Key, Brian (2016) Why fish do not feel pain. Animal Sentience 3(1) (after reading the article, read the abstracts of some of the commentaries too, for contrary vuews).
Only humans can report feeling pain. In contrast, pain in animals is typically inferred on the basis of nonverbal behaviour. Unfortunately, these behavioural data can be problematic when the reliability and validity of the behavioural tests are questionable. The thesis proposed here is based on the bioengineering principle that structure determines function. Basic functional homologies can be mapped to structural homologies across a broad spectrum of vertebrate species. For example, olfaction depends on olfactory glomeruli in the olfactory bulbs of the forebrain, visual orientation responses depend on the laminated optic tectum in the midbrain, and locomotion depends on pattern generators in the spinal cord throughout vertebrate phylogeny, from fish to humans. Here I delineate the region of the human brain that is directly responsible for feeling painful stimuli. 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.
Only humans can report feeling pain. In contrast, pain in animals is typically inferred on the basis of nonverbal behaviour. Unfortunately, these behavioural data can be problematic when the reliability and validity of the behavioural tests are questionable. The thesis proposed here is based on the bioengineering principle that structure determines function. Basic functional homologies can be mapped to structural homologies across a broad spectrum of vertebrate species. For example, olfaction depends on olfactory glomeruli in the olfactory bulbs of the forebrain, visual orientation responses depend on the laminated optic tectum in the midbrain, and locomotion depends on pattern generators in the spinal cord throughout vertebrate phylogeny, from fish to humans. Here I delineate the region of the human brain that is directly responsible for feeling painful stimuli. 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.
Key claims that fish do not feel pain because they lack the neuronal structure required to have pain. Of those possible objections, most boiled down to this question and answer:
ReplyDeleteCritic: "How can we say that fish don't experience another kind of pain with another kind of structure?"
Key: "Because from our experiments, we have an idea of what structures *cause* pain in humans. Because we have this causal mechanism, we can generalize this finding across species."
You can still say that all neuroscience has done is correlate, and that’s not enough. Generalization to species requires causal theories, and those require axioms and proofs. However, even if neuroscience is only correlation, it could still be enough.
Watching apples fall in Cambridge leads me to correlate being unattached with falling. That correlation is enough to lead me to deduce that apples will fall if left in mid-air at McGill. However, that correlation not enough to answer whether or not an apple will fall on the moon - I need a causal theory for that, or to go to the moon.
Is generalizing pain structures in fish akin to generalizing apples that fall at Cambridge to those that fall at McGill or on the moon? We don't know because we don’t have causation. But we do have lots of correlation – and that gives us a guess. According to Key, it's a pretty good one too.
Please look at some of the commentary on Key. It is not about "another kind of pain." It is about pain, felt through another kind of brain structure. There are many neural (and anatomical) structures that generate the same function by different means. Key says fish don't feel pain because they don't have the mammalian structure. Most of the commentators reject this (rightly). Why? Think about it. Don't rush...
DeleteThis video is from one of my other classes and I think it talks about a couple of things that we discussed this term and touches on the hard problem. Putting it on this week's blog in hopes that more people will see it.
ReplyDeletehttps://www.youtube.com/watch?v=HW_4QArGScg&ab_channel=TEDxTalks (watch from 12:55 to the end for most relevant part)
Monti talks a bit about what consciousness is, and favors the idea of evaluating it as a spectrum rather than a clear cut conscious/unconscious and then talks about the key to figuring it out: being able to measure consciousness. I got excited because I thought he might have a solution, unfortunately he said he doesn't. But then he goes on to hint at the Turing test (although maybe more from a hard problem perspective) when he asks if our phones are conscious or whether they could ever be conscious. He said he asked his phone which I found really amusing because that's one of the first things I did when we started talking about whether a robot can feel (although these are probably just computers not robots). I'm glad people much smarter than me have also had that intuition to want to ask our existing AI to see where we're at right now.
After knowing me from 30 hours of Psych 538, I guess most of you can predict my reaction to Monti's TedTalk:
DeleteLike most TdT's it is thuddingly superficial if you can make it to the end without being turned off by the feeling that someone is trying to sell you a car!
1. Lots of C/F weaselling back and forth (can you spot it?). not just between feeling and "being conscious" but even about "waking up" out of a coma (as opposed to just the restoration of the physiology of what is normally the sleep/wake cycle).
2. No, feeling is not a matter of degree; no "spectrum." At any point in time, you are either feeling something (anything) or there is no feeling go on, and hence no feeler. You can feel more kinds of things or fewer kinds of things. The feeling can be more intense or less intense. Feeling can be continuous or intermittent. But, at any point in time, feeling is either being felt or there is no feeling nor feeler. That's all-or-none, not a matter of degree -- even if, 5 minutes later, the feeling does start up again. When people speak about "levels" or "degrees" of consciousness, they are just weaselling about alertness or intensity or quality or continuity; but what is really at issue is whether there is any feeling going on at all.
There is definitely some scope for voluntary control over non-motor actions, such as imagining; and if a paralyzed patient can do that, you can in principal talk to them about anything in a binary code (like morse-code) and they can tell you whether and what they feel. (But you have to be careful for the pitfalls of "facilitated communication.")
3. Chatbot trivia: The less said about that, the better...
PS: The "ask 'how' not 'where'" seems to me to be just a play on words (rather than a call for reverse-engineering). Do you have another understanding of it?
Delete(BTW, this topic belongs in week 10, not week 11.)
It's funny how much perspective plays into this. I watched this video thinking wow a lot of things line up with 538 like considering consciousness as feeling and mentioning that a key part of solving the hard problem is to find a way to measure consciousness. I'm sure you took a more critical approach, looking into how much of it doesn't align with this course. Yeah the spectrum part of the argument sort of felt like the protolanguage argument, like it's either a language or it isn't, it's either feeling or it isn't. But I guess he did mean it in terms of how much feeling but then we'd still run into the problem of determining how much feeling is "enough" feeling which will bring us right back to a yes/no threshold. Yeah I don't think the chatbot trivia is giving us any important answers, I just found it fun.
DeleteI'd like to believe that looking at brain activity, especially when the prompts produce distinctly different activity patterns, would be more objective than guiding someone's hand (I think that's the study the facilitated communication paper was talking about). I've always gotten really excited about the potential ability to communicate with paralyzed patients who are in vegetative states because it brings a lot of hope back to a lot of people but I'm going on a tangent here.
Hahahah yeah he doesn't have the most exciting presentation style I agree. I had to stick around because it was assigned for a class but left the timestamp I thought was most interesting because I didn't imagine many people would
For the ask how not where, it bothered me that he was asking where to begin with. I thought at this point in time in neuroscience, we had abandoned things like "looking for the location of consciousness in the brain". It feels like an unwritten rule in the field that no cognitive ability/function is exclusively "in one place". I think understanding how our neurochemistry gives rise to consciousness (if we can ever actually figure that out) is important for solving the hard problem but I also think that until we have a way to actually measure consciousness we shouldn't be going in that direction, and hence should leave consciousness out of our reverse engineering pursuit to solve the easy problem
Delete"until we have a way to actually measure [feeling]"
DeleteAll we'll ever be able to measure is correlates of feeling -- which just goes to show that the HP is not the fault of the OMP...
"Like most TdT's it is thuddingly superficial if you can make it to the end without being turned off by the feeling that someone is trying to sell you a car!" I very much agree with this about Ted talks!
DeleteThe first idea that I would like to address is nocifensive behaviours and I am going to try to connect this to the hard problem. As the paper introduces, these are the behaviours that occur non-consciously – without feeling (the feeling could come later). In a concrete example that we have also discussed in class, this could be something like retracting your hand from a hot surface. If you accidentally rest your hand on a hot stove-top, you may find that you have withdrawn your hand before you get the feeling of pain. This is one demonstration of why the hard problem is hard. As Professor Harnad mentioned in a reply to one of the skywritings of last week, this seems like a perfect solution for lazy evolution – to go right from the stimulus to the behaviour in one step. So what causal role could the feeling have when this nocifensive behaviour shows that the reflex arc doesn’t require it? I am aware that there are limited examples of nocifensive behaviour and certainly not all of our behaviour is this way, but I think that it illustrates the problem that there aren’t any degrees of freedom left for feeling to explain.
ReplyDeleteOne reply to Key’s paper is that of Chella, and I have taken the following line from the abstract of this response: “This is in line with the thesis proposed by Key: what biomimetic robot fish show is that it is difficult to explain what causal role sentience would add to their robotic capacities.” (Chella, Antonio (2016) Robot fish do not need sentience. Animal Sentience 3(6))
The point that I want to make here is simply that I too cannot explain the causal role of feeling on the “doing” abilities of fish (or humans!), yet this certainly does not lead to the conclusion that fish do not feel (Chella does not arrive at this conclusion either as stated in the last paragraph of the reply). I cannot explain why humans have feelings, but I know that they do! So I am not completely clear on why Chella states this is necessarily in line with Key’s thesis.
Hi Stephanie,
DeleteA quick challenge for your last point - how do you know that others feel? Can you read their minds? Or could it be that they're all just zombies going around?
Knowing is a tall order - it requires being able to have enough or the kind of information that would reduce uncertainty to zero. Fact of the matter is, because of the other minds problem, we don't have that kind of information until we can discover a Searle's periscope for each other. Even despite heterophenomenology's claims, their measurements won't ever produce enough information to make such a claim.
So, maybe we can know some things without being able to explain them, like why do *I* feel. But I suspect the question of why/how others feel is moot without that information.
Although... maybe, if we could answer why/how *I* feel, we'd be a touch closer to solving the Other Minds problem too.
Julian:
DeleteUncertainty zero is just in maths and Cogito. Scientific hypotheses are just probably true.
That said, is your uncertainty about whether other people feel that much greater than that apples fall down rather than up? Babies? Chimps? Dogs? Birds? Mice? Fish? Octopus? (I'd say the uncertainty is only begins to become nontrivial with the simplest invertebrates, microbes and plants -- and or course with toy robots, until we get closer to a T3 robot like Ting).
I would also like to briefly write a comment based on some of the replies I read to Key’s article.
ReplyDeleteSegner, Helmut (2016) Why babies do not feel pain, or: How structure-derived functional interpretations can go wrong. Animal Sentience 3(26)
One of the recurring themes in this class is how studying the structure of the brain is not going to tell us how it functions. As we have also discussed, it is possible to use brain imaging techniques in order to find correlations between certain regions/structures of the brain with their functions (e.g. language, maybe even pain). But why does it need to be a columnar structure in order to give the feeling of pain? These might be great structures to integrate or inhibit certain circuits, but how does integration make me feel pain? As Segner points out, the wulst structure of some birds is not columnar, yet there is evidence that birds have consciousness (feelings). I do not agree with Key that we can say lacking a columnar structure is evidence that an organism cannot feel pain.
The final point I would like to make is related to Mather, Jennifer A. (2016) An invertebrate perspective on pain. Animal Sentience 3(12). After reading about the experiment done on squids where part of their arm was amputated, I looked into the study (see citation below) and found an ethical note saying that “In the US invertebrates are not covered by IACUC regulations; thus, no welfare protocol was required for this study.” The authors then go on to explain what protocols they did follow, but this exclusion of invertebrates from welfare protocols seems very problematic to me. Mather’s response shows that evidence exists supporting the idea that invertebrates could have feelings.
Peripheral injury induces long-term sensitization of defensive responses to visual and tactile stimuli in the squid Loligo pealeii, Lesueur 1821. Robyn J. Crook, Trevor Lewis, Roger T. Hanlon, Edgar T. Walters. Journal of Experimental Biology 2011 214: 3173-3185; doi: 10.1242/jeb.058131
An even more interesting target article on sentience in invertebrates is this one, plus all these commentaries!
DeleteMikhalevich, Irina and Powell, Russell (2020) Minds without spines: Evolutionarily inclusive animal ethics. Animal Sentience 29(1)
Commentaries (including Key's! -- Authors' Response will appear soon):
Levy, Neil (2020) It might not matter very much whether insects are conscious. Animal Sentience 29(2)
Figdor, Carrie (2020) Relationship between cognition and moral status needs overhaul. Animal Sentience 29(3)
Browning, Heather and Veit, Walter (2020) Improving invertebrate welfare. Animal Sentience 29(4)
Vonk, Jennifer (2020) No room for speciesism in welfare considerations. Animal Sentience 29(5)
Cammaerts, Marie-Claire (2020) Invertebrates should be given ethical consideration. Animal Sentience 29(6)
Soryl, Asher (2020) Invertebrate welfare in the wild. Animal Sentience 29(7)
Monsó, Susana and Osuna-Mascaró, Antonio J. (2020) Problems with basing insect ethics on individuals’ welfare. Animal Sentience 29(8)
Balcombe, Jonathan (2020) Intuition and the invertebrate dogma. Animal Sentience 29(9)
Elwood, Robert W. (2020) Do arthropods respond to noxious stimuli purely by reflex?. Animal Sentience 29(10)
Zentall, Thomas R. (2020) Cognition, movement and morality. Animal Sentience 29(11)
Lockwood, Jeffrey A. (2020) Virtue ethics and the likelihood of invertebrate suffering. Animal Sentience 29(12)
Veit, Walter and Huebner, Bryce (2020) Drawing the boundaries of animal sentience. Animal Sentience 29(13)
Mallatt, Jon and Feinberg, Todd E. (2020) Sentience in evolutionary context. Animal Sentience 29(14)
DeGrazia, David (2020) On the possibility of invertebrate sentience. Animal Sentience 29(15)
Woodruff, Michael L. (2020) Whether invertebrates are sentient matters to bioethics and science policy. Animal Sentience 29(16)
Vallortigara, Giorgio (2020) Lessons from miniature brains: Cognition cheap, memory expensive (sentience linked to active movement?). Animal Sentience 29(17)
Baracchi, David and Baciadonna, Luigi (2020) Insect sentience and the rise of a new inclusive ethics. Animal Sentience 29(18)
Marino, Lori (2020) Sentience in all organisms with centralized nervous systems. Animal Sentience 29(19)
Lee, Andrew Y. (2020) Does sentience come in degrees?. Animal Sentience 29(20)
Howard, Scarlett R. and Symonds, Matthew R.E. (2020) Ethical considerations for invertebrates. Animal Sentience 29(21)
Marzluff, John M. (2020) Bridging the empathy gap for invertebrates. Animal Sentience 29(22)
Dietrich, Eric and Fox Hall, Tara (2020) Moral treatment for all. Animal Sentience 29(23)
Key, Brian and Brown, Deborah (2020) Minds, morality and midgies. Animal Sentience 29(24)
Brown, Culum, Prof. (2020) Convergent evolution of sentience?. Animal Sentience 29(25)
Chapouthier, Georges (2020) Invertebrate cognition, sentience and biology. Animal Sentience 29(26)
Broom, Donald M (2020) Brain complexity, sentience and welfare. Animal Sentience 29(27)
Birch, Jonathan (2020) Zones of precaution. Animal Sentience 29(28)
Fitzpatrick, Simon (2020) Avoiding anthropocentrism in evolutionarily inclusive ethics. Animal Sentience 29(29)
Forber, Patrick and Jones, Robert C (2020) Spineless and sentient: a challenge for moral comparison. Animal Sentience 29(30)
Cartmill, Matt (2020) Do beetles have experiences? How can we tell?. Animal Sentience 29(31)
Delon, Nicolas; Cook, Peter; Bauer, Gordon; and Harley, Heidi (2020) Consider the agent in the arthropod. Animal Sentience 29(32)
Clarke, Steve (2020) When some animals are more equal than others. Animal Sentience 29(33)
Gibbons, Matilda and Sarlak, Sajedeh (2020) Inhibition of pain or response to injury in invertebrates and vertebrates. Animal Sentience 29(34)
Hi Stephanie, I agree. As you said, "studying the structure of the brain is not going to tell us how it functions", so how can we know for sure that an animal that doesn't have the same nerve structure as us does not feel pain? What if we discovered a group of aliens who seemed like intelligent beings, but they had a completely different internal structure to ours? Could we really say that since they don't have this columnar structure they definitely don't feel pain? Couldn't there be other internal structures that replicate similar feelings of pain? We still don't understand how and why organisms feel, so why take this great risk and potentially cause suffering to so many animals?
DeleteI have watched recreational fishing once when I was very young, and seeing the hook through the fish’s mouth definitely did not sit right with me. When I asked if the fish was hurt, the answer was no, fish do not feel pain like we do. This answer was confusing - if I had a hook stuck through my lip and was thrown back into the water, I would most certainly feel pain.
ReplyDeleteReading Key’s “Why fish do not feel pain” and the responses to the paper has only strengthened my original intuition that fish do feel pain. Key argues that since fish do not have the same structural features of regions attributed to pain in humans, they must not feel pain. Any behaviours that indicate a fish feeling pain are really just reflexes, and the fish does not feel anything at all. I would argue that Key places too high an importance on his structure-function analysis of pain. Relating back to the “would you kick Ting” question, we have established that we do not need to split someone open to assume that they are feeling pain. If we have sufficient evidence that fish have an ability to sense damage, or nociception, then we should also assume that fish feel pain. We may never be able to feel someone else’s feelings, but if empathy works well enough to “mind-read” other people then the same techniques can apply to fish. Furthermore, from an evolutionary perspective, it makes sense that an organism would want to avoid damage, and pain is a way of telling the organism that something is wrong. Why would the reflex evolve without the feeling of pain? Calling the behaviours to avoid pain a “reflex”, or separating the idea of nociception from the idea of feeling pain, seems to me like a way to avoid an uncomfortable truth. Fish may not have the same neural structures as we do, but all evidence suggests that they feel pain. We should recognize this instead of trying to rationalize the harm we do to living things. It may not lead to a cessation of fishing altogether, but perhaps it would lead to a more mindful and sustainable approach to fishing - and isn’t this a good thing?
Yep, Key was resoundingly wrong on the OMP and the fact that fish feel pain.
DeleteBut "Why wouldn't the reflex evolve without the feeling of pain?" is still the HP, not just for fish, but for us...
I find it interesting that Key chose to investigate on pain, an emotion so strong that I would argue it to be the most obvious indicator of feeling. It's such a strong feeling, observable even in invertebrates, that it's hard to argue that reaction to noxious stimuli is independent of feeling. Yet somehow Key manages to come up with the conclusion that although fish have brains (an essential part in pain perception), they do not feel pain due to their different neural structure. From an evolutionary perspective, isn't it useful to pair the noxious stimuli with the feeling of "ouch" for fish to escape from predators? If Key claims that fish don't feel pain, then what would be the alternative explanation to the fish's motivation to escape? I agree with other commenters that feelings definitely play a role in the "doing" capacity in fish (and in us), which is why Key's dismissal of the presence of feeling is silly.
ReplyDeleteIf feelings play an evolutionarily helpful role for fish, can we argue the same thing for humans? I mean, it must be evolutionarily helpful to us if we can still feel right now, right?
Also to anyone reading this: please read Antonio Chella's peer commentary. I think this rebuttal is the most relevant to our course (the author collaborated with professor Harnad!).
If an organism is sentient (i.e., feels), then pain is obviously an adaptive feeling, and it is far-fetched to argue either that pain is not adaptive for fish, or that fish are not sentient.
DeleteBut the hard problem is to explain how and why sentience itself (i.e. feeling) evolved. (The fact that sentience evolved, and that the nervous system generates sentience (somehow) is not in doubt. But how? and why?)
In Key's article, he argues that other species like fish are sentient if they have the same biomarkers in the human brain that are directly responsible for feeling pain. He concluded that fish are unable to feel pain because they lack the neural structure required.
ReplyDeleteWhile reading the paper, I was reminded of a situation that occurred a few months ago. One afternoon, I noticed that my pet fish was lethargic and had stopped eating. I was in a state of panic because I did not know what caused her sudden behaviour changes. After doing some research on her visible symptoms, I bought her antibiotics to treat her internal infection. Fortunately, she made a speedy recovery!
Due to the other-minds problem, I could only infer what my fish was feeling because she could not communicate her discomfort to me. I was able to recognize from her behaviour that something was wrong. Although fish do not have the exact same neurobiological components as humans, many studies prove they do have the capacity to feel. I did a quick Google search and found some articles written by Dr. Lynne Sneddon. In 2003, she discovered that bony fish adapt their behaviour accordingly to pain. In her experiment, Sneddon injected acid into the lips of a rainbow trout to assess changes in behaviour. The fish reacted to the pain by “rocking side-to-side, rubbing its lips and increasing its gill ventilation rate.” Sneddon then injected morphine into the fish's system, and it significantly reduced pain-related reactions.
"How can we say that the fish was not just reflexively responding to the acid?"
Sneddon: "If that were the case, the morphine would not have made a difference. Therefore, the fish was consciously experiencing pain!"
Sneddon, L.U. (2003). "The evidence for pain in fish: The use of morphine as an analgesic". Applied Animal Behaviour Science. 83 (2): 153–162.
What are your thoughts on experiments like this? I can see how if we utilize this method in other contexts, maybe we can discover more species that are sentient. However, it does seem risky and unethical to perform if the researchers did not take careful consideration of the way the chemical is administered to the animal.
DeleteGood points, Ting. Sneddon was one of the 50+ commentators on Key (2016) and discussed the findings you mention. She also went on to publish a target article of her own in 2018, with several of her collaborators:
DeleteSneddon, Lynne U.; Lopez-Luna, Javier; Wolfenden, David C.C.; Leach, Matthew C.; Valentim, Ana M.; Steenbergen, Peter J.; Bardine, Nabila; Currie, Amanda D.; Broom, Donald M.; and Brown, Culum (2018) Fish sentience denial: Muddying the waters. Animal Sentience 21(1)
I am very troubled by any experiment that hurts or harms animals and I believe that those that do not help save lives are completely unjustified. Is mildly hurting a fish's lip to save many other fish from pain justified?
(There is nothing mild about this question. Yet there are other human-caused hurts and harms inflicted on animals, having nothing to do with saving lives, that are far more troubling.)
I agree with the class’s overall dissatisfaction with Key’s argument, and found a counter example similar to Sneddon’s in one of the commentaries. Striedter’s (2016) response, “Lack of neocortex does not imply fish cannot feel pain,” counters Key’s points on reflexive behavior and lack of behavioral change after ablation. He says “goldfish learn to avoid food that made them nauseous,” and after telencephalon lesions, this conditioned taste aversion is impaired. I thought this was really compelling because Striedter extends the argument (against behavioral changes not following ablation in fish) to imply that learning even occurs to avoid future pain, and some experience of pain had to be necessary for this behavioral change. These behaviors can't solely be reflexive and can even be adaptive.
DeleteHe reiterates many of the comments from the class on cherry picking data too which I found interesting: “I don’t know how these scientists can look at essentially the same data and reach such disparate conclusions, but I suspect that some of them have strong a priori beliefs and, therefore, view the data through differently tinted spectacles.”
The evidence that Key provides in his article is correlational. He then dismisses correlational data that works against his point as ‘erroneous inferences’. I also take issue, as do many of the commentators, with the idea of forcing mammalian anatomy and neurobiology onto fish. Why should we assume that fish don’t feel pain because their structure are not the same as a human? Conceptually, it’s possible they can feel pain using different mechanisms.
ReplyDeleteThe icing on the cake, the point I wish to lambast, is found in the conclusion of Key’s article. Key brings up behavioral changes which are observable in humans after portions of their skulls are removed for brain surgery. He then mentions how damage to the skull in a similar operation done to fish does not lead to comparable behavioral changes. Accordingly, Key, who from the beginning speaks out against the use of behavioral changes as a reason to know if an animal feels pain, uses behavioral data to suggest that fish don’t feel pain. In fact, no matter how many neuroanatomical changes that Key notes from lesion studies or rat experiments, he still relies on the observation of behavior to draw his conclusions.
Hey Matt, I kind of agree that Key’s reading confused me a bit too. My first reaction to Key’s paper is that all of the evidence he brings up concerning the neural structures in the brain is just correlational in the end! Also, it feels like he kind of contradicts himself when he says that people shouldn’t give fish “the benefit of the doubt” because those individuals are just imposing an anthropomorphic perspective onto the fish, but then he goes on to say that since humans have a certain neural structure and fish don’t, they don’t feel, and humans do.
DeleteThis is a point that some of the other commentaries also bring up, like the ones written by Manzotti (2016) and Braithwaite & Droege (2016). From their abstracts, it appears they also point out it isn’t sufficient to conclude that only human neural structures can process conscious pain.
Matt, yes, Key's target article is a striking example of cherry-picking evidence and conclusions to save a theory. It happens in a lot of science, but it's especially bad if it is at the expense of other minds suffering.
DeleteEsther, a lot of science is causal inferences based on correlational data, but here other minds are at stake. (See the discussion of the Precautionary Principle in other Replies.)
Setting aside the tenuous or even meaningless difference between nociception and truly ‘feeling’ the affect of pain that others have brought up, there is one aspects to Key’s logic that I very much disagree with in his paper. The author provides ample evidence to posit that there is a robust correlation between the way that our neural circuitry is configured in the cortex and the affect of pain, so strong as to suggest that it might be sufficient to experience this (though only the abstract of causality rather than being linked by some sort of physical system, as is prescribed in this class). Where Key errs is that he also implies that it is necessary, saying that fish neural anatomy is too dissimilar from our own and vertebrate’s in general such that it wouldn’t be computationally possible. I allude to Fodor in that there’s nothing in the location of parts of the brain (and by extension, the circuits themselves and how they connect), that necessarily explains how a certain cognitive operation transpires.
ReplyDeleteTo bolster his opinion, Key appeals to evolutionary theory to say that rats and humans are homologues in that they both developed a cortex and as such, have the capacity to experience pain. What Key doesn’t really acknowledge in evolutionary theory is the idea of analogues, that beings can develop the same type of function through independent evolutionary means. Just as because the wings of a fly, a bat, and a bird are vastly different doesn’t mean that a bird is the only one with true wings, just because the fish and the vertebrate have vastly different brains doesn’t mean that vertebrates are the only ones to experience pain.
William, "because the wings of a fly, a bat, and a bird are vastly different doesn’t mean that a bird is the only one with true wings"
DeleteNor, even more to the point: that only birds can fly!
I want to connect this paper to the cogito and to some other themes from the course. The cogito was Descartes' seminal work which dealt with what we can and cannot be certain about. This was at the height of philosophical skepticism in which philosophers were questioning whether there was anything we could be certain of. Descartes said that the only thing we cannot doubt is that we feel; it doesn't matter whether the object of the feeling is real or imagined, we cannot doubt that it exists.
ReplyDeleteOkay, so I know that I feel beyond the shadow of a doubt but what about other humans? I am fairly certain that they also feel but because of the Other Minds' Problem I can't know for sure. As a society we are generally less certain that animals have minds and this increased level of uncertainty gives rises to papers such as the one by Key (2016). Key uses comparative neurobiology to argue that because fish do not share the same neural cytoarchitecture that are correlated with pain in humans and mammals then they do not experience pain. Based on this paper, I feel like Key would be inclined to kick T3 and would draw the line at T4 since to pass T3 the successful candidate need not have the same neural architecture as a human. Would he say that a T3 could not feel? I understand that the Turing Test is different since it’s focused on human doing capacities and not fish-doing capacities, but the question remains: is a difference in structure all that it takes for us to be convinced that things don’t feel?
Good points. And there are also differences in behavior, such as Sneddon's analgesic experiments.
DeleteBut analgesics also have effects on plants: What are we to make of those?
I think Key is not only making a large leap by separating nociception from pain perception by arguing that it just a reflex, he is also arguing that assuming that fish feel pain is just as damaging or more so then assuming that they don’t because we would no longer be able to morally justify our massive exploitation of them. In this case, he is arguing we kick Ting so long as she doesn’t explicitly tell us that she is feeling (and doesn’t look like us internally). He argues that humans “intuitively understand that humans share similar functioning nervous systems and behaviours and therefore will experience similar feelings”, seemingly discarding the other minds problem as pure “common sense”. However, this is in no way proof that other humans actually do feel similar feelings, or any feelings at all, and does not even approach a causal explanation of how and why we feel.
ReplyDeleteAylish, We can't ask for a solution to the HP from Key. The question is whether he has done a good job on the OMP.
Delete(And see also the discussion of Pascal's Wager and the Precautionary Principle in other Replies to Week #11.)
I want to approach this in two ways:
ReplyDelete1. I don’t know how content I am with the neuronal argument. Something about it doesn’t sit right with me. It feels overdeterministic of us to only attribute pain to others if we find structural similarities that we deem close enough to ours. There may be multiple ways for an organism to feel pain.
2. Even if we had confirmed evidence that fish don’t feel pain, there is still some pain in the equation. In class, prof Harnad gave us a hypothetical example: let’s say Ting is 100% a robot that cannot feel pain, would we kick her? I don’t think most of us would, because it feels bad to us. Not only that, I’m sure it would upset Ting’s friends and family if we kicked her. In that same sense, there will still be pain arising from killing fish, even if we confirm that it’s not coming from the fish, there are still other sources of pain. Alessia’s skywriting pointed that out very clearly, seeing a hook go through a fish looks very painful, it can manifest pain in others, even if you could confirm that the fish wasn’t feeling anything.
Now for my personal opinion, do I think we should concern ourselves with research that tortures fish simply to figure out whether they feel pain? No, let’s just assume they do. Giving fish the benefit of the doubt isn’t causing anyone pain, although it might be reducing some people’s pleasure. My immediate thought when I think of pain and pleasure is necessity and luxury. Is it necessary to determine that fish don’t feel pain? No. It feels like research in this field is looking to justify a luxury, and honestly, there are so many more important things to put your time and money into than looking for a way to justify your luxury to your conscience.
Hey Lyla, I would like to expand on the point you made in your comment above. I am not satisfied either with the "neuronal argument". Key seems to be ignoring one key notion that we happen to have discussed in this class: underdeterminism. There is a neurobiological model of how pain might be felt and it happens to be a model informed by mammalian neurobiology. The how of this model is vague and I don't take it to be the same kind of explanation that would be required to answer the how of the HP. Nonetheless, Key uses it to argue that organisms who do not have the same feeling machinery as we do cannot experience what we experience. A recurring example of wildly different structures being able to carry out similar functional roles to other structures is cephalopods. They literally have a decentralized nervous system yet manage to behave in surprisingly intelligent ways! The problem with Key's argument is that he forgets that having one model of how pain might be felt does not imply that it is the only model.
DeleteLyla, good points. Yes, for most of us (except sadists and psychopaths) it feels bad to damage a living organism even if it is really just a feelingless zombie (like — I hope — a tree or plant).
DeleteIn my case that even spilled over into objects. When I was about 9 or 10 my mother had a drawer in her office where she let me keep my used candy-wrappers and bus-transfers. Why? Because I felt sorry for them. I felt it was wrong to treat them as if they were just objects , and just throw them away.
But they were just objects!
A few years later I realized where this overgeneralization was coming from: because it felt so wrong to treat feeling animals as if they were just objects. So first I told my (relieved) mother that it was alright to empty the drawer.
And then a few more years and I stopped eating animals (including fish).
Evolution is lazy; it didn’t draw a rigid line between our kin and other sentient beings.
And I’m a slow learner. (The last token didn’t drop until ten years ago, when, in 15 seconds at the beginning of Wolfson’s talk about “dairy” — and that is the greatest shame of my life.)
Solim, good point about underdetermination. Biologists call it "convergent evolution” and "parallel evolution" and in technology and philosophy it is called “multiple realizability.” It is closely related to the notion of “functionalism” in philosophy of mind, as well as to the notion of underdetermination in the philosophy of science. That, in turn, is related to “algorithmic complexity theory” and to degrees of freedom in statistics
Key argues that fish do not feel pain because they lack the requisite neural structures. However, a problem with this seems to be that multiple structures/organizations of structures could perform the same function. In the above thread, Prof Harnad refers to the idea of "multiple realizability," and this seems to be the key (no pun intended) -- multiple structures could give rise to pain, even if the putative pain structures that Key has identified are not found in fish. Moreover, pain likely had an evolutionary function and the capacity to feel pain therefore evolved in multiple organisms. It is likely that pain evolved in fish to serve the same function that pain serves in other organisms, but the structures that fish evolved to accomplish this function were unique or distinct. This hypothesis seems far and away more plausible than to simply say that fish cannot feel pain. Evolution is lazy -- it gives rise to whatever functional trait is most helpful in a given environment over generations; evolution need not instantiate this functional trait the same way (with the same structure) in every organism! If Key is to increase his credibility, he should use better arguments and also lose the sketchy wizard look.
ReplyDeleteAll true -- except you forgot to say what function pain -- the feeling of pain, not just the doing of what organisms do when they are injured -- serves. (That's called the hard problem -- although it is not at all relevant here, because this is all about the OMP, not the HP.)
DeleteIn his article, Key claims that “we vicariously experience the pain of our friends and family members without ever really knowing what they actually feel. We most likely do this because we intuitively understand that humans share similar functioning nervous systems”. However, I believe this is an overstatement. Even little children pick up on pain experienced by those around them but this is hardly due to “an intuitive understanding that they have a similar functioning nervous system”. As others have mentioned, observing the behaviour of other humans/animals plays a strong role in our inferences that others feel pain. Key might argue that this reliance on behaviour is simply a manifestation of the intuitions he mentioned above. However, he also makes strong claims against behaviour as a basis for inferences about whether others feel pain or not. “The possibility that fish lack the necessary neural hardware to feel pain is typically overlooked because of anthropomorphic tendencies that bias interpretations of behavioural observations”. I think the intuitive understanding he alludes to is something more along the lines of : “that human/animal flinched which is something I do when I am scared or in pain. That human/animal is most likely in pain as well”. I don’t see how his claims about different neural structures are relevant. As other students have already mentioned numerous times, I found his argument unsatisfactory and it served only to reinforce my belief that fish do in fact, feel pain.
ReplyDeleteSide note : I second Alex's thought ^^ that Key should lose the sketchy wizard look. He definitely looks like the type of person who would believe fish don't feel pain.
Good points. Why do you think lazy evolution gave us our OMP mind-reading capacity?
Delete“I don’t see how his claims about different neural structures are relevant.”
DeleteWhile I obviously disagree with the ways that assumptions about who does and doesn’t feel pain are used to justify self-serving conclusions, I still think that it is a question that deserves rigorous attention, even if “objective” answers are impossible due to the other minds problem. For instance, you point out that children intuitively interpret certain behaviours as indicating pain, but it is also important to remember that children also often anthropomorphize many non-human and also even non- living things. When I was a kid, I was very concerned about hurting my stuffed animals’ feelings. So I don’t think that an appeal to the power of our mind-reading abilities from childhood is necessarily the strongest argument.
I agree that our tendency to interpret others’ behaviour as involving feelings is much more basic and instinctual that thinking we “share similar functioning nervous systems”, considering this question scientifically, to the extent that this is possible, has everything to do with neural structures. It is very possible that biological differences in other species fundamentally alter the way that they experience the world, including pain.
All of this being said, I think it is important not to abstract this discussion from the real consequences that it can have. When discussing whether other beings feel pain, it is essential to err on the side of caution – the consequences on believing that they do not feel pain when they do is so much greater than being mistaken in the other direction.
11a. Key, Brian (2016) Why fish do not feel pain.
ReplyDeleteKey’s article suggest that fish don’t feel pain because they lack the appropriate neural structure in their brains. The author suggests that the requisite structures for feeling pain in fish would have to be the same as the requisite structures for feeling pain in humans. As professor Harnad mentions, there is a possibility that feeling pain is multiply realizable via different neural circuitry. We know that the brain must produce feeling, but the functional explanation for how it does so is yet unknown. Thus, it seems much too soon to say that feeling pain is caused by specific configurations of the brain. Yes, certain regions may be activated during the reported experience of feeling pain in humans, but the activated regions only correlate with the reported feelings and no functional causal mechanism is given. To assert that fish don’t feel pain, we would have to give a mechanism for feeling and prove that this mechanism is not multiply realizable in the brain of fishes. The multiple realization hypotheses is plausible because of Lazy Evolution which wouldn’t necessarily “come up” with the same exact mechanism for feeling. Furthermore, feeling pain must have had an evolutionary advantage and thus it is plausible that it would have evolved in numerous organisms.
By the way, I think that the comparison between the study about vaccination which made people believe that vaccination caused autism and the studies that gave the benefice of the doubt that fish were feeling pain is an extremely gross comparison.
I really like your connection to lazy evolution here. I think that what we are also witnessing here is that we should not be judging the consciousness of other animals considering we struggle with the Other Minds problem. If the only thing we can be certain of is our own feelings, thanks to Cogito, why then do we have the authority to claim that other animals do not feel? We believe that we can mind read other humans using mirror neurons, which allows us to have good confidence in the feelings of others, but this mechanism has most definitely adapted so they apply best to other humans. I remember reading that the eyes and eyebrows of dogs have evolved to mimic human eyes, so that they seem to express similar emotion. This sounds like a way for dogs to survive given that humans clearly wrong animals that they cannot identify with, like fish. Fish have not evolved any capacity to share or mimic emotions with humans, why would they? But what if fish had eyes and eyebrows like dogs (I know that would be scary)? Then would we be so inclined to kill them?
DeleteMatt your comment made me think about this. Do we really struggle with accepting sentience and feeling in other beings because they are too unlike us? This doesn't seem to be the whole story. The OMP is a first example of that. We struggle with confirming feeling and sentience in other humans, who look (almost) exactly like us! To take Prof H himself as a second example (and me, because I struggled with the same thing as a child), we sometimes attribute feeling to objects, even though we can be very sure they do not feel and although objects can look nothing alike us. However at the end of the day we still do treat objects, even objects we attribute feelings to, kind of like we treat animals we kill and eat - like they exist to serve us, and can be discarded once they outlive their usefulness or importance in our eyes.
DeleteMaximilien,”feeling pain must have had an evolutionary advantage”. Do you have any hunches as to what that might have been? For they could be the solution to the hard problem. Be careful that it's not something that would work just as well for insentient nociception...
DeleteMatt, I'm struck by how similar not only animate movement, but the bodily expression of felt states is across most vertebrate and even some invertebrate species.
Eli, It is easy to cultivate psychopathy in our children: We need only tell them the lie that eating meat is necessary for their survival and health, that since animals also do it without remorse, it is the Law of Nature, and that in any case animals are raised and slaughtered in a "humane" way (you just have to avoid viewing Google images on slaughterhouses). In fact, by exactly the same cultural means we could (again) instil in our young the taste and the justification for rape, torture, slavery, and genocide.
Key writes out an article that while it is well researched and stays in line with the logic that he proposes (Fish cannot feel pain because they lack the neural structure to do so), he fails to really draw a strong conclusion with the evidence he brings up. I believe his biggest error was the fact he would dismiss the behavioural similarities that fish have with mammals when it comes to pain stimuli being applied to them, while not providing an adequate definition of what pain behaviour really is.
ReplyDeleteHe would essentially state that because they are not reacting in the same way that mammals do (while still also showing that it's possible they may be in pain, but the evidence is pointing in the other direction), we cannot infer that they are feeling pain. If he gave a clear definition about what pain behaviour really was (or an even harder task, what is pain) then it would be easier to accept his argument, but this is not the case. The fact that fish have their distinct compensation mechanisms to pain stimuli is enough to suggest that they feel pain, let alone us actually engaging in our OMP mind-reading abilities about what they feel.
Moreover, when applying it to this class, we can see that (as Solim mentioned) Key was seriously ignoring underdetermination, and misunderstanding the OMP. Key attempts to point that the mammalian neurobiological circuitry of pain is the best/only model to properly experience pain, yet, we know this to be false. Not only do we only have correlational evidence that is supporting our ideas (which means we cannot infer causation), our current knowledge of pain circuitry is highly inconclusive, and therefore it is a working model, and not a definite one. Furthermore, this does not also rule out the possibility that we could have a pain model for fish pain-circuitry in the future. In regards to the OMP, his misrepresentation about anthropomorphism not being equal to pain does not hold weight because even if we do not know if a being is feeling pain for certain, my inference that it does feel pain, is not necessarily wrong. The OMP is an issue about inferring for others because I can only feel what I feel. This inference does not rule out the possibility that fish feel pain because they lack the mammalian structure of pain circuitry.
I agree with all the comments describing the problems with Key's article which stem from the assumption that structure determines function. I think another primary piece of "evidence" presented by Key in support of his hypothesis that fish cannot feel pain hinged on observable behavioural changes. Key argues that studies should focus on searching for suppression of identified "normal" behaviours - rather than emergence of abnormal ones - as evidence that an organism is experiencing "pain." He writes in his conclusion, "the lack of suppression...is consistent with these fish not feeling pain after craniotomy."
ReplyDeleteThis statement is placed in the context of fish who normally perform what in humans we would call "care" work - reproductive labor, feeding/nesting/courtship etc. - and who were not observably changed in their performance of this labor after a craniotomy. In contrast, humans do behave differently after a craniotomy (apparently as a result of the reported experience of "pain"). Key's logic along with other scholars, is that because fish do not react behaviourally to an "equivalent" procedure that inflicts pain and related behavioural reactions observable in humans, therefore there is no experience of pain. I think there are so many problematic assumptions in here, but mostly that an organism should behave a certain way and look a certain way to prove pain.
Mothers of human babies can perform tremendous feats of physical strength, that should be impossible and at least incredibly painful, because of the adrenaline that pumps through their bodies when a child is in danger. People can receive injury and not feel the pain until later as a result of adrenaline. Some pain actually feels like growth (physical or emotional). Different organisms in different situations react differently to infliction of harm or damage - it does not always look the same. Fish providing reproductive labor might not have the same structure constituting their bodies as humans, nor the same reactions to similar "damaging" events, but it does not follow that they do not feel or that they do not feel some form of pain. If a person is hit and does not bruise, can we assume they do not feel pain there? No.
This was kind of rambling but if we want to use human pain experience and nociception as the sole barometer for identifying and understanding pain in all life, the hypotheses need to be more nuanced. Human pain is not homogeneously experienced nor evidenced, why should we expect that across species?
I thought it was a stretch to argue that “serious negative implications” could arise by accepting that fish feel pain, Key specifies particularly on “native subsistence fishing, human nutrition and food supply, and economic development.” I really don’t think acknowledging that fish feel pain would be any more detrimental than the commercial fishing industry, and I don’t think that the items he lists are mutually exclusive with empathizing with fish.
ReplyDeleteI definitely agree with you. I think his point is especially absurd in light of the fact that a lot of people accept that other animals we slaughter are sentient and that hasn't even negatively impacted the animal product industry let alone human nutrition and economic development.
DeleteAlso, as has been mentioned above by everyone else, he doesn't really prove that fish don't feel pain in the paper, only that fish don't feel pain with the same mechanism we do and reading the paper lead me to the impression that he's the kind of person that would not only kick Ting the robot but would also excuse others hurting her because she's not a
T4 robot, which leaves me with the impression that he's not necessarily someone I'd like to meet. (Although to be fair, his point is logical in itself. If all beings felt pain exactly through the mechanism humans do, or expressed/reacted to pain like we do, he'd have a point. But not only do we not know that for sure, we have no reason to assume that that is the case.)
Key’s argument feels almost negligent in its assumption that only one possible structure could result in feeling pain, especially when the hard problem of ‘feeling’ is something that we are far from understanding. Not only that but it disregards the gut understanding that is experienced by people on a day to day basis who see and understand animals in pain, including fish and the like (you do not have to travel far on the internet to find worried fish owners posting about the concerning behaviours of their aquarium inhabitants).
ReplyDeleteHaving read Key's article, I think the major flaw in his work is the level of determination. In other words, I think that saying that because fish don't have the structures that become active in correlation with pain in humans, they cannot feel pain, is the result of wrongful overdetermination on Key's part. Even if we assume that the correlation between those structures' activity and pain in humans is in fact causal - that our feeling of pain relies on and is conducted by those structures - this still does not render such structures the only possible way for pain to be felt.
ReplyDeleteIn making this critique, I am drawing in part on class discussions about the Turing test, and which levels are appropriate. If you asked Key, it seems like he would be pushing for a T4 robot (as many neuroscientists and biologists might); yet, as we discussed, insisting on T4 means severely limiting what you consider consciousness, and what minds you deem as eligible to be thinking, feeling, and sentient.
Further, if you take Key's argument to its logical conclusion (which is far further than he takes it), it would be easy to consider many species non-sentient, just because their brains are not constructed precisely the same as ours. In this vein, asking whether a fish feels based on the physical criteria of human feeling appears to be a slippery slope, which would exclude many animals from sentience. This concern is also specified by the context of Key's argument: I would readily agree that given the evidence he presents it is not possible to say definitively that fish do feel pain, yet I would not agree that this evidence is sufficient to justify a definitive belief that they do not feel pain, or, more specifically that noxious stimuli (being hurt) does not "feel like anything to a fish".