Search

The extent to which gender neutral and gendered nouns impact differently upon native French speakers’ gender representations was examined through a yes-no forced choice task. Swiss (Experiment 1) and Québec (Experiment 2) French-speaking participants were presented with word pairs composed of a gendered first name (e.g., Thomas) and a role (e.g., doctor), and tasked to indicate whether they believed that [first name] could be one of the [role]. Roles varied according to gender stereotypicality (feminine, masculine, non-stereotyped), and were either in a plural masculine (interpretable as generic) or gender neutral (epicenes and group nouns) form. The results indicated that the use of gender neutral forms of roles avoided a strong male bias found for the masculine forms, and that both gender neutral and masculine forms used equal cognitive resources. Further, stereotype effects associated with both gender-neutral and grammatically masculine forms were quite small (

When writing a text, the choice between seemingly equivalent connectives, such as but and however, is not arbitrary: aside from preferences linked to the register or modality of the utterance, there are often pragmatic or syntactic constraints that limit the choice to only one particular connective. For French, theoretical approaches have extensively discussed the syntactic and pragmatic constraints that allow only either parce que (‘because’) or puisque (‘since’). However, experimental findings regarding the actual sensitivity of French speakers to these constraints remain inconclusive. In the current study, we examine the sensitivity of different speaker groups to the constraints associated with the use of parce que and puisque. In a controlled offline-task, we systematically evaluate the sensitivity of a crowd-sourced heterogeneous group of native speakers (Experiment 1) and a group of non-native speakers (Experiment 2). Results show that native speakers are more sensitive to syntactic constraints whereas pragmatic constraints are more equivocal. Nonnative speakers, however, are not sensitive to either of these constraints. In addition, the pragmatically complex connective puisque is extremely problematic for non-native speakers, as even the mapping of a seemingly equivalent connective of their L1 did not warrant a good mastery of its pragmatic functions.

While the role of discourse connectives has long been acknowledged in argumentative frameworks, these approaches often take a coarse-grained approach to connectives, treating them as a unified group having similar effects on argumentation. Based on an empirical study of the straw man fallacy, we argue that a more fine-grained approach is needed to explain the role of each connective and illustrate their specificities. We first present an original corpus study detailing the main features of four causal connectives in French that speakers routinely use to attribute meaning to another speaker (puisque, étant donné que, vu que and comme), which is a key element of straw man fallacies. We then assess the influence of each of these connectives in a series of controlled experiments. Our results indicate each connective has different effects for the persuasiveness of straw man fallacies, and that these effects can be explained by differences in their semantic profile, as evidenced in our corpus study. Taken together, our results demonstrate that connectives are important for argumentation but should be analyzed individually, and that the study of fallacies should include a fine-grained analysis of the linguistic elements typically used in their formulation.

A straw man is a form of fallacious argument that involves the distortion of an opponent's view in order to make it more extreme and therefore less acceptable, thus easier to attack. In three experiments, we assess linguistic factors that may influence the acceptability of straw men for addressees. In Experiment 1, we demonstrate that a straw man is more acceptable when the speaker's argument is attacked rather than his standpoint. In Experiment 2, we show that a straw man is more acceptable when it contains two arguments that are simply juxtaposed rather than linked by a causal connective indicating attributed content (i.e. the French puisque). In Experiment 3, we find that a straw man is more acceptable when it echoes the speaker's explicit rather than implicit meaning. Taken together, these experiments strongly indicate that the formulation of a straw man has an impact on its perceived acceptability. Our results thus open a new avenue of inquiry for studies assessing the role of linguistic factors on the acceptability of fallacies.

In this paper, we show how three often used and seemingly different discourse annotation frameworks – Penn Discourse Treebank (PDTB), Rhetorical Structure Theory (RST), and Segmented Discourse Representation Theory – can be related by using a set of unifying dimensions. These dimensions are taken from the Cognitive approach to Coherence Relations and combined with more fine-grained additional features from the frameworks themselves to yield a posited set of dimensions that can successfully map three frameworks. The resulting interface will allow researchers to find identical or at least closely related relations within sets of annotated corpora, even if they are annotated within different frameworks. Furthermore, we tested our unified dimension (UniDim) approach by comparing PDTB and RST annotations of identical newspaper texts and converting their original end label annotations of relations into the accompanying values per dimension. Subsequently, rates of overlap in the attributed values per dimension were analyzed. Results indicate that the proposed dimensions indeed create an interface that makes existing annotation systems “talk to each other.”

Discourse connectives are lexical items indicating coherence relations between discourse segments. Even though many languages possess a whole range of connectives, important divergences exist cross-linguistically in the number of connectives that are used to express a given relation. For this reason, connectives are not easily paired with a univocal translation equivalent across languages. This paper is a first attempt to design a reliable method to annotate the meaning of discourse connectives cross-linguistically using corpus data. We present the methodological choices made to reach this aim and report three annotation experiments using the framework of the Penn Discourse Tree Bank.

Psycholinguistic investigations of the way readers and speakers perceive gender have shown several biases associated with how gender is linguistically realized in language. Although such variations across languages offer interesting grounds for legitimate cross- linguistic comparisons, pertinent characteristics of grammatical systems – especially in terms of their gender asymmetries – have to be clearly identified. In this paper, we present a language index for researchers interested in the effect of grammatical gender on the mental representations of women and men. Our index is based on five main language groups (i.e., grammatical gender languages, languages with a combination of grammatical gender and natural gender, natural gender languages, genderless languages with few traces of grammatical gender and genderless languages) and three sets of specific features (morphology, masculine- male generics and asymmetries). Our index goes beyond existing ones in that it provides specific dimensions relevant to those interested in psychological and sociological impacts of language on the way we perceive women and men. We also offer a critical discussion of any endeavor to classify languages according to grammatical gender.

Coherence relations linking discourse segments can be communicated explicitly by the use of connectives but also implicitly through juxtaposition. Some discourse relations appear, however, to be more coherent than others when conveyed implicitly. This difference is explained in the literature by the existence of default expectations guiding discourse interpretation. In this paper, we assess the factors influencing implicitation by comparing the number of implicit and explicit translations of three polysemous French connectives in translated texts across three target languages: German, English and Spanish. Each connective can convey two discourse relations: one that can easily be conveyed implicitly and one that cannot be easily conveyed implicitly in monolingual data. Results indicate that relations that can easily be conveyed implicitly are also those that are most often left implicit in translation in all target languages. We discuss these results in view of the cognitive factors influencing the explicit or implicit communication of discourse relations.

Connectives are complex lexical items for children, who do not even master connectives used frequently in speech until the end of their primary school years. In this paper, we assess the ability of two groups of 16-year-olds from two different academic levels (i.e., high school vs. professional school students), as well as a group of university students, to use four French connectives bound to the written mode in a constrained production task. These connectives were chosen to assess three potential sources of difficulties: the cognitive complexity of the encoded discourse relation, the connectives’ frequency in corpus data, and their mono- or polyfunctional nature. Our results indicate that teenagers have a lower ability to use these connectives compared to university students, and frequency appears to be the best-suited factor to account for their difficulties. In addition, the teenage group with a higher academic level reached better accuracy with the two more frequent connectives tested. Taken together, our results indicate that teenagers do not yet have mastery over the wide array of connectives from the written mode in their native language, particularly those that are less frequent in corpus data. These connectives are especially problematic for teenagers with a lower academic background.

Discourse connectives are often reported to be difficult for second language learners, yet the causes of these difficulties are still not fully understood. In this paper, we test the ability of German-speaking learners to process and understand a connective with a complex form-function mapping in their L2-French, namely ‘en effet’, a connective that does not have an exact translation equivalent in their L1-German. We assess learners’ competence both in an on-line processing experiment and an off-line judgment task. We argue that one of the interesting specificities of ‘en effet’ is that the two coherence relations that it conveys cannot equally be conveyed implicitly. This case study therefore provides some information about advanced learners’ sensitivity to the necessity of explicitly marking a coherence relation by the use of a connective. Our results indicate that advanced learners do not perceive the difference between relations that need and need not be marked by a discourse connective and have not acquired the complex form-function mapping of ‘en effet’. We argue that these difficulties cannot be attributed to negative transfer effects, but reflect general limitations in proficiency.

I argue that the communication of given information is part of the procedural instructions conveyed by some connectives like the French puisque. I submit in addition that the encoding of givenness has cognitive implications that are visible during online processing. I assess this hypothesis empirically by comparing the way the clauses introduced by two French causal connectives, puisque and parce que, are processed during online reading when the following segment is ‘given’ or ‘new’. I complement these results by an acceptability judgement task using the same sentences. These experiments confirm that introducing a clause conveying given information is a core feature characterizing puisque, as the segment following it is read faster when it contains given rather than new information, and puisque is rated as more acceptable than parce que in such contexts. I discuss the implications of these results for future research on the description of the meaning of connectives.

French connectives have been the object of detailed linguistic descriptions since the 1970s. Several of these earlyworks have paved the way for the development of classical concepts in discourse and pragmatics, such as the distinction between objective and subjective discourse relations and the notion of procedural meaning.

The lexical items like and well can serve as discourse markers (DMs), but can also play numerous other roles, such as verb or adverb. Identifying the occurrences that function as DMs is an important step for language understanding by computers. In this study, automatic classifiers using lexical, prosodic/positional and sociolinguistic features are trained over transcribed dialogues, manually annotated with DM information. The resulting classifiers improve state-of-the-art performance of DM identification, at about 90% recall and 79% precision for like (84.5% accuracy, @k=0.69), and 99% recall and 98% precision for well (97.5% accuracy, @k=0.88). Automatic feature analysis shows that lexical collocations are the most reliable indicators, followed by prosodic/positional features, while sociolinguistic features are marginally useful for the identification of DM like and not useful for well. The differentiated processing of each type of DM improves classification accuracy, suggesting that these types should be treated individually.

In French, a causal relation is often conveyed by the connectives car, parce que or puisque. Since the seminal work of the Lambda-l Group (1975), it has generally been assumed that parce que, used to relate semantic content, contrasts with car and puisque, both used to connect either speech act or epistemic content. However, this analysis leaves a number of questions unanswered. In this paper, I present a reanalysis of this trio, using empirical methods such as corpus analysis and constrained elicitation. Results indicate that car and parce que are interchangeable in many contexts, even if they are still prototypically used in their respective domain in writing. As for puisque, its distribution does not overlap with car, despite their similar domains of use. I argue that the specificity of puisque with respect to the other two connectives is to introduce a cause with an echoic meaning.

The identification of occurrences of like and well that serve as discourse markers (DMs) is a classification problem which is studied here on a corpus of dialogue transcripts with more than 4,000 occurrences of each item. Decision trees using item-specific lexical, prosodic, positional and sociolinguistic features are trained using the C4.5 method. The results demonstrate improvement over past experiments, reaching the same range as inter-annotator agreement scores. DM identification appears to benefit from itemspecific classifiers, which perform better than general purpose ones, thanks to the differentiated use of lexical features.

This paper presents a shallow dialogue analysis model, aimed at human-human dialogues in the context of staff or business meetings. Four components of the model are defined, and several machine learning techniques are used to extract features from dialogue transcripts: maximum entropy classifiers for dialogue acts, latent semantic analysis for topic segmentation, or decision tree classifiers for discourse markers. A rule-based approach is proposed for solving cross-modal references to meeting documents. The methods are trained and evaluated thanks to a common data set and annotation format. The integration of the components into an automated shallow dialogue parser opens the way to multimodal meeting processing and retrieval applications.