This study investigates the restrictions on German werden-passives by tracing them back to their underlying semantic and conceptual properties. Corpus analyses of 432 verbs—including borderline and supposedly non-passivizable verbs—show that passivizability is not determined by a single condition, but emerges from a combination of verb-semantic features and argument-structure properties, with a subject argument exhibiting control as the primary requirement. Crucially, semantic conditions are further grounded in the force-dynamic profile of the construction, integrating passive constructions with and without a subject into a coherent explanatory model and showing that only verbs that conceptually profile an energy transfer at its maximum extent are passivizable. The findings provide a concise account of passivization restrictions and highlight the structural cohesion of passive constructions.

The present paper investigates the grammaticalisation of the so-called recipient (or ‘dative’) passive (Leirbukt 1997; Askedal 2005; Lenz 2013) in different varieties of German. The recipient passive has emerged from a reanalysis of monotransitive active constructions with the meaning ‘x receive y’ (primary predication) that are augmented with a participial object complement (secondary predication). On the passive analysis, the secondary predication is foregrounded as primary: the participle is interpreted as the lexical main verb of the clause, and the formerly primary ‘receive’-predicate undergoes auxiliation. The ambiguity between the predicative and the passive reading of the participle is exemplified in (1). By contrast, if there is no nominal direct object to which the participle could attach (cf. (2)), or if such a reading is ruled out on semantic grounds (cf. (3)), the construction can only be interpreted as passive:
(1) Zu Teeniezeiten hatte ich zwei Paar, die ich bis zum Tod abgelaufen habe. Eins habe ich von meiner Mutter geschenkt bekommen, das andere von meinem Taschengeld bezahlt.
‘As a teen, I had two pairs that I wore to death. One I had received from my mother as a present, the other I had bought from my pocket money.’ [DECOW16B]
(2) Wenn man von klein auf eingetrichtert bekommt, dass die eigene Existenz schädlich für den Planeten ist, dann glaubt man es irgendwann.
‘If you’re told from childhood that your own existence is harmful to the planet, you’ll eventually believe it.’ [DECOW16B]
(3) Die vierzehn Jahre alte Ajla aus Sarejevo hat in der Nacht zum Montag in der Göttinger Universitätsklinik eine fremde Leber eingepflanzt bekommen.
‘Fourteen-year-old Ajla from Sarajevo received a liver transplant at Göttingen University Hospital Sunday night.’ [DeReKo]
The objective of our study is to measure and compare the relative degree of grammaticalisation that the recipient passive has attained in different diachronic stages and synchronic varieties of German. Our quantification builds on Himmelmann’s (2004) understanding of “grammaticization as context expansion” and Diewald’s (2002) notion of “isolating contexts” for the new analysis as a particularly crucial stage within the expansion process: specifically, we assume that (i) the construction progressively spreads to new syntactic and semantic environments, thereby increasing in overall usage frequency over time, and (ii) this overall increase involves a growing proportion of unambiguously reanalysed instances that can only be interpreted as passives.
We test these predictions in a corpus study of the German recipient passive in different synchronic (FOLK, DECOW, DeReKo) und diachronic corpora (DTA). In this, we seek to retrace the diachronic development of the construction through time as well as its synchronic diffusion across more vs. less conservative varieties (with a correspondingly lower vs. higher expected degree of grammaticalisation, viz. written standard, informal web forum discussions and spontaneous conversation). Special attention is paid to the extension of the semantic source scenario RECEIVING to ever less similar types of being affected by directed action, and how this gradience in the data can be operationalised.

The study of lexical and semantic change has seen a rise in the use of data-driven approaches to investigate pathways and mechanisms of change (e.g. Allan & Robinson 2012; Geeraerts et al. 2024). One crucial question in this domain relates to the cognitive underpinnings of semantic change (e.g. Brochhagen et al. 2023; Ramiro et al. 2018). A number of recent studies has investigated the contribution that evolutionary linguistics and experimental psycholinguistics can make in elucidating this question (e.g. Fuchs & Piñango 2021; Smith 2025). Such studies often use artificial or ‘laboratory languages’ as tools for research on language change (Roberts 2017). Specifically, these experimental approaches have the potential to replicate observed pathways of semantic change in controlled settings, to shed light on the contextual and cognitive factors that influence these changes. Here we focus especially on studies that investigate the role of metaphor in semantic extension, a factor that has been recognised as a key driving factor in semantic change (e.g. Anderson 2017; Xu et al. 2017; Ramiro et al. 2018). In this design, participants first learn or interactively negotiate novel form-meaning-mappings. In a second step, these established form-meaning mappings then have to be applied to communicate about an expanded, new meaning space, thereby forcing participants to engage in semantic extension (Smith et al. 2025; Anonymous et al. 2024). For example, participants might first learn the mapping of a novel form to the meaning DOWN. Then, when asked to communicate about an expanded set of meanings, it can be tested whether participants tend to use the form they learned for the meaning DOWN to also refer to the meaning SAD, via the conceptual metaphor SAD IS DOWN (Lakoff & Johnson 1980; Anonymous under review). Through the possibilities of freely manipulating the meaning space in this design, this paradigm can be productively integrated with existing work on patterns in semantic change. Specifically, by determining the original meanings that participants have to learn - as well as the new meanings they have to communicate about using semantic extension - in ways that correspond to documented patterns of change, this paradigm has the potential to replicate processes of semantic change observed in language history. It can therefore provide a valuable methodological bridge between data-driven descriptive work on semantic and lexical change and explanatory accounts grounded in interaction and cognition.

This study explores constructs such as the following:
(1) He still regards himself as a working farmer.ARCHER-1979-NEW
(2) We consider it almost one of the necessaries.ARCHER-1875-ADV
(3) He finds the bromide to be most suitable.ARCHER-1864-MED
(4) She takes me for a milkman.ARCHER-1820-DRAMA
These examples share an underlying Argument-Structure-Construction wherein an Attributor attributes an Attribute to an Attributee. In contrast to other Secondary Predication Constructions (D'hoedt & Cuyckens 2017a), this attribution does not alter the state of the Attributee like the “Resultative-Construction”, but is a subjective evaluation which may not be factual (Halliday 1967. Compare: He considers them husband and wife doesn’t mean the couple is actually married, whereas He pronounced them husband and wife confers marital status.). This is why these constructions, also known as “AGENT–ÆFFECTED–JUDGEMENT”-Construction (Herbst & Uhrig 2009), will be termed Evaluative-Attribution-Constructions in this study, as the “evaluative” or “mental” (D’hoedt & Cuyckens 2017b) versions of the Object-Attribute-Construction from Herbst’s and Hoffmann’s “Constructionist Approach to Syntactic Analysis” (CASA, Object Attribute Construction ATTR: NP & ADJ).
Despite semantic similarities, there is considerable formal variation – in other words, considerably different slot fillers – which is the starting point for delineating different members of the EAC family. Aside from the verb slot, there are differences in the Attribute-slot, which is, e.g., a Noun-Phrase as in (1), (2), and (4) or an Adjective-Phrase (3). Moreover, the EAC connects the Attributee with Attribute-slot with either the prepositions as (1) and for (4), without any filler as in (2) or with to be as in (3).
The study, first, argues against modeling the EAC (a) as a single abstract construction and (b) as multiple allostructions due to its formal variability (cf. Cappelle 2006, Colleman 2011, Herbst & Huber 2022, Perek 2015, Goldberg & Jackendoff 2004). Instead, it advocates for a dynamic constructional family with constructions on multiple levels of abstraction. Constructionhood is hereby motivated empirically and usage-based, so only variants with either high frequency (e.g., in comparison with competing constructional candidates on other levels of abstraction) or some degree of idiosyncracy are considered constructions in their own right (Goldberg 2006).
Therefore, the frequency developments of the EAC variants have to be analyzed. This is done with manually annotated data from the ARCHER corpus (1600-1999). In general, the EAC declines in frequency, especially with Germanic verbs like think (1600-49: 313 hits vs. 1950-1999: 69 hits), in contrast to the rise in French loan words like consider as filler of the verb slot (1600-49: 10 hits vs. 1800-49: 70 hits). Aside from these constructional variants with high-frequency verbs, frequently co-occurring fillers of different slots are promising constructional candidates. The mutual influence of verb and “connector” slot is intriguing due to their low type frequency (verbs: 36, connector: 4) and the resulting likelihood of high-frequent collocations. A covarying collostructional analysis confirmed that especially as and for have a strong attraction to certain verbs (e.g., Gries & Stefanowitsch 2004b), whereas to be does not, so the latter does not seem to have lower-level constructions.

Die vorliegende Studie untersucht die Übersetzung strategisch eingesetzter metaphorischer Ausdrücke im chinesischen politischen Diskurs am Beispiel des 14. Fünfjahresplans und dessen offizieller deutscher Version. Politische Metaphern dienen als zentrale kognitiv-diskursive Werkzeuge zur Veranschaulichung abstrakter Konzepte, zur Übermittlung ideologischer Positionen und zur persuasiven Lenkung (s. Schwarz-Friesel 2013, 2015; Skirl/Schwarz-Friesel 2013). Deren interkulturelle Translation stellt eine besondere Herausforderung dar, da sie nicht nur den sprachlichen Ausdruck, sondern auch kulturspezifische Konzeptualisierungen und intentionale kommunikative Funktionen übertragen muss.
Der theoretische Rahmen der Arbeit integriert die Kritische Kognitive Linguistik (KKL) und das translationsorientierte PEIR-Modell (Political Equivalence + Image Re-presentation). Aus KKL-Perspektive werden Metaphern als sprachliche Ausdrücke verstanden, die auf spezifischen Konzeptualisierungsprozessen basieren. Diese Sichtweise erlaubt es, zwischen konventionalisierten und neuartigen Metaphern zu unterscheiden und deren strategischen Einsatz im politischen Diskurs zu analysieren. Das PEIR-Modell (Yang & Zhao, 2020) operationalisiert diese Erkenntnisse für die Übersetzungspraxis. Sein Kernprinzip ist die politische Äquivalenz, d.h. die getreue Wiedergabe der politischen Kernbotschaft und -wirkung. Als zentrale Strategie zur Zielerreichung dient die Bildreproduktion (Image Re-presentation), also der gekonnte Umgang mit den metaphorischen Bildern der Ausgangssprache im Zielkulturkontext.
Auf der Grundlage einer qualitativen Analyse eines parallelisierten Textkorpus identifiziert und systematisiert die Studie die angewandten Übersetzungsverfahren. Es zeigt sich, dass für kulturübergreifend gut verständliche, konventionalisierte Metaphern häufig die Bilderhaltung (z.B. „harte Technologie“, „Wasserturm“) gewählt wird. Bei kulturspezifischeren oder deliberaten Metaphern kommen flexible Strategien zum Einsatz: die Bildkonversion (Ersatz durch ein funktional äquivalentes Zielkulturbild), die Bildergänzung (durch erklärende Zusatzinfomartionen) oder, bei unüberbrückbaren kulturellen Diskrepanzen, der Bildverzichtigung zugunsten einer expliziten Sinnwiedergabe.
Die Studie kommt zu dem Schluss, dass die Übersetzung politischer Metaphern aus KKL-Perspektive ein aktiver Prozess der kognitiven Rekonstruktion ist. Der Übersetzer handelt als kultureller Mediator, der die intentionalen Konzeptualisierungen und kommunikativen Ziele des Ausgangstextes decodiert und unter Ausnutzung geteilter kognitiver Grundlagen sowie unter genauer Abwägung des zielkulturellen Kontexts eine funktional und politisch äquivalente Repräsentation konstruiert. Die Arbeit leistet somit einen Beitrag zur theoretischen Fundierung einer kognitionsbasierten Translationswissenschaft und bietet praxisrelevante Einblicke für die effektive internationale Vermittlung chinesischer politischer Konzepte.

One Frame, Many Uses: Pragmatic Meaning Variation in German and Chinese Gratitude Constructions

Construction Grammar (CxG) has traditionally excelled at modeling stable form–meaning pairings, yet it still lacks a fully systematic account of how pragmatic meaning shifts emerge from entrenched constructions in context (e.g., Leclercq 2020; Nikiforidou 2009; Cappelle 2017; Finkbeiner 2019). This paper addresses the semantic–pragmatic interface in CxG by examining two high-frequency, structurally parallel gratitude constructions: German “Danke für + X” and Chinese “Xie Xie + X”. Drawing on a corpus-based contrastive design, the study investigates how this conventional politeness formulas instantiate a shared Gratitude Frame while also serving as pragmatically flexible resources for stance, interaction management, mitigation, and (less frequently) ironic reversal.

The German dataset comprises 614 analyzable tokens retrieved from the DWDS Die Zeit corpus (journalistic texts, 1946–2024) using transparent query patterns capturing orthographic variants (Danke für, Danke fürs), with exclusion rules to remove false positives (e.g., dafür). The Chinese dataset is based on the BCC newspaper subcorpus, where an initial keyword search for Xie Xie yielded 6289 concordance lines; after systematic cleaning and construction-based segmentation, including both adjacent and non-adjacent complements where a clear dependency is recoverable, the final sample contains 2460 analyzable tokens. All tokens were manually coded for pragmatic function using a shared five-way taxonomy to maximize cross-linguistic comparability.

The analysis reveals robust commonalities in frame-general meaning (thanking for help, support, information, time, and interpersonal virtues), but also clear language-specific preferences. German strongly favors explicit specification of the gratitude object via determiner-rich noun phrases and evaluative modifiers, and it productively exploits negative or politically charged fillers for pragmatic reversal (e.g., Danke für nichts). Chinese, by contrast, more often foregrounds the recipient through pronouns and especially collective address forms, producing a strong newspaper-specific profile of public-oriented thanking. Quantitatively, pragmatic inversions are much rarer in Chinese (0.2%) than in German (7.5%), suggesting that ironic/critical “thanks” are less conventionalized or less acceptable in formal written Chinese media discourse.

To capture these patterns theoretically, the paper proposes a Usage Scenario Model integrating Frame Semantics with CxG. The model distinguishes frame-general meaning (stable core frame elements anchoring constructional identity) from frame-specific scenarios, where peripheral frame elements and discourse cues systematically recalibrate illocutionary force (e.g., stance-taking, mitigation, topic closure, refusal, or irony). The findings demonstrate that gratitude constructions form a gradient continuum between semantic stability and pragmatic variability and illustrate how a frame-based, usage-oriented approach can provide a replicable method for contrastive pragmatic construction research.

References
Capelle, Bert. 2017. What’s pragmatics doing outside constructions? In Semantics and Pragmatics: Drawing A Line, edited by Ilse Depraetere & Raphael Salkie, 115–151. Cham: Springer.
Finkbeiner, Rita. 2019. Reflections on the Role of Pragmatics in Construction Grammar. Constructions and Frames 11 (2): 171–192.
Leclercq, Benoît. 2020. Semantics and Pragmatics in Construction Grammar. Belgian Journal of Linguistics, Volume 34(1): 225–234.
Nikiforidou, Kiki. 2009. Constructional analysis. In Grammar, Meaning, and Pragmatics, edited by Frank Brisard, Jan-Ola Östman and Jef Verschueren,16–32. Amsterdam/Philadelphia: John Benjamins.

Ziel dieses Beitrags ist es, die Relevanz einer adäquaten Klassifikation von Stance-Objekten und deren Bedeutung für die metapragmatische Reflexion in X-Beiträgen zum Thema COVID-19 aufzuzeigen. Ausgangspunkt ist der Stance-Taking-Ansatz von Du Bois (2007), der ein Stance-Objekt als “what the evaluation is about” (S. 149) definiert. Diese Definition bleibt jedoch vergleichsweise vage und sollte vor dem Hintergrund neuerer Arbeiten zur Positionierung (Kiesling 2022; Kiesling et al. 2018; Merten 2023; Spitzmüller 2013) präzisiert werden. Der Beitrag argumentiert, dass eine differenzierte Klassifikation von Stance-Objekten notwendig ist, um unterschiedliche Objekt-Ebenen im metapragmatischen Paradigma analytisch erfassen zu können.
Anhand eines Korpus von 558 X-Beiträgen von Christian Drosten (Institut für Virologie, Charité), Hendrik Streeck (Uniklinikum Bonn), Tobias Ulbrich (Sozietät Dr. Rogert Ulbrich & Partner), Friedrich Pürner (Europäisches Parlament) sowie institutionellen Akteuren (Charité, Robert Koch-Institut) wird gezeigt, dass drei Typen von Stance-Objekten zu unterscheiden sind. Dazu zählen erstens Themen, die allgemeine Diskursgegenstände darstellen (z. B. COVID-19-Impfstoffe, Infektion), zweitens Ansprüche, die strittige Aussagen formulieren (z. B. Maßnahmen beenden und zurück zur Normalität. Jetzt.), und drittens Frames, die konzeptionelle Rahmungen für das Stance Taking bereitstellen (z. B. Gesundheitskrise).
Die Analyse umfasst Beiträge aus dem Zeitraum von März 2020 bis September 2023. Zur Identifikation und Klassifikation der Stance-Objekte wird Named Entity Recognition eingesetzt. Die automatisierte Annotation erlaubt es, wiederkehrende Referenzen auf Personen, Institutionen, Ereignisse und Konzepte systematisch zu erfassen und den drei Typen von Stance-Objekten zuzuordnen. Die quantitativen Ergebnisse werden durch qualitative Interpretationen ergänzt.
Die vorgeschlagene Differenzierung ermöglicht es, unterschiedliche Dimensionen des Stance Takings präziser zu erfassen, etwa das indexikalische Verhältnis zwischen Personen- und Verhaltenstypen, die im metapragmatischen Paradigma an bestimmte Stance-Objekte gebunden sind (vgl. Spitzmüller 2013). Während Themen als explizite Diskussionsgegenstände in den Beiträgen auftreten, werden Ansprüche argumentativ gestützt oder bestritten und fungieren als Übergang zu Frames, die übergeordnete Deutungsrahmen etablieren. Diese drei Typen von Stance-Objekten werden daher als wechselseitig konstitutive Eckpunkte verstanden, die die metapragmatische Gestaltung von Positionierungen im digitalen Raum strukturieren.
Abschließend zeigt der Beitrag, dass die systematische Klassifikation von Stance-Objekten einen zentralen Baustein metapragmatischer Diskursanalysen darstellt. Sie macht sichtbar, auf welchen Ebenen Stance-Objekte im digitalen Raum verhandelt werden und wie digitales Stance Taking selbst zum Gegenstand metapragmatischer Reflexion wird.

This paper applies methods of Cognitive Linguistic Analysis, specifically Blending Theory and Conceptual Evolution Theory to “Conspiracy Theories” (CTs). First, I outline the conceptual elements of CTs and their function as building blocks for a narratively presented pseudo-argument. Instead of a pragmatic/logical progress from premises to conclusion, CTs project a figurative conspiracy story onto a crisis or catastrophe, which they suspect to be ‘covered up’ by official or ‘mainstream’ discourse.
Although the two basic elements of CTs (i.e. the crisis account and the conspiracy backstory) are shrouded in speculation, they are presented so as to confirm each other, thus providing a seemingly perfect explanation of the crisis and its preferred solution. Blending analysis shows how an integration of diverse input CTs produces emergent blended spaces as well as action scenarios that motivate believers to engage in political activism. As an example, I analyse the integration of logically incongruous mental spaces in the COVID-19-related CTs, on the basis of a corpus of English (US American and British) and German media data. In particular, I show that the CTs’ narrative logic – or rather, coherence – demands the filling of a ‘traitor’ slot, irrespective of evidence or plausibility.
Conceptual evolution theory highlights CTs’ adaptive potential to seemingly ‘fit’ new cultural environments and circumstances. Building on the above examples of COVID-19 CTs, I analyse a special case of their adaptation to the German political context, i.e. the use of these CTs by right-wing parties and extremist groups (AfD, Reichsbürger) as ‘proof’ for another set of nationalist CTs about Germany’s alleged “enslavement” by a “COVID-dictatorship”. Although the various input CTs are logically and historically unrelated, they were successfully integrated by CT propagators into a conceptual scenario that motivated parts of these groups to attempt the “storming” of the parliament to start a popular uprising against the perceived traitor government.
This astounding, if short-lived, success of a blending of multiple CTs in triggering violent political activism illustrates the importance of their adaptability to diverse cultural and historical environments, and it throws up the question of how they can be countered in public discourse, beyond the heuristic target of content-oriented “fact-checking”. It seems necessary to develop a holistic cognitive approach that is able to deconstruct the apparent narrative coherence of CTs and identify the conditions for their construal into action-leading and identity-building scenarios in new cultural environments.