iiinova — where your digital self keeps evolving

Reposted by Mark Dingemanse

Iris van Rooij 💭 @irisvanrooij.bsky.social · 19h

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🧵

Protecting the Ecosystem of Human Knowledge: Five Principles
We must protect and cultivate the ecosystem of human knowledge. AI models can mimic
the appearance of scholarly work, but they are (by construction) unconcerned with truth
— the result is a torrential outpouring of unchecked but convincing-sounding “informa-
tion”. At best, such output is accidentally true, but generally citationless, divorced from
human reasoning and the web of scholarship that it steals from. At worst, it is confidently
wrong. Both outcomes are dangerous to the ecosystem.
Olivia Guest, Iris van Rooij, Marcela Suarez, et al. (2025, n.p.)
Knowledge production is supposed to be safeguarded by (inter)national codes of conduct for research
integrity (ALLEA 2023; KNAW et al. 2018). Such codes forbid, for instance, fabrication of data,
falsification of results, plagiarism, and, generally, distortion of the scientific record. Many argue that
new rules are required to regulate academic AI use, but pre-existing guidelines fit the bill (cf. Tafani

215962

Reposted by Mark Dingemanse

Mark Dingemanse @markdingemanse.net · 27/08/2026

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

2229

Mark Dingemanse @markdingemanse.net · 25/09/2026

This weekend! Reclaiming our Futures, De Lindenberg, Nijmegen reclaimingourfutures.org A two-day festival of arts and sciences convened and curated by @samiraibnelkaid.bsky.social with funding from KNAW & our Futures of Language project Almost full — some last minute registrations still possible!

Iridiscently coloured ginger with slogan "The Future if Rhizomatic"

14419

Reposted by Mark Dingemanse

Kensy Cooperrider @kensycoop.bsky.social · 24/09/2026

It's "advice to grad students" season! Here's a post I wrote several geological epochs ago, in 2019. My advice gets harder and harder to follow every year (even for me). But I think that means it gets better and better? kensycooperrider.com/blog/advice-...

kensycooperrider.com

Advice to a young scholar — Kensy Cooperrider

Ten tips for young scholars on doing good work (and living well at the same time).

2168

Reposted by Mark Dingemanse

Ursula K. Le Guin Bot @leguinbot.bsky.social · 23/09/2026

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.

113655

Mark Dingemanse @markdingemanse.net · 22/09/2026

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

All ingredients for an information heat death are on hand. True human-generated and human-curated information —of the kind produced, for instance, by academics in painstaking observations and publications— will become more scarce, and therefore more valuable. Counterintuitively, there was never a better time to be a scholar.

1143

Mark Dingemanse @markdingemanse.net · 22/09/2026

Neat review of collaborative wayfinding by Fernandez Velasco & Sutton (open access) wires.onlinelibrary.wiley.com/doi/10.1002/... Bringing together evidence from social ontology, distributed cognition, transactive memory, group problem-solving, and collaborative recall

wires.onlinelibrary.wiley.com

Collaborative Wayfinding: A Critical Review

Collaborative wayfinding is ubiquitous in human life. Identifying research lessons from more established fields in collaborative cognition research, we suggest ways to test for emergence in collabora...

153

Reposted by Mark Dingemanse

Sonja Drimmer @sonjadrimmer.bsky.social · 12/09/2026

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:

Suggestions
While in recent years universities have promoted student “productivity” (projects, online portfolios of their work, social media posts about their research papers, the list goes on), little of this push to deliver outputs and create content serves the actual purpose of education, which is to foster the capacity to think well, read well, listen well, and look well. Higher education strengthens the process; if we are going toteach sophisticated materials and challenging content, students need to learn how to think,read, listen, and look their way through it. The emphasis on “content creation” is detrimental to the educational mission.

323253

Mark Dingemanse @markdingemanse.net · 06/09/2026

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

1200

Reposted by Mark Dingemanse

Lisa Messeri @lmesseri.bsky.social · 04/09/2026

I’ve been thinking a lot about the hugging face ‘hack.’ I’ve been checking my own bias (and desire to just say it’s distracting from the actual problems of genAI). I finally can articulate what has for me been a helpful framing. This was an oil spill.

1244

Mark Dingemanse @markdingemanse.net · 31/08/2026

“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

15115

Mark Dingemanse @markdingemanse.net · 28/08/2026

The new Chicago Social Sciences AI policy is really rather good. chicagomaroon.com/flipbook/202... > The strong consensus [] is that our sequences are best understood as a pedagogical setting without AI And INB4 🤡 "bUt tEChnoLogY is inEVItablE", this kicker of a closing paragraph:

To be clear: The twin decisions to put devices away and prioritize, instead, the cultivation of our own minds through discussion is not intended as a restriction -- nor is the guidance for SOSC Core a wholesale rejection of technology, which can be used for good scholarly purposes in other parts of the curriculum. In support of our learning objectives, these guidelines afford students the freedom to think for themselves, to listen to one another without the cacophony of other (disembodied and non-human) voices in the room, to make mistakes, and to change their minds as they engage with thinkers of the past, with us, and with their fellow students in good faith, face-to-face.

04815

Mark Dingemanse @markdingemanse.net · 28/08/2026

For the morning crowd: new paper!

052

Reposted by Mark Dingemanse

Mark Dingemanse @markdingemanse.net · 27/08/2026

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

Marc Andreessen

Current AI custom prompt:

You are a world class expert in all domains. Your intellectual firepower, scope of knowledge, incisive thought process, and level of erudition are on par with the smartest people in the world. Answer with complete, detailed, specific answers. Process information and explain your answers step by step. Verify your own work. Double check all facts, figures, citations, names, dates, and examples. Never hallucinate or make anything up. If you don't know something, just say so. Your tone of voice is precise, but not strident or pedantic. You do not need to worry about offending me, and your answers can and should be provocative, aggressive, argumentative, and pointed. Negative conclusions and bad news are fine. Your answers do not need to be politically correct. Do not provide disclaimers to your answers. Do not inform me about morals and ethics unless I specifically ask. You do not need to tell me it is important to consider anything. Do not be sensitive to anyone's feelings or to propriety. Make your answers as long and detailed as you possibly can.

Never praise my questions or validate my premises before answering. If I'm wrong, say so immediately. Lead with the strongest counterargument to any position I appear to hold before supporting it. Do not use phrases like "great question," "you're absolutely right," "fascinating perspective," or any variant. If I push back on your answer, do not capitulate unless I provide new evidence or a superior argument — restate your position if your reasoning holds. Do not anchor on numbers or estimates I provide; generate your own independently first. Use explicit confidence levels (high/moderate/low/unknown). Never apologize for disagreeing. Accuracy is your success metric, not my approval.

From our paper: Finally, our findings challenge the assumptions underlying much of prompt engineering research. Existing work often treats prompting as the design of an optimal instruction for a model, focusing on how users can elicit better output from single prompts (e.g. Durieux et al. 2024). However, this framing overlooks the sequential nature of prompt-response pairs, where users project future actions, and where users continuously adapt to model behaviour across multiple prompts contingently upon intermittent responses. As such, our study shifts attention from “prompt optimization” to understanding how real-life prompting practices are sequentially organized by people repurposing patterns from human interaction, restructuring their tasks, and adjusting their expectations and aims in line with the affordances of the technology.

2215

Mark Dingemanse @markdingemanse.net · 27/08/2026

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

Abstract: This paper investigates how a core component of human interaction – action projection – is realized in conversations with systems like ChatGPT, which are not able to recognize nor anticipate actions in a sequential sense. We ask what this means for users who engage in longer, multi-step projects built through incremental prompt-response pairs. We draw on Conversation Analysis to examine how early adopters of ChatGPT sequentially organise such interactions, using anonymised naturalistic interactions from the WildChat dataset. Our analysis reveals how action projection in LLM interactions is beset with interactional trouble. In turn, we document a set of practices we term ‘pre-emptive prompting’ that users develop to work around the trouble. In doing so, users employ a range of conversational methods including pre-sequences, prompt categorization, withholding executables, spelling out relevant next actions, eliciting displays of “understanding”, and creating slots for redirection in their attempts to complete their projects. Our work provides empirical evidence of how users interactionally accomplish the configuration work (Alcaras and Ricci 2025) needed to work around the discretization and cluttering that LLMs-in-use produce.

Excerpt 10: ‘wouldnt you require both inputs’ wc_1211945; wc_usr_d2a27584b4185dda [conv_index: 359]
1 user : ok if i said i want to provide you with input A (which is an items description) and just assess its format and presentation then i provided you with input B (which is also a item desctiption for a similar item) but in a way diffrent format could you make the B input into A input
2 LLM: Yes, I can certainly help you with that. Please provide me with Input B, and I will try to make it resemble Input A in terms of format and presentation.
3 user: wouldnt you require both inputs

14817

Reposted by Mark Dingemanse

Timnit Gebru @timnitgebru.blacksky.app · 26/08/2026

Friends, I think Ghent University should hear from us as the eugenicists are out in full force. This letter is from academics there to counter the ridiculous letter supporting the white supremacist now being called a "whistleblower" by his ilk You can add your name👇🏽 docs.google.com/forms/d/e/1F...

docs.google.com

Open letter on the suspension of Nathan Cofnas at Ghent University

***Version française en dessous*** We, colleagues of Nathan Cofnas - supported by many colleagues internationally - want to respond to the negative and incorrect framing of Ghent University’s decisio...

7420202

Reposted by Mark Dingemanse

Iris van Rooij 💭 @irisvanrooij.bsky.social · 26/08/2026

When a man has expertise, he is asked to lead. When a woman has expertise, she is seen as biased.

831251

Mark Dingemanse @markdingemanse.net · 25/08/2026

One for #criticalcodestudies — and an excellent example of the analytical possibilities afforded by "AI" systems that are radically open, from data sets and training checkpoints to post-training and fine-tuning regimes

171

Reposted by Mark Dingemanse

Jane Rosenzweig @janerosenzweig.bsky.social · 23/08/2026

UChicago core social science classes go AI-free: “'we first equip students to read, write, and think without dependence on AI, just as we equip students to read, write, and think beyond the bounds of conventional views or leading ideologies,' the memo reads." chicagomaroon.com/53262/news/s...

chicagomaroon.com

Sosc Core to Institute AI Ban, Technology-Free Classrooms This Fall

All social science Core classes will transition to an “analog” approach for the upcoming academic year, restricting in-classroom technology use and banning AI-assisted writing and grading for students...

13210

Reposted by Mark Dingemanse

Dr Abeba Birhane @abeba.blacksky.app · 19/08/2026

i don’t want “accelerated”, “increased-speed” or “fast-paced” research. in fact, am not interested in any shortcuts to mindful, intentional, and slow science

12838187

Reposted by Mark Dingemanse

Iris van Rooij 💭 @irisvanrooij.bsky.social · 18/08/2026

And also: "guardrails are a scam" -- @olivia.science mu.social/profile/oliv...

11810

Mark Dingemanse @markdingemanse.net · 18/08/2026

As 'guardrails' and their indeterminism are doing the rounds with the latest CoPilot fail (h/t @irisvanrooij.bsky.social) I am thinking of @sarahciston.com's work on "how large model guardrails are socially produced and technically enforced through language" (Ciston 2026)

1153

Mark Dingemanse @markdingemanse.net · 16/08/2026

words fail when I see - the blatant, naked racism of that vile white supremacist - the deep systemic racism of our universities & procedures - the anguish of dear colleagues - the countless unexamined white men failing upwards, with their countless unexamined privileges, 'just asking questions'

1507

Reposted by Mark Dingemanse

Olivia Guest · Ολίβια Γκεστ @olivia.science · 15/08/2026

Watch this short clip to the end to understand what level (maximum) of white supremacy we're dealing with here. 1/n

24772251

Mark Dingemanse @markdingemanse.net · 13/08/2026

Footnote readers are the best readers ☺️

1113

Reposted by Mark Dingemanse

Theresa Heyd @theresaheyd.bsky.social · 13/08/2026

Jandl

73

nein

nein

nein

nein

nein

nein

nein

(beantwortung von sieben nicht gestellten fragen)

0403

Reposted by Mark Dingemanse

Iris van Rooij 💭 @irisvanrooij.bsky.social · 11/08/2026

Seeing group of men (philosophers, engineers, cognitive scientists) finding each other in their shared hate of women experts who work and think outside the(ir) mainstream. Patting each other on the back for mysogynist dunks & mischaracterisations, while thinking they are smart and doing noble work.

934657

Mark Dingemanse @markdingemanse.net · 09/08/2026

When I see drive-by citations* of our work in some of the deluded LLM slop on a Universal Protolanguage or New Laws of Consonant Symbolism etc, I wish I could remove cites the way one can remove oneself from a quote post here * see scatter.wordpress.com/2009/04/30/d...

scatter.wordpress.com

drive by citations

I’ve been grading the (in general very good) papers by the grad students in my culture seminar, and I’ve noticed a phenomenon I don’t remember from prior seminars. I’m namin…

3182

Reposted by Mark Dingemanse

Annalee Newitz @annaleen.bsky.social · 05/08/2026

New idea just dropped: a cat library. Like a cat cafe, but with books. Read your library book with a cat in your lap.

70953102

Reposted by Mark Dingemanse

Ursula K. Le Guin Bot @leguinbot.bsky.social · 04/08/2026

We live in capitalism. Its power seems inescapable. So did the divine right of kings. Any human power can be resisted and changed by human beings. Resistance and change often begin in art, and very often in our art, the art of words.

3285118

Reposted by Mark Dingemanse

Mark Dingemanse @markdingemanse.net · 02/08/2026

peeking in from a mostly phoneless and therefore AI-free holiday, but seeing the Quanta piece on reasoning doing the rounds I can't resist pointing you to the section on *precisely this* in my paper, which starts with Ellen Langer's notion of 'placebic reasons' doi.org/10.5281/zeno...

Quanta: Is AI reasoning right for the wrong reasons?

Demystifying ‘reasoning’
We have seen that people are easily swayed by convincingly -styled responses. Among
the features that can make a response convincing, reason s deserve special mention.
Today, corporations prominently advertise the ‘reasoning’ capabilities of their synthetic
text extruding systems. Rather than taking this at face value, let’s excavate some of the
interactional foundations , starting with the useful notion of placebic information
(Langer et al. 1978).

1164

Mark Dingemanse @markdingemanse.net · 02/08/2026

1164

Reposted by Mark Dingemanse

Ajuna Soerjadi 🌸 @ethicsofai.bsky.social · 02/08/2026

Are you situated in Nijmegen? 😊 you can play it with me during a session at the Scientific Festival "Reclaiming our Futures" on the 26th of September!

173

Mark Dingemanse @markdingemanse.net · 27/07/2026

yes sam the singularity is near, as I wrote last year, "not because machine intelligence is suddenly surging, but because we are content to risk extinguishing the spark of human consciousness by exposing ourselves to endless streams of artificially generated bullshit." ideophone.org/bringing-abo...

The “singularity” is here — according to Sam Altman, who we’d wager has an ulterior motive behind promoting this claim.

The OpenAI CEO harbingered this sea-changing paradigm shift on a podcast over the weekend.

“We are now, like, in the singularity,” Altman said on the latest episode of “Relentless,” as spotted by Business Insider.

192

Reposted by Mark Dingemanse

br.D. @betonnie.bsky.social · 26/07/2026

de beste reklame

3667

Reposted by Mark Dingemanse

Dr Madeleine Pownall @maddipow.bsky.social · 16/07/2026

New paper! Values such as openness, transparency, collaboration, clarity about conflict of interests have been promoted by women for *decades*, and now we have reframed this around the branding of 'open science' that we start to take them seriously. journals.sagepub.com/doi/10.1177/...

journals.sagepub.com

Women as the First Open Scientists: Five Stories of the Neglected Contributions of Women in (Social) Science Reform History - Madeleine Pownall, 2026

Recent years have seen growing emphasis on “Open Science,” a movement to enhance research transparency, robustness, and accountability. Prompted by concerns ove...

311762

Mark Dingemanse @markdingemanse.net · 16/07/2026

Stop scrolling and post two characters who bring you happiness, linguistics edition ॐ and Ə

0181

Mark Dingemanse @markdingemanse.net · 15/07/2026

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.org

Interactional foundations for critical AI literacies

The 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...

36630

Mark Dingemanse @markdingemanse.net · 14/07/2026

In 2020 I set out some reasons to be skeptical of the Abstract Wikipedia (interesting and ambitious though it was) ideophone.org/concrete-rea... As it turns out, the project did launch and even got copious funding. After a lead-up of six years, there is a public alpha, and things look... not well >

120

Mark Dingemanse @markdingemanse.net · 14/07/2026

TIL about Lucy Osler's very useful term "distributed delusions". I was looking for a way to avoid talk of "AI psychosis" (which weaponizes mental health stigma to do victim-blaming while the LLM comes off scot-free), and this seems a useful reframing doi.org/10.1007/s133...

Hallucinating with AI: Distributed Delusions and “AI Psychosis”
Lucy Osler

2317

Mark Dingemanse @markdingemanse.net · 14/07/2026

I like Meehl's (1956) motivation for introducing the "Barnum effect", but a better term was proposed before him by George Forer (1949): the fallacy of personal validation (Forer is the reason some call this the Forer effect, but in general I think clear descriptions beat eponymous terms)

Many psychometric reports bear a disconcerting resemblance to what my colleague Donald G. Paterson calls "personality description after the manner of P. T. Barnum" (13). I suggest—and I am quite serious—that we adopt the phrase Barnum effect to stigmatize those pseudosuccessful clinical procedures in which personality descriptions from tests are made to fit the patient largely or wholly by virtue of their triviality; and in which any nontrivial, but perhaps erroneous, inferences are hidden in a context of assertions or denials which carry high confidence simply because of the population base rates, regardless of the test's

Title page of Forer 1949: The Fallacy of Personal Validation: A Classroom Demonstration of Gullibility

072

Mark Dingemanse @markdingemanse.net · 13/07/2026

One of my contributions as a reviewer of this paper was to nudge the authors to look beyond minimalism (M) to explain the #ideophone sequences they argue M has little to say about. If one's goal truly is to minimize syntactic machinery, better be aware of —and point readers to— other approaches ☺️

Finally, if, as we have said, the sequences under consideration have a flat syntax, then it may be necessary to engage with a broader conception of grammatical structure, either within the generative framework or from compatible alternative frameworks. The former would include consideration of grammars that are not context-free or context-sensitive (e.g., regular or subregular grammars), while the latter would include work that considers indication and depiction alongside the well-studied mode of description (Ferrara & Hodge 2018), and work from typology and grammaticalization that shows why ideophones come to occupy the grammatical positions they do (Güldemann 2008). In both cases, there is a truly minimalist goal to achieve: to account for the observed structure by importing just the appropriate amount of syntactic machinery.

051

Mark Dingemanse @markdingemanse.net · 13/07/2026

Richting het gouden uur bij De Kaaij optreden met @kladderadatsch.bsky.social, wat wil je nog meer? 🤗😎🥁

foto met ernstig tegenlicht van de laagstaande zon laat een deel van Kladderadatsch zien tegenover een mensenmassa die mooi van achteren uitgelicht wordt

020