iiinova — where your digital self keeps evolving

Mike Anderson @mljanderson.bsky.social · 05/09/2026

Even engineers ought to know that screens in cars are dangerous, and harder to use than physical buttons and dials.

Mike Anderson @mljanderson.bsky.social · 27/08/2026

tldw; it’s better to understand brain-behavior relationships by tracking our causal embeddedness in the world than by searching for representations of it. And we have the tools to do it, in real time, with complex stimuli and ecologically relevant tasks.

000

Mike Anderson @mljanderson.bsky.social · 31/07/2026

Indeed. You have to teach them to think about *something*, formal logic notwithstanding.

030

Reposted by Mike Anderson

David Spurrett @doctorspurt.bsky.social · 30/07/2026

Here's a really nice paper by @edbaggs.bsky.social on "James J. Gibson’s affordance examples". The affordance idea is often glossed as about "possibilities for action". Anyone who actually reads Gibson will see many examples that don't fit that schema. psycnet.apa.org/fulltext/202...

psycnet.apa.org

APA PsycNet

3167

Mike Anderson @mljanderson.bsky.social · 02/07/2026

Much, much longer, I’m afraid

000

Mike Anderson @mljanderson.bsky.social · 26/06/2026

This is basically a visualization of language AI. So smooth. Until it isn’t.

050

Mike Anderson @mljanderson.bsky.social · 21/06/2026

Correct. Not an effort to police deployment of the word “representation”. But there are very many very different things at play here: registrations, emulators, mathematical models, diagrams, constraints, mechanical regulators, mental entities. Calling them all by the same word is not useful.

110

Mike Anderson @mljanderson.bsky.social · 21/06/2026

It’s a valiant effort, Nick. But if *every* transformation of A—>B requires representation, and every covariance is a representation, and cognitive scientists who try to sort this out (Ramsey and Shea I’m guessing?) are just spewing hot air, then this guy has placed himself out of reach of argument.

110

Mike Anderson @mljanderson.bsky.social · 20/06/2026

If he really doesn’t believe in the distinction between representations and causes, then it’s much odder than this…

000

Reposted by Mike Anderson

languagemit.bsky.social @languagemit.bsky.social · 20/06/2026

—— Open Mind is MIT Press’s Diamond Open Access Cognitive Science journal. We now have our first Impact Factor: 2.9 as of June 2026 Submit your papers to Open Mind direct.mit.edu/opmi

118372

Mike Anderson @mljanderson.bsky.social · 20/06/2026

If you don’t “believe in” the difference, then you can’t make sense of the distinction between direct and mediated perception. Which is fine. You can stick to your own playground. But then don’t make pronouncements about the impossibility of a phenomenon you haven’t taken the time to understand.

000

Mike Anderson @mljanderson.bsky.social · 20/06/2026

Sorry. That’s just a bad analysis. Please review the difference between causal and epistemic mediators.

100

Mike Anderson @mljanderson.bsky.social · 20/06/2026

We started with direct perception and now we are on direct action, whatever that is? But Ok. If that’s what direct means, then, yeah, no relationship in the actual world is direct. Who thought otherwise?

100

Mike Anderson @mljanderson.bsky.social · 20/06/2026

Thanks, Nick. I wasn’t going to bring that up because character limits and attention span. But, yes. Eco ideas have been influential in robotics in particular. As you say, part of the issue is that the concept of computation has become paper thin in order to cover all the ground.

010

Mike Anderson @mljanderson.bsky.social · 20/06/2026

This is a great example, since Gibson formulated his ideas while working for the Army Air Force trying to understand what information pilots used to land planes. (BTW, I worked as a mechanical engineer and was an AI scientist before my current gig. So I know a fair bit about devices)

100

Mike Anderson @mljanderson.bsky.social · 19/06/2026

And here I need to point out that computational devices were *not* built to solve problem X. They were built to compute missile trajectories and crack codes. So that we don’t have Ecological Devices strikes me as a poor argument. Do we actually have cognitive devices? Very unclear.

110

Mike Anderson @mljanderson.bsky.social · 19/06/2026

That strikes me as a substantive position. But I’m not sure it favors your view unproblematically. If I can build a device A that solves problem X using principles {Y}, that’s an existence proof. But to make an analogy to system B, you need reason to suppose it faces X. That {Y} solves X is not one.

100

Mike Anderson @mljanderson.bsky.social · 19/06/2026

I have no idea what “just an academic issue” means. It’s about getting a *theoretical* and *empirical* grip on a phenomenon. I think the Gibsonian approach has both, while the Helmholzian or Shannon approach misunderstands the problem.

100

Mike Anderson @mljanderson.bsky.social · 19/06/2026

“Without constraints, biases and filters it doesn’t work”. Indeed, it doesn’t! But “it” is an ill-posted problem of inverse inference that organisms don’t have to solve. Telegraph operators do to extract messages from signals. Not brains.

110

Mike Anderson @mljanderson.bsky.social · 19/06/2026

Right. “A residual trace, a signal, (yes, a representation!)” uses three non-synonymous terms in a row. It’s this slide that is troublesome.

111

Mike Anderson @mljanderson.bsky.social · 19/06/2026

Perceiving “chairness” is straight-up Plato. To @adw.bsky.social comment, it would appear that more attention to the theoretical/philosophical underpinnings of the “engineering” hypothesis you advocate would be useful here.

000

Mike Anderson @mljanderson.bsky.social · 19/06/2026

I’m less hung up on the vocabulary (if SCENE gets a lot of attention it creates niche awareness, as the marketers say) than the concepts. If you think organisms “represent” affordances then it’s really unclear what your IV is going to be. Not affordances, in any case.

000

Mike Anderson @mljanderson.bsky.social · 19/06/2026

But of course that’s not Gibson’s view at all. There are no “forms” or any hint of metaphysical or epistemological dualism. There’s a structured world that structures energy arrays that are picked up by other structures (sensory systems) in organisms navigating the structured environment. No mystery

240

Mike Anderson @mljanderson.bsky.social · 19/06/2026

That’s a super interesting take. It supposes Gibson to be a platonist claiming we directly perceive Forms. Even Plato thought we only did that while dead, and so it’s natural to suppose the senses are highly imperfect guides to the world, and any claim to the contrary is nuts.

100

Mike Anderson @mljanderson.bsky.social · 19/06/2026

It seems like a classic case of fitting the problem to the tool. “We can solve the problem of model inference in partially observable environments…. So perception is a problem of building models from incomplete data”

020

Mike Anderson @mljanderson.bsky.social · 18/06/2026

Should have registered the trademark while there was still such a thing as IP.

030

Mike Anderson @mljanderson.bsky.social · 18/06/2026

Kudos for advocating for ecologically valid environments and closed-loop behavioral tasks in neuroscience (see eg @culhamari-lab.bsky.social ). But why are we starting by *assuming* that inference is fundamental to perception-action coupling? Neither animals nor babies are “little scientists”.

170

Mike Anderson @mljanderson.bsky.social · 17/06/2026

At least they cite Gibson? But naturally nothing in recent Ecological Neuroscience, in which there is a growing body of research. Well, it’ll be a useful foil for The Cambridge Handbook of Ecological Neuroscience.

010

Mike Anderson @mljanderson.bsky.social · 17/06/2026

At least they cite Gibson? But naturally nothing in Ecological Neuroscience, in which there is a growing body of research. Well, it’ll be a useful foil for The Cambridge Handbook of Ecological Neuroscience.

130

Mike Anderson @mljanderson.bsky.social · 04/06/2026

If you haven’t encountered Dan Dennett’s “Quinjng quaila” you should most definitely check it out

100

Mike Anderson @mljanderson.bsky.social · 04/06/2026

This looks more like a move in a language game.

100

Mike Anderson @mljanderson.bsky.social · 04/06/2026

Not if you’re more of a Wittgensteinean philosophical neurosciency person.

110

Mike Anderson @mljanderson.bsky.social · 02/06/2026

Makes me realize I don’t know. Am I experiencing dreams in real time or am I remembering? And if I’m remembering does that mean the experience had to be conscious in the moment?

140

Mike Anderson @mljanderson.bsky.social · 02/05/2026

Dare we publish the “philosophical success metric”: [cleared up confusion/caused more confusion]?

040

Mike Anderson @mljanderson.bsky.social · 30/04/2026

Ironically, your Mark Twain quote is most probably Josh Billings (1874)

010

Mike Anderson @mljanderson.bsky.social · 23/04/2026

Not to mention the first sentence is false.

110

Reposted by Mike Anderson

Dio D. Vicen @diovicen.bsky.social · 25/02/2026

Luis Favela, @dewitmm.bsky.social, and yours truly talked to @braininspired.bsky.social about ecological neuroscience. Among other things, we tried to comment on how to make "affordances" and "ecological information" helpful for neuroscientists. Check it out! youtu.be/p987y_v-_PE?...

youtu.be

BI 232 How Should Neuroscience Integrate with Ecological Psychology?

YouTube video by Brain Inspired

082

Mike Anderson @mljanderson.bsky.social · 04/02/2026

Btw, do you mean the Turvey perception book (2018)?

110

Mike Anderson @mljanderson.bsky.social · 04/02/2026

Of course, keep in mind if Hohwy can make something Cartesian, he will…

010

Mike Anderson @mljanderson.bsky.social · 04/02/2026

Moreover it’s deeply mired in inferentialism and internal model building.

110

Mike Anderson @mljanderson.bsky.social · 04/02/2026

What’s your estimate for how many times we need to learn something before we learn it?

000

Mike Anderson @mljanderson.bsky.social · 21/01/2026

I get the sentiment, but it’s an overly broad judgment. Sometimes a student, in conversation for instance, offers a crucial insight and you invite them to help finish a paper.

110

Mike Anderson @mljanderson.bsky.social · 11/01/2026

The mistake seems to come from citing Cisek, who *does* say affordances are “represented”. But he’s using the neural meaning of representation which is much looser than the CTM definition. What Cisek says is compatible with neural resonance, which is a bona fide Gibsonian notion.

020