MTMaite Taboada

Maite Taboada

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Maite Taboada has a total of 90 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Linguistic approaches to fake news research are growing and maturing: commentary on a special issue

    by Maite Taboada

    This first special issue of Linguistics Vanguard on the language of fake news offers multiple perspectives on the linguistic analysis of fake news and misinformation, with a range of approaches and methodologies, showing the sophistication and maturity of the field. The study of fake news is a flourishing area of research, with contributions from communication, media studies, data science, and natural language processing. The growing body of research in linguistics contributes careful qualitative analyses of the social context of production and interpretation of misinformation, coupled with well-documented techniques from corpus linguistics. This commentary provides a short retrospective of linguistic approaches to fake news and highlights the contributions from the papers in the issue.

  • Conflicting feelings: sociological and computational sentiment in workplace sentiment surveillance

    by Evan Donahue, Matthew Canute, Adjua Akinwumi, Philippa R. Adams, Maite Taboada, Wendy Hui Kyong Chun

  • Open Access

    Attitude in Reported and Non-reported News: A Critique of Sentiment Analysis in Corpus Pragmatics

    by Monika Bednarek, Maite Taboada

    This study uses natural language processing (NLP) tools to examine a large Canadian English-language news corpus with respect to quotation and positive/negative sentiment. Specifically, we analyse sentiment in reported/quoted speech in comparison to non-quoted speech, testing the hypothesis that quoted speech contains negative sentiment and is more subjective. Our study explores whether NLP tools that simplify pragmatically complex concepts (such as attitude/evaluation/stance) can be used to test hypotheses that derive from discourse analytic or pragmatic studies of news discourse. We show that sentiment analysis is not suitable for accurate analysis of the news values of Positivity and Negativity and cannot be used to test hypotheses that derive from assumptions about these news values. At the same time, some of the insights from the sentiment analysis confirm our hypotheses (and are in line with other corpus studies), and sentiment results can be a starting point for additional qualitative analysis. Finally, we suggest a range of possible developments for sentiment analysis which draw on linguistic considerations.


  • Podcasts as an emerging register of computer-mediated communication

    by Katharina Ehret, Laurens Bosman, Aminat Babayode, Nicole Chan, Ivan Fong, Noelle Harris, Alissa Hewton, Danica Reid, Rebekah Wong, Maite Taboada

    Podcasts, a relatively recent audio medium, have risen in popularity since their initial appearance in the mid-2000s. Yet, little is known about their lexico-grammatical characteristics and their relation to other computer-mediated and traditional registers. Addressing this gap, we apply Biber-style multidimensional analysis (MDA) to a representative sample of Spotify podcast transcripts and selected computer-mediated registers (e.g., informational blog, interview) as well as traditional spoken registers (e.g., broadcast, conversation). We compare their lexico-grammatical characteristics to those of other registers along the emerging dimensions. We find that, while podcasts share some linguistic characteristics with traditional spoken registers such as broadcast discussion and scripted speech, they are unlike any of the analyzed registers. In fact, their most striking characteristic is their considerable internal variability, likely related to their versatility but also due to their mixing of features and very diverse nature. In short, podcasts are an emerging register of computer-mediated communication.


  • Open Access

    Reported speech and gender in the news: Who is quoted, how are they quoted, and why it matters

    by Maite Taboada

    News stories have a well-defined generic structure, consisting of components such as headline, lede, and body, with reported speech a prominent feature, especially in hard news stories. Reported speech serves multiple purposes, from providing evidentiality and intertextuality to contributing to the construction of newsworthiness and to the context creation of news. It is also a site of potential bias in who is cited and how, including with respect to the gender of sources. Using a large corpus of English-language news stories for all of 2023 from the main five mainstream news outlets in Canada (over 370,000 articles from news websites), I examine the gender distribution of those quoted, the syntactic variation in the structure of quotes, and the types of reporting verbs. The study provides a comprehensive overview of the extend of gender bias in contemporary Canadian news, at the same time offering insights into the nature of reported speech in modern news and how it endures and evolves, including in news meant for digital-only publication.


  • Open Access

    The ‘adverb-ly adjective’ construction in English: meanings, distribution and discourse functions

    by Maite Taboada, Cliff Goddard, Radoslava Trnavac

    We investigate a class of adjective phrases composed of a deadjectival adverb ending in -ly and an adjective head (e.g. staggeringly incompetent, absolutely terrific, fiscally responsible), a compact construction whereby two adjectives may jointly contribute to evaluative meaning. Using corpus methodologies on more than 1 million examples and relying on semantic analyses of about 1,000 instances, we propose that the construction can be divided into different semantic subtypes, including Degree (deeply disturbing), Focus (utterly ridiculous), Manner (delightfully performed), Reaction (strangely compelling), Topical (historically inaccurate) and Epistemic (intuitively obvious), among others. Using this typology, we investigate the relative distribution of each subtype across several registers of written English. We found a high frequency of the Reaction subtype in book, film and art reviews, and we suggest a discourse-functional explanation for this, linked to the perceived value of originality in expressive writing. This investigation reveals the power of semantically informed, corpus methodologies to shed light on the distribution of specific constructions.

  • Misinformation Detection in News Text

    by Fatemeh Torabi Asr, Mehrdad Mokhtari, Maite Taboada

    We address the difficulty of automatic misinformation detection through state-of-the-art natural language processing techniques. Machine learning and natural language processing are often touted as the perfect solutions for the problem of detecting misinformation at scale. We argue, however, that current approaches tend to fall short because of the unavailability of reliably annotated data. Given the scarcity of quality labelled data, we first conduct a data collection effort by leveraging fact-checking websites. Second, we perform a comparative feature analysis of the news articles with true vs false content. Finally, we conduct a set of text classification experiments using a variety of methods and show that the quality of the training data in terms of its labelling system and balanced coverage of topics directly affects the classification accuracy on test data (news articles that the classifier did not see during training). Feature analysis experiments show specific trends of linguistic patterns in fake news articles. In predictive classification, some of these features such as n-grams and semantic features help considerably in recognising false from true news articles and some features such as readability features tend to be less helpful. We also compare deep learning classification models against feature-based models and show that, given the small size of currently available data, feature-based models are more capable of cross-topic generalisation. This result points to the need for automatic classification methods that are informed by linguistic, corpus linguistic, and stylistic research. Finally, we show that using data labelled based on the reputation of the sources does not result in accurate classification of test data, which, in turn, motivates future data collection with reliable tagging and on diverse topics.


  • Open Access

    Radar de Parité: An NLP system to measure gender representation in French news stories

    by Valentin-Gabriel Soumah, Prashanth Rao, Philipp Eibl, Maite Taboada

    We present the Radar de Parité, an automated Natural Language Processing (NLP) system that measures the proportion of women and men quoted daily in six Canadian French-language media outlets. We outline the system’s architecture and detail the challenges we overcame to address French-specific issues, in particular regarding coreference resolution, a new contribution to the NLP literature on French. We also showcase statistics covering over one year’s worth of data (282,512 news articles). Our results highlight the underrepresentation of women in news stories, while also illustrating the application of modern NLP methods to measure gender representation and address societal issues.


  • Open Access

    Radar de Parité: An NLP system to measure gender representation in French news stories

    by Valentin-Gabriel Soumah, Prashanth Rao, Philipp Eibl, Maite Taboada

    We present the Radar de Parité, an automated Natural Language Processing (NLP) system that measures the proportion of women and men quoted daily in six Canadian French-language media outlets. We outline the system's architecture and detail the challenges we overcame to address French-specific issues, in particular regarding coreference resolution, a new contribution to the NLP literature on French. We also showcase statistics covering over one year's worth of data (282,512 news articles). Our results highlight the underrepresentation of women in news stories, while also illustrating the application of modern NLP methods to measure gender representation and address societal issues.