This talk explores intra-speaker malleability in the realization of the first person possessive in the North East of England ([maɪ], versus [mi] and [mɑ]). Previous apparent time research on the possessive in Northern England has shown age-graded effects – older speakers realising higher rates of standard [maɪ] – which have been interpreted as a corollary of linguistic marketplace pressures (Childs 2013). The scociolinguistic enterprise assumes that post-adolescent speakers exhibit retrenchment towards the standard in middle age (Downes 1984), resulting in a U-shaped curve. My analysis relies on a panel corpus that covers the entire adult lifespan, providing novel insights into how individual speakers’ longitudinal linguistic choices pattern relative to observed community trends.
The dynamic set-up allows me to explore the workings of the linguistic marketplace, which to date have been mainly theorized based on apparent-time data or panel datasets covering short snippets of the lifespan (see Buchstaller and Beaman 2022). My data-set relies on 30 speakers aged 22-79, who are representing a wealth of socio-demographic trajectories and who were recorded two (or three) times at five-year intervals. I report on an analysis of more than 3000 tokens of the first person possessive that were analysed in R (Core Team 2022) using a variety
of statistical methods.
My data supports Sankoff et al.’s (1985) claim about the lifespan impact of pressures felt by professionals of the language, whose “lives and work would require them to have access to the language variety legitimized by the dominant ideology” (p. 108). Some speakers, however, do not converge with the traditionally assumed age grading as marketplace pressures model (Wagner 2012): Those in working-class occupations retain high rates of the reduced variants and there is little evidence of them changing towards the prescribed norm. In contrast, it seems as if they use the variant to project a locally authentic persona and to index a sense of belonging by foregrounding their roots in the local community (cf. Sundgren et al. 2021).
Overall, my research provides evidence of socially and linguistically differentiated patterns when speakers are traced across the entire adult life-span. The nuanced panel corpus reveals that the hypothesized U-shaped curve is an oversimplification of the complexities involved in change across the life-span. To paint a socially realistic model of linguistic malleability across the adult lifecourse we need to ask what being an adult means in terms of prescriptive pressures, presentation of self and social networks (Milroy & Llamas 2013, Tetreault 2017).

Statistical learning in language acquisition was originally conceived as a gateway to speech segmentation in the absence of pre-existing knowledge about the language to be acquired. Initial support for this hypothesis came from corpus analyses that showed that Transitional Probabilities (TPs) tend to be higher within words than at word boundaries and that adults and infants can segment continuous streams of stimuli in various perceptual domains by calculating TPs (Saffran et al., 1999; Kirkham et al., 2002; Pelucchi et al., 2009). However, recent corpus studies have questioned the universality of TPs by showing that the most successful statistical segmentation strategy for a given language depends on the specific language under question. While Italian child-directed speech (Verb-Object language) is best segmented with Forward-TPs (FTPs), Hungarian (OV language) is best segmented with Backward-TPs (BTPs) (Gervain & Guevara, 2012; Saksida et al., 2016). This suggests that speech segmentation using TPs may depend on language-specific knowledge. Here we argue that TPs in the linguistic input do not only depend on the specific language to be acquired, but also on the age of the child, the input is directed to. We argue that
linguistic input to infants (IDS) or children (CDS) differs from that of adults (ADS) at the prosodic, phonological and syntactic level (Cristia, 2013).
To investigate how speech register influences statistical regularities in the input, we compared IDS and CDS from 7 languages (German, Estonian, Italian, Dutch, English, Greek, Hungarian) to children of varying ages (range 1.5 to 75 months). We compared the segmentation performance of FTPs and BTPs using a relative algorithm (word boundaries are assigned when TPs between two syllables are lower than TPs of surrounding syllable pairs) and an absolute algorithm (word boundaries are assigned when TPs drop below a fixed language-specific threshold). We find that while the type/token ratio of syllables remains constant (ß=0.019, SE=0.014, t=1.32, p=.186), the type/token ratio of words increases (ß=0.04, SE=0.012, t=3.04, p <.001) as children become older. The increase of different word types as children become older increases the number of ways existing syllable types are combined and decreases the overall value of TPs (ß=-0.4, SE=0.1, t=-7.23, p< .001). Furthermore, the performance (i.e., F-score of segmented words) of the absolute algorithm increases in language-specific ways. In Object-Verb languages (Dutch, Hungarian) segmentation with BTPs improves with age, outperforming FTPs as children get older and in Verb-Object languages (English, Estonian, Italian, Greek) FTPs improve with age outperforming BTPs as children get older. The results show that the statistical make-up of linguistic input changes as children get older, with better segmentation observed with language-specific and not universal statistical regularities.

The aim of this research is to juxtapose and evaluate competing code-switching theoretical frameworks by examining their predictions on morphosyntactic integration. In this study, two main approaches are compared, namely, Poplack’s (1998, 2018) Nonce Borrowing Hypothesis (NBH) and Myers-Scotton’s (1993, 2009) Matrix Language Framework (MLF) model. These models’ predictions on morphosyntactic integration are compared by an analysis of Hungarian-English code-switching examples. In case the predictions match or differ from actual manifestations, we can get closer to an objective claim about which model is superior if their number of “successful” predictions is different.
The examples are taken from informal podcast talks in Hungarian contexts. The participants are communicating in Hungarian with the occasional use of English. By analysing the integration of the English words or phrases we can shed light upon the validity of the above mentioned models’ explanations.
According to the NBH, language contact takes place between the donor and the recipient language, where the donor language provides free morphemes or lexical items that preserve their original grammatical features (word order as well) in the case of code-switching. As for the word order it is unsure whether the rule applies only for the code-switched elements from the donor language or for the whole clause/ sentence. Furthermore, the question of how the identification of the recipient and donor language happens is also unanswered. According to Poplack (2018), during code-switching, usually more than one word is inserted into the recipient language. As claimed by the theory, the case should be considered as borrowing, when (mostly) a single (lone) item is inserted into the other language, this item adapts to the grammar of the recipient language, thus the elements of the donor language behave as they were of the recipient language morphosyntactically. In this research the recipient language is Hungarian, whereas the donor language is English.
According to the MLF model, the matrix language provides the grammar for the embedded language items. The main difference between this model and the one mentioned above is that this model considers all manifestations code-switching where items are inserted from one language to the other and the identification of the matrix language is well defined (presence of system-morphemes, number of morphemes is usually higher).
In the present study, I present the predictions of the two theories and their approach to the different language contact manifestations. The following example sentence and some explanations below are an illustration of attempting to solve the research problem:
Example (1): ” Mindenki nagyon clean akar lenni.”
Everyone.PRON very.ADV clean.ADJ want.V be.AUX
“Everyone wants to be very clean.”
According to Poplack’s view: Hungarian is the recipient language; English is the donor language; the lone donor language item [clean] shows that it is a borrowing; etc.
According to Myers-Scotton’s view: Hungarian is the matrix language; English is the embedded language; in this code-switching process there is no surface integration, however the syntactic functions are parallel; etc.
By the concordance (if there is any) of the predictions and the manifestations we can draw the conclusion of one model being superior to the other in terms of examining the phenomenon.

En este trabajo, expondré los avances realizados hasta el momento de la monografía «Representaciones discursivas en Textos Políticos y Guitarra Negra de Alfredo Zitarrosa» enmarcado en el seminario «Discurso, poder y medios», dictado en 2021, de la Licenciatura en Lingüística de la Facultad de Humanidades y Ciencias de la Educación. El objetivo general del proyecto es estudiar cómo se representa discursivamente a los actores sociales de la última dictadura cívico-militar uruguaya en dos álbumes de Alfredo Zitarrosa. Los objetivos específicos son los siguientes: 1) describir las representaciones discursivas de los actores sociales que aparecen en estas obras; 2) analizar las representaciones discursivas de estos actores, y 3) caracterizar las representaciones discursivas de las acciones sociales que estos actores realizan en las obras. La metodología consiste en el estudio del léxico y de la sintaxis de las canciones pertenecientes a los álbumes Textos Políticos (1980) y Guitarra Negra (1985), en aras de analizar los actores y las acciones sociales que aparecen creados por el autor, a través de la propuesta teórica conformada por Van Leeuwen (1996; 1995). El fundamento principal para seleccionar estos álbumes radica en que ambos son elementos artísticos que reflejan un episodio fundamental de la historia uruguaya por medio de la óptica de uno de sus autores más importantes, como lo es, en este caso, la figura de un cantautor que vivió el inicio de la caída de la democracia hasta el retorno al país luego de su exilio. Las canciones que han sido analizadas, a la fecha de este resumen, han sido las pertenecientes a Textos Políticos (1980). Hasta el momento, los actores sociales representados en este álbum son actores sociales relacionados con el gobierno (presidentes, militares y empresarios); actores sociales no relacionados con el gobierno (movimientos sociales, sindicatos, el pueblo y trabajadores de diversos rubros), y Alfredo Zitarrosa — representado como un actor social que se posiciona como parte de los actores no relacionados con el gobierno en algunas obras y como un sujeto individual en otras. Los mecanismos de representación mayormente empleados han sido la nominalización, la categorización, la indeterminación y, en el caso de Alfredo Zitarrosa, la sobredeterminación (overdeterminaton).

Preregistration is the practice of systematically documenting the methodological decisions involved in the design of a study. As the prefix -pre suggests, such documentation must occur prior to any data collection or analysis. Importantly, preregistering a study is one of the practices put forward by the open science movement. To put it briefly, open science is an umbrella term that encompasses a series of proposals for the production and dissemination of scientific knowledge. A primary concern of the movement is to advocate the adoption of more transparent, rigorous, and reproducible practices in research. In this context, preregistration is an important element for improving the robustness of scientific results due to its potential of reducing the occurrence of three well known problems in the scientific field: (i) publication bias, (ii) analytical flexibility, and (iii) the distinction between
exploratory and confirmatory research. Publication bias refers to the fact that more studies are conducted than published and that there is a preference for publishing positive results; thus, the peer-reviewed literature is biased towards positive and novel findings. Analytical flexibility refers to the researchers’ degree of freedom on the use of statistical procedures and data transformation, as well as, on the selective report of confirmatory results. Consequently, research outcomes may be misleading due to the possible presence of type I or type II errors. Finally, the distinction between exploratory and confirmatory research is important to facilitate the self-correctness process of science and to avoid the promotion of a distorted model of science. Considering such problems often present in the research cycle, the practice of documenting the methodological decisions of a study before the data collection or analysis has the potential to reduce publication bias, to avoid analytical flexibility, and to distinguish exploratory from confirmatory research. Having this in mind, in this work we argue that the language sciences would also benefit from the practice of preregistration. Considering the complexity of our object of study, linguists investigate language and its social manifestations through different perspectives and methodologies. Therefore, research on human language needs to be transparently and rigorously conducted. This is why we believe preregistering research in the language sciences is essential to assure the robustness of our findings and the impact that the knowledge we produce has in our society.

Embora línguas como o espanhol e o alemão tenham duas conjunções adversativas distintas (pero e sino; e aber e sondern, respetivamente), o português tem apenas o morfema mas para veicular os valores de correção e de negação de expectativas. Tendo em conta as várias construções em que mas pode ocorrer e a dificuldade notada por Canceiro (2019, 2021) em estabelecer uma ligação clara entre as construções e os valores de mas, analisaremos neste trabalho os valores associados a frases como (2)-(5) e as configurações sintáticas que lhes estão subjacentes. Quanto a frases como (1), que envolvem coordenação frásica, é geralmente aceite que o seu valor é de negação de expectativas. Os casos em (2)-(5), por outro lado, não são consensuais e serão o alvo da nossa análise.
(1) Está a chover, mas o João vai à praia.
(2) O João não deu um livro à Maria, mas à Ana.
(3) O João deu um livro à Ana, mas não à Maria.
(4) O João deu um livro não à Maria, mas à Ana.
(5) O João deu um livro à Ana, mas à Maria não.
Dada a dificuldade em associar os dois principais valores de mas a estas construções, construímos uma Tarefa de Compatibilidade Contexto-Frase (Escolha Forçada), em que os participantes têm de avaliar qual dos contextos ((6) para correção, ou (7) para negação de expectativas) é mais adequado às construções (2)-(5), ou alternativamente se os dois são adequados.
(6) Alguém foi corrigido por dizer que o João tinha dado um livro à Maria.
(7) Esperava-se que o João tivesse dado um livro à Maria e isso não aconteceu.
A tarefa foi apresentada na plataforma desenhada por Lourenço-Gomes (2018) e os resultados do teste piloto mostraram uma preferência, em todas as construções, pela interpretação de negação de expectativas (a recolha de dados está em andamento). Os dados obtidos com a tarefa irão permitir perceber se a conexão assumida por alguns
autores entre os valores da conjunção e as configurações sintáticas é robusta. Em relação a frases com valor corretivo, por exemplo, Vicente (2010) propõe uma análise de coordenação frásica e elipse, enquanto Toosarvandani (2013) e Steindl (2017) assumem que embora seja possível adotar a análise de Vicente, não existem argumentos suficientemente fortes para excluir uma proposta de coordenação de constituintes sub-frásicos.
A nossa análise para o PE assume que não existe uma ligação direta entre o valor semântico e a configuração sintática, embora tipicamente a coordenação de sintagmas esteja associada ao valor corretivo. Assim, apresentaremos evidência para que as frases (2)-(5) sejam analisadas da seguinte forma: (i) para frases com negação frásica (cf. (2)) assumimos a possibilidade de coordenação frásica + elipse (no segundo termo), ou coordenação de dois PPS, sendo que apenas o primeiro está sob escopo da negação, uma vez que mas bloqueia o escopo para o segundo termo; (ii) frases como (3) e (4), por corresponderem a casos de Contraste Sintagmático, serão analisadas como coordenação de sintagmas (sem elipse), mesmo quando os sintagmas são CPs; por fim, (iii) frases como (5) correspondem a casos de Elipse de TP com partículas de polaridade (cf. Saab, 2008; Matos, 2017), ou seja, coordenação frásica + elipse.
Este trabalho pretende, assim, fornecer uma análise da coordenação adversativa, tendo em conta dados experimentais que clarificarão valores semântico/pragmáticos, e apresentar configurações sintáticas que reflitam as diferentes propriedades destas construções, considerando ainda a hipótese de existirem dois mas em PE.

Os sentidos não-normativos de gênero e de sexualidade sempre foram colocados num lugar de marginalidade pelos sentidos hegemônicos que constituem a nossa sociedade e, quando alinhados a outros sentidos que desafiam o corpo social branco-cêntrico e cis-heteronormativo, como os sentidos racializantes e (homo)sexualizantes, o lugar ao qual é destinado é o de invisibilidade, de apagamento, de silenciamento. Nesse sentido, sabedores do quão urgente é refletir e discutir sobre o que, durante muito tempo, foi silenciado, buscamos, através deste trabalho, compreender como
se interseccionam as discursividades de gênero, de sexualidade e de raça. Para o desenvolvimento desta pesquisa, ancoramo-nos, como suporte teórico-analítico, na Análise de Discurso de filiação materialista, tomando como premissa as ideias de Michel Pêcheux ([1975] 2009) e de Eni Orlandi (2000, 2012), nas teorias de gênero e de sexualidade (BUTLER, 2006, 2015; FOUCAULT, 1969) e nas reflexões sobre as discursividades racializantes (MODESTO, 2018), considerando o caráter interseccional dos objetos de pesquisa, isto é, o entrecruzamento das discursividades de gênero, raça e sexualidade. Mobilizamos, como material de análise, a produção cinematográfica Madame Satã (2002), com o intuito de, por meio do movimento pendular entre descrição e interpretação: i) identificar as intersecções constitutivas dos sentidos de/entre gênero, raça e sexualidade nas quais se inscreve a personagem da obra supracitada; ii) analisar a relação de sentidos entre essas discursividades na produção fílmica
selecionada e iii) compreender como esse funcionamento discursivo significa as relações sócio-históricas nessa produção, ao mesmo tempo em que forja a posição-sujeito dissidente da masculinidade negra (e) gay. Como desenvolvimento desta pesquisa, analisamos que a masculinidade negra é forjada, nas duas obras, pela subversão dos sentidos hegemônicos de gênero e sexualidade alinhados aos de raça. Através da materialização dos sentidos de gênero e sexualidade em não conformidade com as expectativas da concepção biologizante de corpo-gênero, numa formação social branco-cêntrica, nas corpos-subjetividades homem negro e gay, surgiram os gritos de autoafirmação e autodeterminação que constitui esse sujeito.

The goal is to show the design of a digital portal of written corpora, mainly uses and descriptions of verbal predication in Portuguese national varieties, as well as their comparison with some Romance language varieties. The proposal deals with the life cycle of such database and its research team. It integrates the plan of scientific interests and actions in theoretical, descriptive-comparative and applied linguistics of a Franco-Brazilian project. The directive is to solve the problem of accessing data and qualified descriptions of expression of states of affairs via verbal predications in Romance languages and their varieties, especially in workspaces with Portuguese (non-) mother language.
It is hypothesized that there are convergences and divergences to be detected based on the comparison between languages ​​and varieties of Romance languages ​​that can feed national and international work bases related to certain communicative practices in the hyperconnected world, such as teaching, translation, interpretation, dubbing. The plausibility of this hypothesis is evident in the fact that, in Brazil, the use of foreign digital data platforms is increasingly growing and in the fact that many predication constructions are not yet considered in practices and Portuguese teaching materials (among them, constructions with support verb). The methodological strategy is based
on the guidelines of empirical analysis of uses/corpora of Sociolinguistics and Construction Grammar, which weigh a network of formal and functional factors (in these, semantic, discursive, pragmatic, cognitive and sociocultural ones). The aim is to deliver the digital portal to scientific and non-scientific publics.
By multivariate and collostructional analyses of datasets of Romance languages, we focus on the problem of diasystemicity based on the detection of the Romance languages ​​and their varieties´ stable and common constructional patterns of predication (diaconstructions) and their particular patterns (idioconstructions), as well
as of their internal constructional patterns that align by similarity (allostructions/constructional variants) in an area in the constructional network (metaconstruction) resulting from analogy and, then, from the process of neutralizing the attributes that, in other contexts, serve to differentiate them. We deal with the relation between these theoretical-descriptive generalizations about perspective and representation predications and spaces of Portuguese as (non-) mother language.

A breve apresentação é um recorte da tese intitulada O pensamento linguístico de João de Barros e as escolas de ler e escrever na América portuguesa quinhentista, (LOURENÇO, 2022). Esta, inserida no campo da Historiografia da Linguística (HL) e da Linguística Missionária (LM). Dentre seus objetivos, destaca-se o de analisar o pensamento linguístico de João de Barros, em relação à alfabetização latino-portuguesa humanística. As obras Cartinha com os Preceitos e Mandamentos da Santa Madre Igreja e Gramática da língua portuguesa, ambas de João de Barros (1539; 1540, respectivamente), são prováveis fontes para o ensino jesuítico na colônia, assim como influentes no clima intelectual do recorte temporal compreendido entre 1549 e 1556. Autores e pesquisadores como Pierre Swiggers (2009, 2013), Ronaldo Batista (2013, 2016, 2019) e Otto Zwartjes (2011) compuseram parte do aporte teórico-metodológico da pesquisa, auxílio imprescindível em reflexões que passaram pelos pressupostos da HL – tais como os parâmetros de cobertura, perspectiva e profundidade – e da linha de pesquisa da LM. Em uma análise qualitativa de fontes documentais, foi possível comprovar, em nossa investigação e através dos resultados obtidos, o desenvolvimento de instituições de ensino da época do período missionário, no território que viria a ser o Brasil; assim como a implantação do vernáculo português; a alfabetização intercultural, de base latino-portuguesa, pela catequese; e a gramatização da língua indígena.

It is urgent to gather scientific efforts around the transdisciplinary construction and management of a national online repository of our multilingual and multicultural heritage. This is one of the proposals of the ABRALIN Project and the ANPOLL Sociolinguistics WG entitled Platform of Brazilian Linguistic Diversity, which involves researchers from different areas of knowledge and partnerships with other institutions, such as the Portuguese Language Museum. In this space, we want to highlight the relationship of the construction and implementation of this platform to the area of Digital Humanities. The Digital Humanities consist of a relatively new scientific territory, arising from the intersection between the Humanities, Social Sciences and Digital Technologies. This challenging contemporary way of producing science involves the production, storage and circulation of knowledge in the same
digital environment. These parameters are in line with the GO (Global Open) FAIR BRASIL movement, an initiative that aims to make fragmented and disconnected data Findable, Accessible, Interoperable and, therefore, Reusable (FAIR) by machines and people. These discussions involve ethical issues and work horizons for open, diverse and socioculturally referenced science and education via mapping of languages and varieties in/from Brazilian territory. Thus, in this event, we aim to share with society such reflections and potentialities of (re)use of linguistic data and knowledge produced in the Brazilian territory. These discussions are also aligned with Open Science, a movement that proposes structural changes in the way scientific knowledge is produced, organized, shared and reused. It is a new way of doing science, more collaborative, transparent and sustainable.

The title of this paper alludes to the title of Heimann’s (2009) influential book on the (unsuccessful) history of Czechoslovakia. In this paper, I will focus specifically on one aspect of this failed state, namely on its failed project of Czechoslovak language. I will apply the concept of Language Making (Krämer, Vogl & Kolehmainen LMLL) and present the Czechoslovak language as one of the manifestations of the state ideology of Czechoslovakism (Hudek, Kopeček & Mervart LMLO).
Founded in the aftermath of the First World War in 1918, the First Czechoslovak Republic was, like its predecessor Austria-Hungary, a multiethnic, multiconfessional, and multilingual state. Beside the two titular nations of Czechs and Slovaks, large minorities of Germans, Hungarians, Ruthenians, and Poles also lived in the new republic. Nevertheless, the Czechoslovak Constitution, adopted on 29 February 1920, established Czechoslovakia as the nation state of Czechoslovaks. Not of Czechs and Slovaks, but of Czechoslovaks. Following this Czechoslovakist principle, the Language Act (122/1920 Sb.), adopted on the same day as the constitution itself, stated that the state and the official language of the republic is the Czechoslovak language and Czech and Slovak were declared to be its two branches, just as Czechs and Slovaks were viewed to be the two branches of a single Czechoslovak nation.
I argue that this act constitutes an example of Language Making, defined as the “conscious […] human processes in which imagined linguistic units are constructed and perceived as a language, a dialect or a variety” (Krämer, Vogl & Kolehmainen LMLL: d). In this case, Czech and Slovak as two separate West-Slavic languages stopped being perceived, at least officially, as two distinct units, but were merged, following the one state-one nation-one language principle, into one discursive unit.
Just as the Czechoslovak language was made in February 1920, it was unmade in May 1948. The newly adopted post-war constitution (150/1948 Sb.) avoided any reference to the Czechoslovak nation or to the Czechoslovak language. Instead, Czechoslovakia was re-defined as the nation state of two separate, although “brotherly” nations: Czechs and Slovaks. The Czechoslovak language, discursively made by being given a name in 1920, was effectively unmade by not being mentioned anymore (Krämer, Vogl & Kolehmainen LMLL: Oh).
In sum, this paper uses the examples of Czechoslovakia, the Czechoslovak nation, and the Czechoslovak language to illustrate the interplay of political history and linguistic reality and to show that just like states, also languages can fail.

When perceiving a sentence like “the mother gave the candle the daughter”, we must infer whether the intended message was (1) the literal, implausible one, or (2) the plausible alternative “the mother gave the candle to the daughter” but the word “to” dropped during transmission over a noisy channel. Such inferences rely on assumptions about the channel, the language, and likely meanings, but also about the producer’s identity (e.g., their tendency to misspeak or refer to different events). We tested whether noisy-channel comprehension is influenced by a speaker’s
perceived gender, given that gender stereotypes influence many cognitive processes. Do implausible messages made by men vs. women differ in how often they are interpreted non-literally as plausible alternatives?
We used the stimuli and paradigm from Gibson et al. (2013, PNAS). Participants (N=365 on Prolific) read 20 implausible sentences (60 fillers) each presented alongside a question (“did the daughter receive something/someone?”). Their answer indicated whether they interpreted the sentence literally. Before the experiment, participants saw a photo of either a man or a woman (out of 8 possible photos) with text introducing them as the research assistant running the study. The photo remained visible throughout the experiment with the text “this is my experiment”. Participant gender was crossed with the producer’s gender (“man photo” condition: 97 men, 84 women; “woman photo” condition: 101 men, 83 women).
To replicate a central prior finding, each implausible item had two versions: a double-object (“the mother gave the candle the daughter”, suggesting an omission of “to”) and a prepositional-object (“the mother gave the daughter to the candle”, suggesting an addition of “to”) construction. Each participant read 10 items of each kind. Because omissions are more likely than additions, non-literal interpretations should be more frequent for DO vs. PO sentences.
Interpretation (literal vs. non-literal) was modeled via a mixed-effects logistic regression with the maximal converging structure. Fixed effects included participant gender, producer (photo) gender, construction, and all interactions.
Replicating prior findings, participants made more non-literal inferences for DO than PO sentences (b=1.47, SE = 0.12, c2 (1)=15.5, p<10-4). Critically, a man photo led to more non-literal interpretations compared to a woman photo (b=-0.79, SE=0.23 c2 (1)=11.8, p<10-3). Thus, when producing implausible sentences, men are “given the benefit of the doubt” more often: their implausible utterances are interpreted non-literally as plausible alternatives. We avoid drawing conclusions about the reason (perhaps participants think men are more likely to mistype; perhaps they think women are more likely to say implausible things; etc.). Regardless of the reason, we demonstrate a gender bias in noisy-channel comprehension.