rigorousintuition.ca

This week we are joined by ethologist and best selling author Jonathan Balcombe to discuss his new book ‘What a Fish Knows‘. Balcombe’s research reveals just how wrong our assumptions of fish intelligence have been and uncovers the miraculously complex behaviour of our underwater cousins

Cordelia » Wed Aug 17, 2016 10:50 pm wrote:Sorry to hear about Spike 82.

When I was a kid, my brother became allergic to my cats; as compensation, my parents let me collect a menagerie of parakeets, finches, lizards, turtles, toads, goldfish etc.....

My mother was a fastidious housekeeper and I still don't understand how she tolerated all my birds and reptiles. Our living room had a cathedral ceiling where I'd let the parakeets fly. My bedroom smelled like a fucking zoo and once our neighbor freaked when she found my escaped horned toad in her garden. My brother remembers when, like one of the Spikes, one of my goldfish jumped out of its bowl and into the kitchen sink just as my mother turned on the garbage disposal; the fish went down the disposal but was thrown back into its bowl uninjured. So I guess not only do fish have feelings and loyalty; sometimes they can just be pretty damn lucky.

I think insects are also sentient.

Insect Awakenings.

The study looks almost perfect, if ridiculously low-tech: the kind of thing an undergrad might do with a budget of $3.50. All you need is a mirror, a piece of Plexiglas, and a bunch of ants.

Oh, and blue paint. The whole thing comes down to blue paint.

Start with the Plexiglas. Put Ant A on one side, Ant B on the other. Ant A shows no reaction to its buddy at all. So far so good.

Stick it in front of a mirror. Now it pays attention. Goes up to the glass, taps its reflection, shows interest it never showed with the real ant on the other side of the plexi. Interesting.

Maybe it’s reacting to something in the mirror, some chemical in the silver backing perhaps. So put a dot of blue paint on its head and put it in front of the mirror again. This time it checks out its reflection and starts grooming its head, as if to get rid of that weird-ass dot that just appeared there. It never tries to groom its reflection, which is where it actually saw the paint.

This is starting to get creepy.

Okay, um, maybe it could just feel the paint up there. Maybe it itched or something. So try a speck of brown, ant-coloured paint, something that won’t be visually obvious in reflection.

No grooming.

Put a speck of blue paint on the back of the head, where the ant can’t see it in a mirror. No grooming.

I’m not one to jump to conclusions, but I’m having a hard time interpreting these results in any way other than: ants recognize themselves in mirrors. Which means they pass a test frequently used as an index of self-awareness, a test that even some higher primates fail.

The stats seem sound, generally returning P-values of less than 0.001 (for the statistical neophytes in the crowd, that means the odds of getting those results by random chance are less than 1 in 1000). But the remarkable thing is, the researchers didn’t even do stats on most of their results. They couldn’t do stats, because there was no variation in the data. All the face-painted ants groomed their faces once they saw themselves in a mirror; none of the unpainted ones did. I don’t think I’ve ever seen such clean data in a behavioural study before.

What is this simple-yet-profound experiment, this rock-solid research with the batshit crazy results? Why, it’s “Are ants capable of self recognition?“, by Marie-Claire Cammaerts and Roger Cammaerts. Where will you find it? In the Journal of Science, a publication whose website veritably screams JunkWoo. The title bar on its website looks like a banner ad for generic penis pills. Just below that you’ll see, for some reason, a stock photo of a smiling dude in safety goggles and a yellow hard hat. “Instruction to Author” is either a typo or a tacit admission that every paper in the journal is written by the same person under different pen names. Even the journal’s name seems designed to encourage confusion with more respectable platforms (“the journal, Science?” “American Journal of Science?” “Journal of Science Education?”) while simultaneously discouraging investigation into its actual pedigree. (Google the phrase “Journal of Science”: you’ll get 67,000,000 hits. I scrolled through the first 300 and couldn’t find a single link to the actual journal.)

The Cammaerts are not flakes. They’re well-published in respected, peer-reviewed journals. I don’t know what they’re doing in the Journal of Science, unless they lost a drunken bet at a party somewhere. Or maybe their results are just so incredible that no one else would publish them.

I’m thinking maybe it’s that second thing. If you go to Wikipedia’s page on “Mirror Test”, pull back the curtain and read the backstage discussion, you’ll see editors and commenters stating that they “flat out don’t believe those results”, even while others praise the methodology that produced them. The idea that ants can self-recognize just opens too big a can of worms.

And yet, Cammaerts and Cammaerts are not entirely alone. Way back in 2010, writing in the top-of-the-line Proceedings of the Royal Society B, Björn Brems described research suggesting that fruit flies have Free Will. It wasn’t “free will” in the classic sense— it basically amounted to any behaviour complex enough to make you unpredictable to predators— but as anyone conversant with the literature will tell you, that’s pretty much the only kind of free will we humans can lay claim to as well (albeit with more bells and whistles). And just this year, Proceedings of the National Academy of Sciences ran a piece called “What Insects can tell us about the origins of consciousness“, by Andrew Barrona and Colin Klein.

Barrona and Klein argue that vertebrate consciousness is seated in the midbrain, which acquires information about both the organism’s internal state and its external environment. It integrates these into a model that generates behavioural goals— if your internal state is too hot then move somewhere cooler, that sort of thing— and relays those goals to the motor system. The midbrain contains all the elements necessary to sense, navigate, and survive in a given environment. Barrona and Klein argue that such integration is the root of consciousness, and point out insect brain structures serving the same functions; they conclude that insects should experience comparable levels of awareness. (Nematodes, lacking comparable structures, would not.)

It’s important not to go off the deep end here. There’s a huge difference between consciousness and self-awareness, between sentience and sapience (and a belated thankyou to Leonid Korogodski for hammering that difference home to me many years ago); an organism can have conscious experiences without consciously reflecting on their own existence. And the Mirror Test has always struck me as a questionable metric for self-awareness anyway (for one thing, it’s easy to envision an algorithm that recognizes the self without being aware of the self). But the traditional view of insects as mere computer programs wrapped in chitin, utterly deterministic in their behaviour, appears to be wrong. Stimulus A does not always provoke Response B, as you’d expect from purely deterministic reflexes; sometimes the insect is focused on other input, sometimes it can be distracted. We are learning that insects pay attention to things. It seems increasingly likely that their experiences are conscious ones, to at least some extent. Consciousness may be far more ancient, far more widespread than we ever suspected.

Which means that suffering is, too, by the same token.

I’m not sure why, but I bet that explains a lot.

BREMERTON — Chum salmon appear to have a cockroach-like resistance to pollution.

Scientists at Washington State University’s stormwater research center in Puyallup recently made the surprising discovery that chum are unaffected by the same levels of toxic road runoff that quickly kills their coho cousins.

“Overall, the results were very surprising because I thought chum would show some signs of getting sick, like we saw with coho — but that didn’t happen,” WSU ecotoxicologist Jenifer McIntyre said.

Also known as dog salmon, chum are the most plentiful salmon species in Puget Sound. They’ve maintained relatively healthy populations while coho and chinook numbers have plummeted.

Since 2012, the center has taken coho raised at a Suquamish Tribe hatchery and exposed them to stormwater captured from busy roadways. The coho showed almost immediate signs of illness and were dead within 24 hours.

The same toxic cocktail appears to have little affect on chum.

“Leading up to their deaths, (coho) grew lethargic and seemed confused, swimming erratically near the water’s surface and turning onto their sides,” McIntyre said. “(Chum) remained healthy looking and alert. Even their blood chemistry was relatively unaffected.”

The chum also were from the Suquamish Tribe’s hatchery on Grover’s Creek near Indianola.

The results match a decade of urban creek monitoring. Coho in Seattle-area creeks were observed dying at high rates after heavy or light rainfall. Rain picks up oil, heavy metals and other containments from roofs, roads and parking lots and funnels them into creeks.

Chum’s runoff resistance is a mystery.

“We do not have the answer, but I’m incredibly curious,” McIntyre said.

Chum are considered the most robust of the salmon species. Even as newborns, they leave protected freshwater streams and head to open water earlier than other salmon species.

They appear generally less sensitive to water quality. Even the problem of low dissolved oxygen, which has long plagued coho and other marine species in Hood Canal, is taken in stride by the resilient chum.

“Anybody that works with Pacific salmon will say chum are really tough fish,” she said.

They might be tough, but they’re not especially tasty. With flesh some describe as “mushy,” chum are unpopular with anglers and have a comparatively low market value. The ailing coho, on the other hand, is a prized catch.

McIntyre hopes to learn how other salmon species react to stormwater pollution. Their survival rates could be severely impaired as the sound’s urban areas continue to expand, McIntyre said.

But there is a solution. A four-year study by WSU’s stormwater center showed that simple filtration systems can greatly reduce stormwater runoff’s toxic effects. None of the coho exposed to runoff that trickled through a gravel, sand and compost filter died. Such filtration systems are often incorporated into rain gardens, bioswales and other landscape features.

“It was remarkably efficient,” said McIntyre after her study was completed last year. “It made the runoff nonlethal, and it didn’t even make the fish sick anymore. It shows us the remarkable potential of runoff filtration.”

Virtual Appearances, Virtual Realities

“Nothing is real.”
—John Lennon

So. The latest variant on the classic double-slit experiment continues to support what all other such experiments have supported in the past, namely that whether something behaves like a particle or a wave depends upon whether it’s being surveilled, which demonstrates in turn that (in the words of the study’s author) “reality does not exist if you are not looking at it”, which in yet another turn means that nothing makes any fucking sense whatever. I’ve always clung to the belief that it all does make sense, but — because stuff that happens on quantum scales has no relevance to the process of natural selection up here in the classical world— our brains simply aren’t wired in a way that lets us grok such things intuitively. But let’s put that aside for now, because I think I might have come up with an explanation for all this quantum dumbness that actually does make sense at classical scales:

Nick Bostrom is right. We’re all living in a simulation. More, these dual-slit experiments suggest that (and here’s my little contribution, which probably has its head up its ass because I don’t know anything about this stuff but bear with me) we’re living in a simulation with a really low budget.

Most of you already know Bostrom’s argument. For the rest of you, his reasoning boils down to 1) if it’s possible to create simulations with self-aware inhabitants, and 2) if some advanced species is inclined to actually build such sims, then 3) there are probably way more simulated universes (>>1) than real ones (=1). Which means, based purely on the odds, that we are far more likely to be living in one of a myriad simulations than we are in a singular baseline reality.

I’ve always liked this thought. It fits in nicely with the whole Digital Physics paradigm that seems to have taken root in the Physics community. It jibes with the way reality seems to kinda “stop” below a certain scale of resolution (Planck Length and Planck Time may be no more than pixel dimensions and clock cycles). And if enough studies like Bean et al come down the pike— and if they hold up— the Simulation Hypothesis might even find its way out of the it’s-untestable-so-it’s-not-science swamp that’s mired String Theory for so long.

So what do we know about our own baby steps into building simulated realities? We know that when you’re playing Fallout 4, the graphics engine doesn’t waste energy rendering the landscape behind you. We know that when you put on your Oculus Rifts, they don’t paint the vista at the back of your head. Why should they? Nobody’s looking there. Oh, they keep all the pointers and variables on hand, sure. They’re completely capable of rendering the world to your left the moment you turn your gaze in that direction. But they don’t actually solidify any part of the world until someone looks at it.

You see where I’m going with this, right?

First-person gaming is a pretty good metaphor for quantum indeterminacy; nothing is real until observed. Maybe it’s more than a metaphor. Maybe we’re living in a cut-rate Bostrom sim, one that can’t afford the computing power to render everything in hi-res detail all the time— so it cuts corners, saves cycles only for those parts of the model that are being observed.

Maybe there’s nothing insane or counterintuitive about quantum indeterminacy after all. Maybe it’s perfectly understandable— depressingly familiar, even— to anyone who’s lived on a grad-student budget.