Mark Dingemanse

Mark Dingemanse

@markdingemanse.net

Language, interaction, tech • papers https://markdingemanse.net • blog https://ideophone.org • more colourful on fedi https://scholar.social/@dingemansemark/ • POSSE: Publish on Own Site, Syndicate Everywhere

3.2K followers557 following4.6K posts

Pinned

Revised! Interactional foundations for critical AI literacies doi.org/10.5281/zeno... Gods speak, spirits speak, computers speak. Oracular ambiguity or statistical probability provides loopholes, and discrepancies are expunged by Faith.” — Ursula K. Le Guin, The Left Hand of Darkness (1969)

doi.orgInteractional foundations for critical AI literaciesThe ubiquity and ease of use of large language models makes it easy to overlook the interactional and interpretive processes at play. To understand the attraction of this technology we need to trace i...

Proofs and coffee, my favourite combo. Coffee is a Blommers roast of Costa Rica Aquiares, soundtrack is Some day my prince will come (Davis/Coltrane), paper is 'Overlap in conversation' w/ first author Ada Lopez and Andreas Liesenfeld, soon in a journal near you

Fijn om deze gereviseerde gedragscode te zien, met heldere tekst over technologie. Wat opvalt is hoe stabiel de waarden zijn — precies dezelfde als in de vorige versie: wetenschappers streven naar werk dat eerlijk, zorgvuldig, onafhankelijk, transparant, en verantwoordelijk uitgevoerd is.

Iris van Rooij 💭@irisvanrooij.bsky.social

"Universiteiten en hogescholen dienen hun onderzoekers te beschermen tegen druk en inmenging. Wetenschappers moeten voorzichtig zijn met AI. Dat staat in de nieuwe gedragscode wetenschappelijke integriteit." www.voxweb.nl/nieuws/nieuw...

Of zoals de universiteit van Aalto nuchter schreef, 'Nothing new here, actually! The rule of research integrity and responsible conduct of research still apply!' (geciteerd in mijn bijdrage in de bundel Als wij denken, net verschenen, hier te lezen doi.org/10.5281/zeno... )

Mark Dingemanse reposted

Why do some colleagues from psych, cog neuro, etc., accept AI nonsense that all or almost all our science will be automated? I'm a bit flabbergasted. Yes, of course we'll need observations, always. But to say theory and modelling practice can be automated by AI is preposterous. Some basic points...

Mark Dingemanse reposted

The first genuinely honest ad for generative AI I've ever seen, from Microsoft / CoPilot A mediocre white guy sees his non-white colleague getting praised for good work, he asks AI to *literally copy it*, and then he gets all the praise instead: www.ispot.tv/ad/SoDi/micr...

Mark Dingemanse reposted

one thing I'm gonna say is: more horn sections and drumlines at protests please.

Greg Pak@gregpak.net

#NoKings Vote Early parade in Bay Ridge with @bkindivisible.bsky.social and the @brassqueensnyc.bsky.social ! Vibes: impeccable! Vote early, friends! nokings.org

The writeup is a bit imprecise —what I write about the Claude code leak was not directly related to the notion of epistemic diversion — but certainly a central reason this tech bamboozles so effectively is design choices that invite the attribution of understanding and that conceal the lack of it

AI researchers warn technology 'months away' from being able to log into manuscript management system to submit peer review

This post is unavailable.

Mark Dingemanse reposted

Appreciate the sharing of this - a favorite page I always use in talks, and I appreciate the alt text, well done! And it gives me insights that this is read as a grid, which I see now, though I intended it as infinitely forking paths already chosen. Let’s keep towards “what might be…”

Mark Dingemanse@markdingemanse.net

the best world for these people is one where everything is automated. this is progress — to buy into that bleak vision is to "lean into the future" Made me think of this page from Unflattening: Forgotten the wonder of what might be. / In its place, a single chorus... / This is how it is.

Won't link DHH here but here's a recent bit from his blog that is really telling about his technofascist utopia > We're about to see an absolute bloom in software development as the price of development plummets and everyone realizes how much automation we still have left to do in this world.

the best world for these people is one where everything is automated. this is progress — to buy into that bleak vision is to "lean into the future" Made me think of this page from Unflattening: Forgotten the wonder of what might be. / In its place, a single chorus... / This is how it is.

Never heard about these atrocities at school. Learnt of the Banda massacre much later, when I read Wallace and found that today, Bandanese is only spoken on the Kei islands 400km away from Banda, where some survivors fled Down with JP Coen statues!

Last heartbreaking example because it should be brought to light what vile things the Dutch have done and are still largely proud of. > The population of the Banda Islands prior to Dutch conquest is generally estimated to have been around 13,000-15,000 people...

> While […] prescriptive technologies are often exceedingly effective and efficient, they come with an enormous social mortgage. > […] that we live in a culture of compliance, that we are ever more conditioned to accept orthodoxy as normal, and to accept that there is only one way of doing “it”.

Freire in Education for Critical Consciousness (2002/1973): > If [people] are unable to perceive critically the themes of their time, and thus to intervene actively in reality, they are carried along in the wake of change. They see that the times are changing, but they are submerged in that change

BTW I changed 'men' to 'people'. The casual and seemingy unexamined sexism of the language in Freire is quite grating and/but whenever I see it, I think of what bell hooks said on this point (in a conversation with 'Gloria Watkins' in Teaching to transgress, hooks 1994:49) #feminism #sexism #freire

Mark Dingemanse reposted

Replying to @maartenkossmann.bsky.social

derived from generally attested Berber terms for body parts He also showed that the metaphors make sense in the landscape that is described. Put otherwise: the question to what extent the old language of the Canary Islands was a form of Berber is now settled: indeed it was! @phdnix.bsky.social

Mark Dingemanse reposted

Today at NACAL there was a fantastic talk by Massinissa Garaoun on pre-Hispanic toponomy in the Canary Islands. Based on extensive fieldwork, checking pronuniations and the topographical features the terms are connected to, he showed that there are many toponyms that are clearly...

Revisiting this in 2026: I decided to preprint my contribution doi.org/10.5281/zeno... It's like a time capsule except for the references (which have been updated in 2023): an early take on the language of perception in Siwu, eastern Ghana

doi.orgThe language of perception in SiwuThe language of perception in Siwu, a Kwa language of eastern Ghana, is described based on linguistic elicitation, ethnographic field research, and a standardized set of stimuli targeting the perceptu...
Mark Dingemanse@markdingemanse.net

Ten years ago, fresh out of my PhD, I completed three papers. One I submitted to a regular journal; it came out in 2012. One was for a special issue; it took until 2017 to appear. One was for an edited volume; the volume is yet to appear. >

Mark Dingemanse reposted

Dear @erc.europa.eu , we DO already have common standards & frameworks. 'AI' likes to pose itself as 'special' but we do not need to reinvent the wheel. We already have established codes of conduct for research integrity. We just need to apply them. See below + more resources in thread🧵

There is a need to agree on common standards & frameworks for AI models used across all aspects of research evaluation. These should ensure & encourage trust, confidentiality, & transparency, & should be defined through collaborative discussions between research funding and performing organisations.

Mark Dingemanse reposted

Replying to @markdingemanse.net

As we conclude, > [LLM] users encounter a fundamental asymmetry: they must singlehandedly supply much of the sequential structure that is ordinarily co-constructed by multiple participants ... This hidden labour challenges accounts of LLM use that focus on performance and optimization

This was incredible. Sharing some shots of the book table that was up during the event for inspiration — ft work by Bender & Hanna, Butler, Césaire, Fanon, Lorde, de Kom, O'Neill, Sheldrake, Simard, Stengers, Woolf, and many more

Mark Dingemanse reposted

How men feared women! she thought, walking among the late-flowering roses. Not as individuals, but women when they talked together, worked together, spoke up for one another - then men saw plots, cabals, constraints, traps being laid.

As I wrote in 2020 about the same topic, all ingredients for an information heat death are on hand. ideophone.org/large-langua... I disagree with the quoted post's indiscriminate use & ascription of "intellect" and its conflation of textual products & cognitive processes; this is self-defeating

Oleg Sobchuk 🇺🇦@sobchuk.bsky.social

I don't agree with some things in this (very sharp) blog post, but it certainly offers a new angle on "AI and academia". Papers will become less valuable, prestigious institutions - more. But also (and here is me thinking) - direct communication unmediated by text. giorgio.gilest.ro/the-weimar-o...

Mark Dingemanse reposted

Mark Dingemanse reposted

It’s excellent getting to say to a group of students, We’re gonna read X & we’re gonna spend 75 minutes talking about it, & ideally the way you think about the world & yourself is gonna be a bit more nuanced, complex, & satisfying, & that’s today’s “product.” Or, as @cnygren.bsky.social & I said:

Tapping the sign

Mark Dingemanse@markdingemanse.net

When an "AI" provider announces a new result, remember they have a product to sell. Watch for the careful and systematic ways human labour is elided at every step and claims about the model and the 'discovery' process are shielded from direct inspection Today's specimen: openai.com/index/model-...

I remember wondering whether this would survive typesetting but CUP clearly understood the assignment and kept my unicode ッ intact

Mark Dingemanse@markdingemanse.net

I love ¯\_(ツ)_/¯ so much that I used it in print in an academic paper 🌞 doi.org/10.1017/lang...

“There appears to have been a profound shift, beginning in the 1970s, from investment in technologies associated with the possibility of alternative futures to investment technologies that furthered labor discipline and social control” — David Graeber, The Utopia of Rules, 2015 #quote #technology

For the morning crowd: new paper!

Mark Dingemanse@markdingemanse.net

Very proud to share this new work led by the amazing Marlou Rasenberg (not on here), with Joel Fischer and me doi.org/10.5281/zeno... We use #EMCA methods to study empirically how LLM users shoulder the burden of solving sequential trouble and end up restructuring their own aims & activities

Mark Dingemanse reposted

Replying to @markdingemanse.net

While much work in so-called 'prompt engineering' obsesses over the perfect prompt (Andreessen's monster prompt comes to mind), we find that real-life prompting practices are sequentially organized, feature repurposed patterns of human interaction, and result in users adjusting expectations & aims

Very proud to share this new work led by the amazing Marlou Rasenberg (not on here), with Joel Fischer and me doi.org/10.5281/zeno... We use #EMCA methods to study empirically how LLM users shoulder the burden of solving sequential trouble and end up restructuring their own aims & activities

View full thread

As we conclude, > [LLM] users encounter a fundamental asymmetry: they must singlehandedly supply much of the sequential structure that is ordinarily co-constructed by multiple participants ... This hidden labour challenges accounts of LLM use that focus on performance and optimization

Very proud to share this new work led by the amazing Marlou Rasenberg (not on here), with Joel Fischer and me doi.org/10.5281/zeno... We use #EMCA methods to study empirically how LLM users shoulder the burden of solving sequential trouble and end up restructuring their own aims & activities

We identify a set of practices we call 'pre-emptive prompting', "through which users creatively circumvent the system’s lack of action recognition and action projection". Users stumble upon these through trial and error, in exchanges like this:

While much work in so-called 'prompt engineering' obsesses over the perfect prompt (Andreessen's monster prompt comes to mind), we find that real-life prompting practices are sequentially organized, feature repurposed patterns of human interaction, and result in users adjusting expectations & aims