When it comes to animals, it seems what people are most interested in is if animals are “conscious.” However, I’ve criticized the word ‘consciousness’ as vague and pointing to many different semi-related concepts. So instead of addressing the question “Are Animals Conscious?” I’m going to instead address the question “Do Animals Experience Qualia?”
To feel qualia isn’t necessarily the same thing as to be conscious. People we consider unconscious or semi-conscious often feel qualia. (e.g. Ambien zombies and people in REM sleep.) And people that we consider fully conscious sometimes lack certain kinds of qualia. (e.g. people that feel no pain or have pain asymbolia) (See this post here for details.) Yet when people ask if animals are conscious or not, I think a good deal of the time they really just want to know if animals have an inner life (i.e. aren’t automatons) and if they can feel things like pain, suffering, jealousy, grief, etc.
In this post, we are going to address this question. Using Critical Rationalism as our guide, we can even tentatively answer the question.
Does the Existence of Universal Explainers Mean Animals are Not Conscious?
Many fans of David Deutsch believe that Deutsch’s concept of “Universal Explainers” implies animals are not “conscious” which to some of them implies being automatons (i.e. experiencing no qualia.) This is common enough of a belief that I feel the need to address their arguments.
Dennis Hackethal made such an argument in his interview on the Do Explain podcast. I covered criticisms of some of his arguments in this post. But Dennis’ argument is fairly straightforward. He notes that consciousness requires error correction and so he conjectures that the two go together. In my previous post, I noted that he has the arrow of association backwards. Error correction is required for all learning, conscious or unconscious. Dennis’ own example of riding a bike demonstrates how you can constantly error correct to keep your balance without any conscious effort. This is how all skills and inexplicit knowledge work, so this argument for why animals can’t be conscious does not work.
Another argument I’ve heard is that it would make sense if three mysteries –- consciousness, qualia, and the universal explainer algorithm — all turned out to be a single mystery. And I admit that that would be rather elegant, if true. But an argument from elegance like this is the basis for an early speculative conjecture only, not a good hard-to-vary explanation.
So there is no reason – that today we know of anyhow – why the existence of a single universal explainer algorithm would need to imply that animals do not feel qualia or are automatons. 
Can Apparent Animal Qualia Be Explained as Anthropomorphizing?
David Deutsch has, in the past, claimed that belief in animal qualia is a cultural and even supernatural belief.  But I suspect the real main reason why people believe that (at least higher) animals feel qualia is because such animals act like they feel qualia. We all have seen pet dogs in pain, for example, and it looks just like a human being in pain.
I can see why this might cause skeptics to feel that the main basis for belief in animal qualia is therefore easily explained as anthropomorphizing. We are used to crediting certain kinds of behavior (wincing, screaming in pain, etc) as being caused by pain because that’s how we know we react to pain. So it would be natural for us to ascribe a dog wincing and yelping due to pain as the dog feeling pain as well. What makes this even harder is the fact that some such behaviors we associated with pain – specifically flinching – exists in all animals, even ones that have no nervous systems. Single-celled animals flinch when touched. (See “A Theoretical Computer Science Perspective on Consciousness,” Manuel and Lenore Blum, p. 22)
So can we simply explain all animal behavior solely as a case of anthropomorphizing?
No, we can’t.
To make it clear why a blanket claim of anthropomorphizing is actually a bad explanation watch this video here of this experiment done with monkeys. This is more powerful if you actually watch it for yourself, but I’ll describe what the experiment is below.
The experiment was to give to monkey’s a cucumber or grape in exchange for a rock. If both monkeys got cucumbers, they were happy to complete the task in exchange for a cucumber as a reward. But if one of the monkeys got a grape for the task then the second monkey got jealous and angry when he only received a cucumber for the same task. This was true even when (as we see in the video) the monkey had at first accepted the cucumber. It was only upon seeing the other monkey receive a grape that the first monkey became jealous and upset.
Can you explain what you are seeing in this video by calling it “anthropomorphizing?” No, you can’t. That’s really not explaining the behavior of the monkey at all. We still have an observation in need of an explanation and trying to root the explanation in the observer will do no good.
So how do you explain the monkey’s behavior? Do you find that you are like me and you do not feel you can offer a good explanation without using words like ‘jealous’, ‘angry’, and ‘upset’ that imply that the monkeys have an inner life at least somewhat analogous to ours? Can you offer an good alternative explanation that doesn’t reference those feelings that isn’t just an ad hoc explanation? (Popper wisely said that we must not save theories using ad hoc explanations because they can be had for the asking.)
Or in other words, can you offer an alternative explanation that is not easy-to-vary? I find I can’t. So that is why I have no qualms with explaining the monkey’s behavior referencing qualia such as jealousy and anger. For it is currently the only non-ad hoc explanation available to me that allows me to explain the behavior.
So this is the first surprise when it comes to determining if animals feel qualia – that we need to conjecture the existence of animal qualia to explain many kinds of animal behavior and that without doing so the behavior has to remain unexplained.
So let’s make explicit the tacit theory I’m apparently being forced into to explain the monkey’s behavior:
Proposed Theory of Animal Qualia 1: The reason (at least higher) animals act like they experience qualia is because they do experience qualia.
What’s surprising is just how hard-to-vary this explanation actually is.
Let’s take a simple example of the hard-to-varyness of this explanation. How do you explain why dogs act like they are in pain – in a way that humans can easily recognize – if dogs do not in fact feel pain? If you are taking the above theory as your view, then you can now tie it to one of our deepest explanations, biological evolutionary theory. Animals developed pain before humans (and thus before the advent of universal explainers) ever came on the scene. Since our brains contain much of the same machinery as the animals we are related to, then it’s not really that surprising that our brains contain the very same machinery as a dog’s brain by which pain is experienced. That is why dogs acting like they are in pain is so recognizable to us.
Now imagine you are trying to take the alternative view that dogs do not feel pain and are really automatons. Here is the proposed competing theory:
Theory of Animal Qualia 2: Animals are automatons that feel no qualia at all. All instances of behavior that seem to be experiencing qualia are just signaling behavior with no qualia being experienced.
How does this second theory explain that dogs react towards pain in ways humans recognize so easily? This theory nearly forces you into an ad hoc explanation that maybe evolution, just by coincidence, happened to make a dog’s signal for damage (as well as basically all higher animals!) the same as the one that humans happen to use when they actually do feel pain. If the human reactions to pain evolved only since humans became universal explainers then there is no reason our signals should happen to be just like dogs’. So this second theory leaves a significant explanation gap that our first theory does not have: how do you explain the coincidence?
Dog’s and Pain Killers
In fact, we can now take the two competing theories and easily test them. Consider how humans react to pain when given painkillers. There are two kinds of painkillers, of course: local and general. Local anesthetics wouldn’t test our two proposed theories well because you deaden the signal to the brain. In that case, both theories predict that neither humans nor automatons would detect that they are damaged and thereby signal pain responses.
However, the story is entirely different with general painkillers. If you give humans opioids in just the right amount they can still feel some pain, so they know when they are being damaged, but they stop suffering. If dogs were actually automatons then suffering wouldn’t be possible for them. So giving them opioids in the right amount would not stop their brains from registering as needing to signal they are damaged. So here then is the proposed experiment to differentiate between the two proposed theories:
Proposed Experiment: Give dogs opioids in small amounts. If they continue to act as if they are suffering, then they are automatons. If they stop acting like they are suffering, then they feel qualia.
Now I’m sure most of you know that this experiment has been done probably millions of times. We can and do give opioids to dogs and they will stop suffering. This is right in line with Theory 1, but the opposite of what Theory 2 predicted. So you will again be forced to use an ad hoc explanation to save Theory 2. You’ll need to claim that opioids simply behave differently on dogs than humans. In humans it causes suffering to end while in dogs is disables the pain signal. But this ad hoc explanation now introduces a new significant explanation gap: why does the drug affect humans and dogs differently in a way that just happens to coincidentally make it look like dogs do feel suffering? At a minimum, this experiment means that Theory 1 survived a test that could have in principle falsified it and that Theory 2 is now even easier-to-vary since, to save the theory. you have to claim this isn’t a good test in the first place.
The Theory of Grief
As it turns out, nature has supplied us with a series of incredibly good test cases via our theory of grief. The idea is this: is grief an adaptive behavior?
If biological evolution is a correct theory, then the answer has to be ‘yes’. Otherwise humans would not feel grief in the first place because evolution never would have endowed us with such an emotion. But how can that be? Grief certainly seems to be a non-adaptive feeling that reduces survival value, not increases it. A human being that is feeling grief may become depressed and refuse to eat, sleep, or take care of herself. A person that is feeling grief can literally die due to either lack of selfcare or through suicide. Grief itself can’t be adaptive as a standalone emotion. So why does it exist at all?
Our best theory is that grief is indirectly adaptive. Nicholas Christakis, in his book Blueprint: The Evolutionary Origins of A Good Society explains:
There is much evidence that grief is physiologically harmful and can even increase one’s subsequent risk of death, so ancestors of ours who did not feel grief should have out survived those who did [all else being equal.] (p. 298)
[Our best theory] is that grief is a by-product of a psychological system that evolved to make humans feel bad when separated from their living kin, since staying together was presumably adaptive. In this view, grief relates to social cohesion, and it is the price we pay for social intimacy. (Christakis, p. 299)
Christakis also explores related theories such as the theory that fear of grief was evolution’s way of getting kin (particularly parents) to protect each other for fear of how they will feel if they lose family members. (p. 299)
Both of these theories are closely related. What they have in common is that they are explanations of grief that conjecture grief to be explicitly a form of qualia. So if animals are automatons then animals won’t feel grief and should not act in non-adaptive ways when ‘signaling grief’ like refusing to eat or take care of themselves. But if animals do ‘feel grief,’ then our best theory of grief will now require us to conjecture that grief-filled animals have qualia.
So do animals feel grief?
In fact, scientists have recorded many examples of animal grief and even done a series of studies to test alternative explanations. Christakis (who believes animals feel emotions) documents a number of these examples and follow up studies:
Christakis gives an example from researcher Jane Goodall about a chimp that literally died of grief over the death of its mother:
[A chimp named Flint] suddenly left the group and raced back to the place where Flo [his mother] had died and there sank into ever deeper depression. By the time Fifi showed up, Flint was already sick, and though she groomed him and waited for him to travel with her, he lacked both the strength and the will to follow. Flint became increasingly lethargic, refused most food and, with his immune system thus weakened, fell sick. The last time I saw him alive, he was hollowed-eyed, gaunt and utterly depressed, huddled in the vegetation close to where Flo had died… The last short journey he made, pausing to rest every few feet, was to the very place where Flo’s body had lain. There he stayed for several hours…He struggled on a little further, than curled up – and never moved again. (Christakis, p. 300-301)
So here we have a documented example of an animal feeling so much grief that it died. This is clearly non-adaptive behavior so it can’t be simple automaton signaling. When combined with our best theory of grief, we are forced to accept that this chimp was experiencing the emotions of grief — right in line with Theory 1 but at odds with Theory 2. To save Theory 2 we’re forced to now claim that the death of this animal was unrelated to grief and just a coincidence.
But coincidences can be tested!
Another group of researchers came up with a clever way to test the hypothesis that actions like the above really were signs of grief. They had the idea that if primates were experiencing deep emotions, similar to humans, that there would also be signs of hormonal changes and those hormonal changes would reduce if the primates received social support as a way to reduce their grief — just like in humans.
Goodall… was not alone in witnessing and recording deep emotional and physical reactions to loss in primates. One study, using fecal samples collected from a troop of more than eight baboons in Boswana, found that glucocorticoid levels (a measure of stress in both baboons and humans) were higher in female baboons most closely related to a deceased animal. Even more interesting, these females appeared to reduce their stress by increasing the frequency of grooming and the number of grooming partners. … The grieving female baboons who expanded their grooming networks had substantially reduced glucocorticoid levels compared to grieving females who did not. (Christakis, p. 301)
So some animals, when ‘signaling grief’ not only have the same hormonal response as humans but resolve it just like a human would by seeking companionship. This is consistent with Theory 1 but not Theory 2.
Elephants and Grief
Elephants are another animal that have been recorded by scientists as dying from grief.
[Elephant researcher] Joyce Poole became convinced that elephants felt grief when she saw… an elephant mother [watching] over [the body of] her dead newborn: “every part of her spelled grief” she said. …One matriarch “stopped eating and starved herself to death after her protégé died in childbirth.” (Christakis, p. 302)
The fact that elephants feel grief may explain a long-standing puzzle that many researchers have recorded about elephants: Why they have rituals for their dead?
Roger Penrose, after watching an episode from David Attenborough’s television show Echo of the Elephants (1993), noted the oddity of elephant behavior towards their dead, expressing his opinion that they must feel grief and have feelings similar to humans:
The leader of a heard – a female, whose sister had died some five years earlier – took the heard on a long detour to the place of her sister’s death, and when they came upon her bones, the leader picked up her skull with great tenderness, whence the elephants passed it from one to another, caressing it with their trunks. (Penrose, Shadows of the Mind, p. 407.)
Attenborough is not the only person to have noticed this sort of behavior in elephants. Zoologist Cynthia Moss reports a herd of elephants visiting the bones of its deceased matriarch, Emily. (Elephants are matriarchal, so the former leader of the group.):
They stepped closer and very gently began to touch the remains with the tip of their trunks, first light taps, smelling and feeling, then strokes around and along the lager bones. Eudora and Elspeth, Emily’s daughter and granddaughter, pushed through and began to examine the bones…All elephants were quiet now and there was a palpable tension among them. Eudora concentrated on Emily’s skull caressing the smooth cranium and slipping her trunk into the hollows in the skull. (Christakis, p. 302)
Joyce Poole describing another incident said:
The family approached [Jezebel’s] remains and then suddenly stopped and became silent…and then spent the next hour turning the skull, the jaw, and the long bones over and over. The elephants, who appeared to be in a sort of trance, neither interacted nor vocalized and seemed to focus only on the dead elephant. Jolene, Jezebel’s daughter, appeared to be the most absorbed of the group…Why would an elephant stand in silence, over the bones of its remains for an hour if it were not having some thoughts, conscious thoughts, and perhaps memories. (Christakis, p. 302)
Whales and Chimps have also been seen displaying similar behavior. 
Theory 1 neatly explains such complex animal behavior the same way it does for humans. But how will Theory 2 explain an elephant matriarch taking a long detour just to visit the bones of a deceased love one?
The theory of elephant grief is also backed up by studies that show that:
Elephants may even have a kind of collective mourning, like humans do. Large groups of elephants have shown symptoms associated with grief, including abnormal startle response, depression, and extreme aggression. (Christakis, p. 303)
Dogs and Anti-Depressants
But we don’t need to go into the wild to study animal grief. It is well documented that dogs experience grief over the death of an owner or companion dog. The movie Where the Red Fern Grows wasn’t taking artistic license when it comes to depicting dog’s experiencing grief. It actually can be dangerous to their lives. That is to say, dogs can die of a broken heart not unlike humans.
Grief for dogs is dangerous enough that vets will prescribe anti-depressants to dogs to help them over their grief. It’s interesting to note that they prescribe the very same drugs to dogs that they do for humans because the drugs operate the same on a dog as they do on a human. Why would that be the case? Theory 1 easily explains this. Theory 2 can’t.
Animals form attachments to each other, not unlike how humans do. But do animals have feelings of love and friendship like humans do? The existence of animal grief suggests that they do have feelings for each other of some sort. Another observation that requires us to conjecture that animals feel something similar to friendship is this study that was done with monogamous voles:
…voles who are experimentally isolated from their [monogamous] partners display hormonal and behavioral changes consistent with depression. In another experiment, scientists suspended voles from their tails and noted their reactions. If a vole was pair-bonded with a mate, it struggled. If it was not pair-bonded, it also struggled. But a pair-bonded vole whose partner had been taken away acted as if it were depressed… and it did not struggle. (Christakis, p. 169)
There are many documented cases of animals seeming to form attachments of love and friendship with other animals or with humans. Obviously dogs are the most famous for this and explains their popularity as pets. Given the arguments above, we should likely embrace Theory 1 and accept that these displays of affections are likely due to the animals actually experiencing emotions similar to humans.
Dogs are famously far more excited to see people they are close to that they haven’t seen in a while. Elephants have a complex greeting ceremony that increases in intensity the longer their friend has been away. Elephants can even become attached to humans and will perform this ceremony for them. 
Why do humans feel feelings of friendship and why is that adaptive? Scientists have conjectured that the same reasons why humans benefit having such feelings is why chimps do as well:
Chimpanzees that groom each other today are much more likely to form a protective alliance in a few weeks and be sharing food in a few years. … Primatologists theorize that emotions may serve as a kind of cognitive accounting for each exchanges over longer-term interactions in the wild. [Similarly] We feel good about other people in proportion to how nice they are to us, and [such emotions are used to] keep balance sheets of those sentiments and remember past interactions. (Christakis, p. 231-232)
Human Emotions Come From Their Animal Brains
A final observation that may have some bearing on this question is whether or not human emotions come from their ‘animal brain’ or ‘human brain.’ Our best theory of how brains evolved — consistent with how we know evolution works — is that evolution built newer areas of the brain, such as the neocortex (long believed to be the seat of human intelligence) on top of an evolutionary older and simpler animal brain. For example, the basic brain stem in a human — that keeps you breathing and your heart pumping — is more or less identical to the brain stem in an animal and some animals only have the brain stem.
Michael Golding, a Critical Rationalist Psychiatrist, was interviewed on the Do Explain podcast and he said the following:
For example, if I stimulate an almond-shaped little bit of brain tissue in a particular way — called the amygdala — I, as an adult, will feel a profound sense of rage or fear. Furthermore, we know that if I do surgery on that part of the brain, for example, if I remove the mesial temporal region… the individual [will have] virtually no anger… [and] doesn’t experience anger or rage….
So we know what part of the brain activates rage and fear and can even, on command, stimulate it in humans. And this area is in the animal portion of the brain, not the neocortex. This leads Michael to conclude:
Clearly, our mind… shares space, shares the theater of the mind with animal parts of our brain which are able to create [qualia] experiences [such as rage] for us as well.… So that, yeah, I like those examples so, so you could say we are… universal [explainers] running on mammalian hardware. I think that’s… accurate.
I would also note that there is actually extensive studies on this that have been done by Mark Solms that he documents in his book The Hidden Spring: A Journey to the Source of Consciousness. I will do a future post about this book. But the important takeaway is that he has been able to isolate which parts of the brain create affect/qualia and it is not the neo-cortex but is the parts of the brain we share with most animals. So Mark’s studies back up Michael’s view.
Far from being a mere cultural belief or superstition like Deutsch suggested, belief in animal qualia is actually an explanation forced upon us by the existence of some animal behaviors that can’t be explained without first conjecturing that they experience qualia. Pet owners who think their dog loves them, or is in pain, aren’t merely anthropomorphizing; they are explaining their animal’s behavior the only way available to them. This is the essence of what we mean by a ‘best explanation.’
However, some caution is warranted here. As I discussed in this post, it used to be the best explanation that animals had limited explanatory knowledge. But the later discovery of the theory of behavior parsing threw doubt on this theory. (Though I also documented examples of animal behavior that it’s not clear behavior parsing can explain.) So we should remember that all theories, even ‘best theories’ are always only tentatively held.
Moreover, it’s not true that all best theories should be embraced whole-heartedly. The theory that animals feel qualia may be currently necessary to explain some animal behavior, but it is not at all comparable to, say, biological evolution or the other four strands. This is because this is not yet a deep theory. For that we still await a full theory of qualia. So we should accept, for now, that we can’t explain animal behavior without referencing an inner life, but stay open to all ideas. 
I try to imagine what it would be like to have someone present to me a theory of animal consciousness where they could explain all of the above experiments and behaviors but without referencing an inner life for animals. I don’t know what such an explanation would look like, so I can’t really imagine it right now. That’s typical of situations like that. You don’t really know what a good alternative will look like until you actually see it.
But let’s say someone did present to me such an explanation. If this theory was deeper than the one I just explained, I think I’d likely jump to that new theory quite easily. I seem to feel no desire to insist on animal consciousness of any kind. I simply want to know what the truth of the matter really is and right now I happen to not know how to explain much animal behavior without reference to some sort of inner life. But I’m not even convinced this limited ‘animal inner life’ deserves the word ‘consciousness.’ My feelings about ‘animal consciousness’ simply don’t go beyond the fact that I don’t have a good alternative explanation to jump to yet.
 Deutsch seems to believe animals most likely do not experience qualia, but hasn’t ruled it out either. Here are the relevant tweets I could find.
In this tweet, he claims qualia isn’t needed to explain animal signaling and aping behavior. This seems to be referencing the idea that behavior parsing explains apparent animal explanatory knowledge, as I discussed in this post. But I’m unclear why he’s here linking that argument with qualia. In this tweet, he agrees some caution is warranted because animals may be able to suffer. In this tweet, he agrees with me that the word “conscious” has such a range that it’s pointless to ask if animals are conscious. However, he apparently considers it unlikely that animals have a form of consciousness that has moral ramifications. On the other hand, this tweet seems to indicate that maybe he does believe animals experience qualia. And finally, he claims animals do nothing that can’t be explained by current artificial intelligence. A conjecture I refuted in this post.
I believe this post raises examples of observations that Deutsch’s above conjectures can’t explain, thus causing problems for his original theory that most likely animals do not experience qualia.
However, more recently, on my podcast, Deutsch actually seemed to express the possibility that dogs wouldn’t be able to act like they have emotions without actually having them. Ultimately he expresses caution in drawing any conclusions and suggests we must allow for many opinions on this subject and not be premature in jumping to conclusions. (See his full answer here.)
 From Christkas’ book Blueprint:
Cetaceans, primates, and elephants, like humans, appear to treat bodies of deceased love ones gently, not as mere inanimate objects. One orca whale mother, assisted by her pod, was observed buoying up her dead baby for three days in a matter one expert described as follows: “They know the calf is dead. I think this is a grieving or a ceremonial thing done by the mother… She doesn’t want to let go.” Primatologists have seen adult female chimpanzees (both related and unrelated) long after the corpses have started to decompose. Chimpanzees have even been filmed cleaning a corpse’s teeth in a mortuary-ritual-like behavior. (p. 301)
 From Christkas’ book Blueprint:
Ethologist Joyce Poole spent years observing African elephant societies. Once, after a prolonged absence, she eagerly returned to them….
The elephants were less than two meters away from us when Vee stopped dead in her tracks and, with her mouth wide, gave a deep and loud throaty rumble. The rest of the family rushed to her side, gathering next to our window, and with their trunks outstretched, deafened us with a cacophony of rumbles, trumpets, and screams until our bodies vibrated with the sound. … What we had experienced was an intense greeting ceremony usually reserved only for family. (Poole, p. 217)
Poole and others researchers have observed countless displays of compassion and friendship. … [Gives an example of elephants helping sick elephants and walking besides them to help them along.] Poole recognizes the likely evolutionary explanation for such friendly and altruistic behavior: elephants may be hardwired to assist members of their clan simply to increase their own genes’ chance of survival. “But if that is the case,” she writes, “why do elephants assist injured elephants and injured humans, but not, as far as we know, other species?” That point suggests a set of evolutionary guided abilities that transcend simple kin selection. Elephants seem to form opinions and make judgments about altruism. They recognize when a human is actually in need of help. (p. 218-219)
…elephants also form friendships with entirely unrelated individuals… highlighting the fact that social interactions and assistance are not solely between kin. (p. 221)
 The fact that our best theory is that animals do feel qualia does have some moral implications. However, in a future post, I will cover why our best theories of animal qualia actually suggest that animals have nothing like the moral status that humans do.