Maite Taboada
banner
mtaboada.bsky.social
Maite Taboada
@mtaboada.bsky.social

Distinguished Professor of Linguistics @SFU. Discourse analysis, corpus linguistics, computational linguistics, NLP, sentiment analysis, social media language, language of misinformation. she/her

Computer science 55%
Psychology 15%

Sounds like fun! I don't know AT at all, but open to learning more about it!

The paper:
Kolhatkar, V., N. Thain, J. Sorensen, L. Dixon and M. Taboada (2023) Classifying constructive comments. First Monday 28(4). doi.org/10.5210/fm.v...

We found good indicators of a constructive contribution *in news comments*:
- The comment provides evidence or a solution
- Promotes dialogue
- Addresses specific points (in the news article)
- May include personal stories or address personal experience as evidence

Hey, thanks for the shout-out! Yes, we have worked on models of constructive conversations, with the folks at Jigsaw (the flipside of the toxicity detector).

This was a group effort with colleagues at SFU's @digitaldemocracies.bsky.social

And yet, we see sentiment analysis everywhere. In this paper, we analyze the rhetoric of one particular company using sentiment analysis to gauge the employee happiness. You can imagine how that's going to go wrong. Bossware at its worst.

I have done research on sentiment analysis for years, trying to explain how it needs linguistically-informed methods. How you can't just "do sentiment analysis". How you can't just aggregate values and claim insight.

On how not to do sentiment analysis:
doi.org/10.1007/s001...

Last week, I visited @citius.gal, the Research Centre on Intelligent Technologies at the @usc.gal, my annual visit as a member of CiTIUS' Scientific Advisory Board.

Always great to see how well the Centre is doing.
And always such a pleasure to enjoy beautiful Santiago de Compostela!

Reposted by Maite Taboada

Congratulations to @sfulinguistics.bsky.social professor, @mtaboada.bsky.social, a finalist for the SSHRC Insight Award! The Insight Award recognizes outstanding achievement arising from SSHRC-funded initiatives which contributed to knowledge and understanding about people, societies and the world.

I'm a finalist for the 2025 #SSHRC Insight Impact Award! What an honour to be selected as finalist!
Thank you, @sshrc-crsh.canada.ca, for supporting social sciences and humanities research across Canada.

sshrc-crsh.canada.ca/en/news/2025...
And we’re also thrilled to announce that the next RaAM Conference will be held in Pamplona, Spain! Save the dates!

Amber Rynearson and I presented today at @raammetaphor.bsky.social, on our work trying to integrate different frameworks for metaphor identification.
What a great conference and so well run. Thank you, Hamad @al-azary.bsky.social and team!

Reposted by Maite Taboada

It has been a pleasure to present our research in #IPrA2025 at The University of Queensland (Australia).

Thanks, @cliff-goddard.bsky.social and Helen Bromhead (panel organizers), for your interest on our @uneduniv.bsky.social #EUPlainTech project (@ageinves.bsky.social) . 🤗

#PlainLanguage
#NLP

IPrA in 2027 - Helsinki

Nice idea of the organizers of #ipra2025: signs with the first IPrA we attended

In Brisbane for #ipra2025!

Reposted by Maite Taboada

📢 19th International Pragmatics Conference
🗓️ June 22–27
🌏 Brisbane
🔗 pragmatics.international/page/Program...

M. Bermúdez, I. da Cunha (@uneduniv.bsky.social) & M. Taboada (@sfulinguistics.bsky.social) will present our #EUPlainTech project (@ageinves.bsky.social).

Oh, wow, Laura and Carmen! Two such lovely people. Say hi!

Talk by Barbara Dancygier, one of the most prominent scholars in cognitive linguistics. And this one is about memes!

E.g., here they are compared to other spoken registers. The mean factor scores are similar, but the spread for podcasts (purple box) is huge.

We show that podcasts are an emerging register of computer mediated communication, with lots of internal variability.

Jussi Karlgren, then at Spotify, gave us the idea of working on podcasts and was kind enough to zoom in to our class to give us background on their dataset.

The funny bit of background is that we also brainstormed on that other social media site, appealing to the community of discourse experts to help us come up with interesting data to analyze.

The paper started at a graduate seminar in Fall 2022. All the students in the class collaborated to brainstorm research ideas, collect data, think of methodologies, run analyses, & write up the results. Katharina Ehret, our resident MDA expert, joined later to help us run more analyses.

Another paper a long time in the making:
3️⃣ years
🔟 authors
1️⃣ graduate seminar
9,789 podcast transcripts
1️⃣ paper: doi.org/10.1075/rs.2...

You're welcome! GeMeCo (the Spanish language dashboard) has some IP restrictions because of potential DoS attacks.
the ups and downs are sometimes temporary. The trend line is at about 30%