50 Comments
User's avatar
Garrison Lovely's avatar

Yeah, the evidence for fish pain is very strong — I wrote a piece on this topic for Vox a while back! https://www.vox.com/future-perfect/23639475/pescetarian-eating-fish-ethics-vegetarian-animal-welfare-seafood-fishing-chicken-beef-climate

Expand full comment
salvador's avatar

Fishes suffer up to 22min of intense pain when taken out of water, this is a fact

Expand full comment
Vidur Kapur's avatar

Thanks for looking into this - I covered very similar ground a couple of years ago and ended up at 60% that fish feel pain. I've updated a bit in your direction.

Expand full comment
hamsters's avatar

Hi, pescatarian here (potentially full vegan in the future), need to do more research.

1) Call me biased but I find the arguments that fish reacting to negative stimuli to be quite weak. I don’t think smth like seeking antibiotics is proof of pain, that is, suffering from these negative stimuli. It’s very much possible that they don’t have the capacity to process the negative stimuli as pain (as posed by the cortex arg), and feel only something that must be acted upon.

2) Question for you, I know ur a utilitarian but say I could press a button and painlessly transform a cow into 900 steak dinners. Does the life of the cow as a standalone have any bearing on your choice to go vegan? And do you think the utility of 900 steak dinners outweighs any potential utility possibly taken by the cow. I think you see what I’m getting at, is any form of (practical) humane fishing permissible by your ethic?

Expand full comment
TheKoopaKing's avatar

> But if a creature behaves in various ways that make far more sense if they experience pain than if they don’t, then you should think they probably feel pain.

BB, your computer behaves as if it has the nuclear launch codes. It probably uses a password, has security mechanisms to prevent illegal accesses to processes, and has buttons to press that result in further computations down the line. Your should probably believe your computer has the nuclear launch codes.

Expand full comment
Matt Whiteley's avatar

My question would be what it actually means to talk about feeling pain without 'knowing' that you are feeling pain. Even if we assume this behavioural evidence indicates consciousness, at what point is the 'qualia' of a pain experience just a qualia that exists as a kind of fact rather than a moral responsibility we are beholden not to create? Suffering is existentially bad in humans, it's not just bad because pain qualia equals bad on some kind of moral qualia scale.

Expand full comment
Bentham's Bulldog's avatar

They just seem different. Having an experience is different from being aware that you are a creature in the world who has that experience on account of annually feeling pain. I imagine early infants don't have self awareness but can feel pain. Sometimes when I'm in pain, while I'm in some broad sense aware that I exist, that fact isn't on my mind.

Expand full comment
Matt Whiteley's avatar

Right, but in a fish I would assume there is no conscious meta connection between experiences, so in what sense are they not just isolated facts rather than moral content? Can you really say "they feel pain" or isn't it just "there is pain", and if the latter, can we use the word "conscious" in the same way?

Expand full comment
Bentham's Bulldog's avatar

I don't know what you mean by conscious meta connection. I don't see how this is different than infants.

Expand full comment
Matt Whiteley's avatar

I don't remember feeling any pain as an infant, do you?

Expand full comment
Bentham's Bulldog's avatar

You don't think infants feel pain?

Expand full comment
Matt Whiteley's avatar

I think my moral treatment of an infant has to do with their personhood, not to do with the degrees to which they may experience pain qualia.

Expand full comment
J P's avatar

While you may be conscious of memories, memories aren’t required for consciousness

Expand full comment
Rajat Sirkanungo's avatar

Well done, stickman! I love posts like this from you. I wish you were more active on youtube.

Expand full comment
J P's avatar
1dEdited

Overall, very nice reply. The apparent responses to and seeking-out of analgesia are potentially quite compelling as you describe them (though I would guess the results are probably messier than your clean summary, like most of neuroscience, & biology more generally). I would now peg my probability of at least some fish experiencing pain as ~60-85% (I find confidence point estimates fairly useless).

A few notes:

>> First of all, no, slime molds don't have very impressive behavior.

You and Diggles seem to lack an appreciation for the impressive optimization computations slime molds demonstrate. These are difficult math problems for humans that they naturally solve with ease.

Eg, https://www.sci.news/biology/slime-mold-problems-linear-time-06759.html.

More here: https://www.perplexity.ai/search/how-difficult-are-the-problems-sGbpORJyQ6aabfsGpzij1Q#0

Not rocket science? Well, in fact, lots of rocket science is built upon solutions to very difficult optimization problems, as are many applied mathematical sciences and engineering fields.

>> If fish didn't respond to analgesics, that would clearly be evidence they didn't feel pain.

Not true: it would be evidence that the substance doesn’t serve as an analgesic for the fish, either because they’re not in pain or because it simply doesn’t relieve *their* pain. There are incredibly varied responses to pain medicines just in humans, let alone across different animal classes. For example, all pain medicines work for some people but not for others. Eg: https://www.npr.org/sections/health-shots/2011/09/26/140705557/when-it-comes-to-pain-relief-one-size-doesnt-fit-all.

Certainly one would expect some amount of evolutionary conservation, and the fact that they apparently respond to the provided analgesics is consistent with that expectation. However, if they didn’t, it might be simply because you haven’t provided what would be an analgesic *for them.* The lack of evidence would make the case for pain far less compelling, but it wouldn’t be *positive evidence* for a *lack* of pain.

>> gamma ray oscillations underlie pain

Not sure if this was a joke, but, if not, it should just be gamma oscillations (high frequency, generally population-level oscillations in brains). Gamma rays are high-energy photons.

>> This even holds for behaviors that seem to only fulfill conscious functions, like playing and dreaming. Such a result would be quite surprising!

The function of dreams is an active research question. There’s evidence that it helps with memory consolidation. One interesting recent hypothesis proposes it acts as a regularizer (in the machine-learning sense) to help animals avoid overfitting and aid in generalizing (https://www.sciencedirect.com/science/article/pii/S2666389921000647). Of course, one can have memory and learning without consciousness, like in AI systems.

Play in animals of course has classically been thought to help learn useful skills. I’m not saying either of these don’t fulfill conscious functions, just that assuming so (let alone that it’s all they do) is begging the question.

For example, a null robot-like model might still evolve something that looks like dreaming for memory and learning, and something that looks like playing for skill acquisition and development. One would have to argue that dreaming and playing as observed in animals isn’t consistent with this null model, rather than taking it as a given.

Expand full comment
TheKoopaKing's avatar

> In favor of mid-brain centric theories:

Are you really gonna make me read a whole book you motherfucker

Expand full comment
TheKoopaKing's avatar

>In favor of cortex-centric theories:

Trojan source alert:

"The principal structural features of this region are identified and then used as

biomarkers to infer whether fish are, at least, anatomically capable of feeling pain. Using

this strategy, I conclude that fish lack the necessary neurocytoarchitecture, microcircuitry,

and structural connectivity for the neural processing required for feeling pain."

Expand full comment
TheKoopaKing's avatar

>>for pain to exist in an organism, the brain of the organism must sustain the sorts of neurological features that allow for consciousness and pain perception. In humans, this is the cortex,

I feel like cortical remapping undermines most of the claim that pain processing has to be situated in a specific unchangeable location in the human brain. The debate will turn on whether or not fish can realize a subset of the same mechanisms involved in enabling feeling pain. I imagine studying cortical remapping would involve more of the "bare minimum" needed to feel pain but idk how that shit works.

Expand full comment
TheKoopaKing's avatar

> If a creature felt pain you’d expect them to:

I would expect a functional decomposition of their brain states that shows they are relevantly similar to how my brain states function when I am in pain. There are of course other ways to construct measures for this, like by behavioral proxies, but only if those proxies are in fact connected to pain.

Consider that if feeling pain is something that evolves after the capacity to have a central nervous system and to mechanically engage in pain behavior does, then there would have been creatures that could display the totality of pain behavior without feeling any pain. Those creatures would trivially fail the behavioral proxy criteria for feeling pain, because I so stipulated it. What's relevant for resolving whether they do feel pain is what the nature of that putative evolution from mere pain behavior to feeling pain was and whether fish's contemporary brains are relevantly like it, which is harder to establish purely via behavioral proxies.

Expand full comment
Bentham's Bulldog's avatar

You could, of course, think that pain serves no functional roles that differ it from non-pain behavior and that it evolves late. But I don't know why you'd think any of those things.

Expand full comment
TheKoopaKing's avatar

>pain serves no functional roles that differ it from non-pain behavior

The functional role feeling pain plays is that of communicating a shared neuronal structure to various other parts of the brain that are specialized in other things besides nociception like language or memory. It's in virtue of being shared among various modules of the brain that pain is conscious and not just mere motor inputs and outputs, like your brain's regulation of your heartbeat. And presumably the specialized module would evolve before being integrated into the shared superstructure that would allow it to become conscious. The feeling of pain isn't an extra something that attaches itself to random bits of pain behavior for no reason, as your commitment to dualism would probably make you think. Although this is of course a simplified heuristic model and evolutionarily the story could be totally wrong with respect to something like protopain.

Expand full comment
Vikram V.'s avatar

> But even a 10% chance that fish feel pain—and that we annually painfully slaughter a population roughly ten times the number of humans who have ever lived—is enough to make it a serious issue.

Fish pain is easily outweighed by the expected value of quark suffering. I simply don't have time to care about the foibles of the fishies, given the enormously serious situation that the quarks are (not) in.

> Various fish are also capable of transitive inference and pass the mirror test.

This is certainly more impressive form of memory. Reading the article, I concede that some level of self-reference and understanding of rivalry after memorizing the mirror could support at least a chance of consciousness.

> Various studies have found that when acid is injected into the lips of fish, they rock back and forth, rub their lips against the side of the tank, and eat less.

Evidence that fish can sense things. Yes! Evidence of consciousness? No, not really.

> The evidence comes from the fact that in scenario after scenario, fish respond exactly the way you’d expect them to if they felt pain.

This argument is wrong.

A. Its logic proves too much. You can generalize from pain to any other stimuli. Are we seriously saying that if an entity consistently responds to *any* stimuli, like seeing the color green, that's evidence of consciousness? No! Whether or not something can be stimulated is not coextensive with the existence of consciousness.

B. There is a ready alternative: a creature can react to main without being conscious through a purely mechanical process. This doesn't have to be an AI or purely procedural process to be non-conscious. If there's nothing indicating subjective qualia specifically, as opposed to just chemically induced reactions, then you don't have evidence of consciousness.

C. All of the 8 things you list as evidence are entirely compatible with non-conscious pain responses. Fish obviously react to pain. There just isn't a consciousness to subjectively experience it.

> For example, if a person’s DNA being at the crime scene is evidence they committed the crime, then their DNA being absent must be evidence they didn’t do it (though not necessarily proof). If fish didn’t respond to analgesics, that would clearly be evidence they didn’t feel pain. So the fact that they do must be evidence that they do feel pain!

Ok, but this is more like a confession. Obviously, if a person confesses to a crime, that's strong evidence of guilt. However, the lack of a confession is not very strong evidence for innocence. I would contend that the criteria you describe are in a similar mold, but overwhelmingly weaker. So a very very small credence bump from this argument.

Expand full comment
Michael's avatar

There’s a lot of undue skepticism here, but I will stick to the rubbing situation. We know it was pain because they didn’t rub in response to the control non-painful but uncomfortable saline injection (for which they showed signs of mere sensation) and they rubbed substantially less when given both bee venom and anesthetics or acetic acid and anesthetics.

Expand full comment
Vikram V.'s avatar

Clearly, venom produces a different biological reaction than saline when injected. All the systems in the fish that prime it for action/rubbing activate. That biological difference leads to a uniquely powerful response, as compared to saline.

What does any of this have to do with consciousness?

Expand full comment
Michael's avatar

Because it’s behavior associated with pain? And opioids caused the behavior to cease. Opioids modulate pain. Like, how much more pain-indicative could this situation be? Similarly, fish flap around violently when taken out of the water, what we might recognize as writhing in pain. When a fish is doing everything in its power to indicate that it’s in pain, why not treat it like the fish is yelling “help me, I’m in pain!?”

Expand full comment
TheKoopaKing's avatar

>fish flap around violently when taken out of the water, what we might recognize as writhing in pain.

Seems to me more like they're suffocating. I've also seen fish get punctured, play dead, and then writhe around. Idk how to think about the fish's brain in these states.

Expand full comment
J P's avatar

While I agree with your previous responses, you fall directly into the anthropomorphizing trap with the flapping-about argument. They may or not be in pain. It could be a simple escape response, as they might do in the ocean when they reach a very low-oxygen location to more-quickly find a higher-oxygen zone. If analgesics reduce this behavior, that would be a more convincing argument

Expand full comment
Michael's avatar

I’m not sure that’s been tested since most research on the level of pain fish experience when flopping due to CO2 accumulation take for granted that fish sentience is a proven fact (as this is the majority opinion in the field). But studies have been done quantifying fish pain while flopping.

Expand full comment
Vikram V.'s avatar

Again, I am not disputing that chemical signals are being sent, and that reactions to them are happening. I am questioning why those reactions are indicative of *conscious* experience. Obviously opioids that interfere with neurotransmitters will change that phenomena, but that still does not show subjective experience.

> When a fish is doing everything in its power to indicate that it’s in pain, why not treat it like the fish is yelling “help me, I’m in pain!?”

The fish is not doing things to indicate that it’s in pain. The flopping is not a signaling mechanism. The flopping is an automatic muscle response to stimuli. There is no agent deciding that they are going to “indicate” that pain exists.

> “ why not treat it like the fish is yelling ‘help me, I’m in pain!?’”

This is absurd on its face. A fish obviously could not comprehend this statement even if it was conscious. The statement is also not an automatic response to pain, but rather a communicative mechanism that indicates a subjective second-level awareness that the self is experiencing pain, as opposed to the first order experience of pain. If a fish were screaming this, I would be open to saying it’s conscious. They are not.

Expand full comment
Michael's avatar

Is a second order awareness of pain necessary for a first order experience of pain? (No)

Expand full comment
Vikram V.'s avatar

No. Which is why I pointed out that equating what a fish does to communicating a second order awareness is unnecessary and absurd.

It seems like you just introduced an extraneous concept, then attacked my refutation as being irrelevant. You are the person who made this an issue!

Expand full comment
Charles Amos's avatar

Can't I carry on eating wild caught Cod because the counter factual is them being suffocated in the stomach of another fish; which I imagine is worse than suffocating?

Expand full comment
J P's avatar
1dEdited

This seems utterly unconvincing. Glenn assumes the existence of fish is net-negative, and that more fish is therefore further negative. The logical conclusion to this would be for us to eliminate all wild fish, and only keep those around we can coddle to have net-positive experiences. Utter rubbish—we would destroy all wildlife with such logic.

Note the silliness here: if their lives were so miserable that they shouldn’t even spawn, then they wouldn’t avoid deadly circumstances, since dying for them would be better (unless they are in fact not conscious and such evasion is simply programmed behavior, which is against your thesis of fish consciousness).

In fact, such an argument could be used in the other direction: the earlier these net-negative creatures die, the less suffering and therefore the better for each individual creature. Utilitarianism is such a silly ethical theory that we can’t use ridiculous “total suffering,” integrated over all animals, to guide us here.

But even if we did, if one were to cut a fish’s suffering in half by eating it at half its expected longevity, the extra fish that will last the other half of its life will have the extra half of the suffering. Ie, one could argue you’ve neither increased nor decreased total suffering.

Perhaps there’s a better argument against Charles Amos’ suggestion, but Glenn’s is very much not anything close to a good argument

Expand full comment
Bentham's Bulldog's avatar

If you think that fish live good lives then the argument against eating wild fish is far more straightforward!

Expand full comment
J P's avatar
1dEdited

🙂 Not entirely. Then the argument flips per Glenn’s logic: if animal 1’s life is net positive, then killing it (esp without making it suffer) early to replace it with animal 2 can double the “happiness.”

Also, they provide health benefits (eg, cardiovascular health from omega 3s) that so far have not been replicated synthetically (supplements seem to lack the protection, and they’re derived from fish anyhow). Unclear how to compute the ethical effects of that.

But yes, in principle I entirely agree that we should minimize and eventually eliminate our dependence on meat. However, I think that, practically, that requires massive amounts of cheap, high-quality lab-grown meat (we’re getting there but still a long row to hoe). Until then, we have simply co-evolved far too closely with such food sources for far too long to eliminate them without such a technological substitute, even if some small fraction of the population is admirably able to make the sacrifices that take away from their net happiness.

In fact, I would argue that anyone making those sacrifices for such principled ethical considerations is barking up the wrong tree. If you truly have a deep conviction that we can’t be a properly ethical species until we give up meat, then you shouldn’t dedicate your life to veganism, you should commit to being part of the solution, and work in the lab-meat space (or other areas that space needs advances in). Until we solve that, 90%+ of humanity will continue consuming animals and their products (indeed, if my first-choice career doesn’t work out, I will seriously see if I can work in that space!).

Expand full comment
akash's avatar

> if their lives were so miserable that they shouldn’t even spawn, then they wouldn’t avoid deadly circumstances, since dying for them would be better

I think most living beings, once alive, do not wish to die. This is true for humans as well. From a evolutionary biology perspective, this makes complete sense. If animal species would commit suicide based on the expected utility of their lives, those species wouldn't survive; but evolution optimizes for gene propagation, not pain/pleasure.

> The logical conclusion to this would be for us to eliminate all wild fish, and only keep those around we can coddle to have net-positive experiences. Utter rubbish—we would destroy all wildlife with such logic.

Why would that be the only logical conclusion? One possibility could be just moral cluelessness given how complicated and vast wild animal suffering is (basically what I practically endorse for the time-being); another could be, as David Pearce et al. suggest, bioengineering beings so that they suffer less or predate less. Yet another could be, reducing diseases in fish that make their lives more miserable (we already try to do this in farmed and certain conservation settings).

I also think you are indexing too much based on how the conclusion makes you feel. Certain moral claims might sound completely unintuitive but could still have some truth value!

Expand full comment
J P's avatar
1dEdited

I advise against mind-reading of people in general, let alone of a stranger you’re only interacting with over the internet. My argument has nothing to do with how I feel about Glenn’s conclusion.

>> I think most living beings, once alive, do not wish to die.

You are arguing for my point. If they really “don’t wish to die,” than having more of them would be good for those that would come into existence, which is exactly contrary to Glenn in the link Bentham posted as a response. Perhaps you should read it

>> One possibility could be just moral cluelessness

I don’t see how moral cluelessness is a possible conclusion from Glenn’s argument. He certainly doesn’t think so either from what I read.

You’re certainly right in principle re: technological solutions, but I would be extremely cautious about such solutions applied at scale in nature, as they can easily go awry and lead to negative and even catastrophic consequences, especially with genetic engineering. Such consequences are already observable from applying these in the farming scenarios you point out (eg, massive dosing of antibiotics is helping lead to the reduction of efficacy of precisely those antibiotics and to the growth of the related superbugs). “Predate less” would generally mean fewer of the predatory species—is that positive? Unclear. You could of course feed them lab meat in principle, but that would require a massive global undertaking to have any meaningful effect. Moreover, doing so at scale for long enough would evolutionarily change those species, with unforeseeable consequences, like with genetic engineering.

In summary, I don’t find any of this plausible at scale anytime soon, and if any of it is ever undertaken, it should only be done with extraordinary caution and very, very slowly. I could certainly see an argument for such artificial interventions in that heavily constrained light. But, frankly, I think people who argue for such interventions mostly demonstrate a deadly combination of complete arrogance of humans’ abilities to control nature and total naïveté regarding the immense complexity of nature and thus the nearly unfathomable difficulty of such an undertaking. A sort of species-level Dunning-Kruger effect, if you will.

Expand full comment
akash's avatar

Likely not Cod, but someone on the Forum recently made a case for eating sardines and anchovies https://forum.effectivealtruism.org/posts/MvXbFB2Hhgq46toye/a-vegan-case-for-eating-sardines-and-anchovies

I am sympathetic to the idea that high-welfare beef is ok if you must consume meat but also care about animal welfare

Expand full comment
Tom Hitchner's avatar

The headline makes it seem like you’re pro-fish pain

Expand full comment