How animals feel pain

On Saturday, I wrote this bit about whether animals feel pain, and I said then I’d follow up that day or the next. I didn’t! I’ve been doped up on painkillers and my brain is all soft around the edges. I finally decided to give up on them this morning when I woke up, had breakfast, and then fell asleep for a couple of hours. Enough already. Now I have to recover from some potent drugs as well as my back pain.

Anyway, where were we? Oh, right, we had taken down William Lane Craig’s argument that humans (and maybe some other primates) are the only creatures on the planet that actually feel pain, that other animals are mere meat robots who act out a superficial script that looks like they are in pain, but really, they have no consciousness to experience the pain.

I don’t think he is making this argument to warrant or excuse animal torture, but rather he’s trying to justify human exceptionalism. See, humans have this special god-granted ability to perceive suffering and pain, which is why we have souls and animals don’t, and why we have to worry about that Final Judgment, since we can sense and appreciate the harm we do to others. At least, that’s my charitable assessment.

Curiously, in order to make this argument, Craig feels a need for some scientific backing, some recognizable neuroanatomical feature that shows humans are special. I don’t get it. He already believes in something invisible and intangible, the soul, so why not just say humans posess an invisible magical flimflam that the scientists can’t see or experiment on, neener neener, therefore humans are unique in having a conscience or ghost or pneuma that gives them the abilty to really truly deeply feel pain and suffering?

Often the physical substrate for feeling pain is determined in a backwards sort of way: we find some feature of the human brain that is only found in us, or is more pronounced in us, and we decide that, aha, that must be where this higher ability resides. Some of the common culprits are our enlarged pre-frontal cortex (PFC) or more narrowly, the anterior cingulate cortex (ACC). The ACC was a favorite of Francis Crick, for instance, who thought that not only was it the seat of awareness, but also of free will (I think free will is a non-existent phenomenon that makes no sense, but I’ll defer on discussing that to another time).

I don’t think the hypothesis is far out of line — there is evidence that lesions in this area do cause sensations of dissociation, and it is entangled in a lot of higher brain functions. But on the other hand, other animals have these structures, so how do you use this phenomenon to make humans exceptional? You can’t.

This leads me to an article in a journal called Philosophical Psychology, Against Neo-Cartesianism: Neurofunctional Resilience and Animal Pain by Halper et al.. Mixing philosophy with psychology and throwing in a lot of ethology and neuroscience sounds like a potent way to address the issue, don’t you think? Especially with a feisty abstract like this one:

ABSTRACT: Several influential philosophers and scientists have advanced a framework, often called Neo-Cartesianism (NC), according to which animal suffering is merely apparent. Drawing upon contemporary neuroscience and philosophy of mind, NeoCartesians challenge the mainstream position we shall call Evolutionary Continuity (EC), the view that humans are on a non-hierarchical continuum with other species and are thus not likely to be unique in consciously experiencing negative pain affect. We argue that some Neo-Cartesians have misconstrued the underlying science or tendentiously appropriated controversial views in the philosophy of mind. We discuss recent evidence that undermines the simple neuroanatomical structure-function correlation thesis that undergirds many Neo-Cartesian arguments, has an important bearing on the recent controversy over pain in fish, and puts the underlying epistemology framing the debate between NC and EC in a new light that strengthens the EC position.

In one corner, the Neo-Cartesians like William Lane Craig (there are also secular philosophers who like this idea).

In the other corner, the Evolutionary Continuity camp, which I would happily attach myself to, which argues for “a non-hierarchical continuum with other species and are thus not likely to be unique”. Yay Team EC!

The paper quickly dispatches part of the argument. All mammals have a prefrontal cortex, although the human PFC is relatively larger. So now you’re going to have to argue, if you think the PFC is the seat of awareness, that a quantitative difference leads to a qualitative change, and you’re also going to have the problem of a continuum of PFC sizes. Where do you draw the line?

Well, maybe it’s the anterior cingulate cortex, rather than the whole PFC, that matters. They have a case study that refutes that. Patient R is a man whose brain was devastated by herpes simplex encephalitis, yet survived with normal intelligence and anterograde amnesia (loss of the ability to form new memories). Most importantly to this point, he has retained full self-awareness.

Patient R, aka Roger (Philippi et al. 2012), provides us with a novel angle for assessing certain versions of the NC hypothesis. Roger has extensive damage to his PFC as well as his ACC and insula, bilaterally. He has been probed for self-awareness (SA) in numerous ways, some standard, some novel with positive results for all probes. The authors of the 2012 study concluded that SA is likely a function of the interaction of multiple brain regions, with some redundancy, rather than dependent upon one particular region.8 Roger’s case, like others described in the literature (Damasio et al. 2013), seems to demonstrate fairly conclusively that the PFC, the ACC, and the insula are not needed for SA, including SA of a fairly sophisticated sort, a sort we need not presume animals to have to make our case here.

Patient R also has a normal physical and emotional response to pain — if anything, he now reacts more strongly.

So now if you want to argue for a discrete localization of self-awareness in the brain, you’re either going to have to pick a different brain region or claim that Patient R is a p-zombie. Or, perhaps, that the brain has a lot more flexibility than was thought. I like this last idea, but then, that makes the pursuit of a feature unique to humans futile.

Further, these cases suggest an alternative to the rigid structure-function correlation thesis. Resilience of function following brain damage suggests the existence of degrees of freedom in the relationships between certain functions and neuroanatomical structures (Rudrauf, 2014). Anatomical regions and networks normally supporting central psychological functions (like the emotional appraisal of pain) may simply be the usual defaults. In patients such as Roger, functional resilience after such extensive and irreversible anatomical damage cannot be explained by structural plasticity, in the sense of anatomical restoration or large-scale “rewiring” (e.g., dendritic sprouting and axon regeneration) to restore structural connectivity. Large-scale functional networks supporting key psychological functions, however, can be maintained or restored even when the integrity of normal anatomical networks is massively and irremediably compromised.

This suggests that a different concept of flexibility is more apt, namely what Rudrauf (2014) has called “neurofunctional resilience”. This concept is based on the phenomenon of preserved function in spite of large-scale architectural changes, is not limited to one specific class of mechanisms or levels of observation, and indicates a relative openness in implementation at various levels of a functional hierarchy. The neurofunctional resilience framework, while in need of elaboration and refinement, makes more sense of lesion cases like Roger’s, observed variation in structure-function relationships in imaging studies, interspecific structure-function variations, the relative unimportance of lesion locales versus lesion size vis-à-vis functional deficits in the developing brain (Pascual, 2017, 5; Battro, 2000), and better fits general theoretical considerations about multiple realizability drawn from computational neuroscience. In realizing that a crude, “phrenological” localizationist structure-function paradigm (even one incorporating plasticity) is unable to account for these observable phenomena, one need not, of course, embrace the old holistic, “equipotentiality” theories of brain function (see Finger, 1994, ch., p. 4). A new, subtler approach is needed.

My first thought would have been that Patient R is an amazing example of neuronal plasticity, but the author is right: this is something more impressive. Big chunks of the brain are just gone; minor self-repair mechanisms, like neurons regrowing around a damaged pathway, are not sufficient. This is as if your car had the electronic engine timing system blown off, so the wires were rerouted to make use of circuits in your car radio instead. Be impressed! Brains seem to have a biological imperative to assemble themselves into some kind of cognitively functional structure, in spite of massive damage.

But never mind human brains — they’re too complicated, and you can’t do experiments on them. What about fish brains? Do they feel pain? And what about cephalopods?

As Godfrey-Smith (2016, 94 f.) notes, flexible behaviors and preference changes related to pain avoidance and analgesia-seeking (observed in both fish and chickens) in entirely evolutionarily novel situations and perhaps certain grooming and protecting behaviors associated with bodily damage are arguably best explained in terms of the presence of consciously experienced negative pain affect. And when one considers the overall behavioral, affective, and cognitive repertoire of, for example, cephalopods, as Godfrey-Smith does at length, the notion that such an animal, whose nervous system is so different from ours, does what it does in the absence of consciousness begins to look implausible and continued commitment to it perversely skeptical.

It seems more reasonable to follow the approach of Segner (2012, p. 78) who, in considering fish pain, looks at seven relevant properties: (1) nociceptors, (2) pain related brain structures homologous or analogous to those found in humans, (3) pathways connecting peripheral nociceptors to higher brain regions, (4) endogenous opioids and opioid receptors in the CNS, (5) analgesic-mediated reduction of response to noxious stimuli, (6) complex forms of learning, including avoidance learning of noxious stimuli, and (7) suspension of normal behavior in response to noxious stimuli. Humans and fish, Segner concludes, unequivocally share all but item (2), which is partially shared: we share subcortical structures with fish but not the neocortical structures. However, given the evidence reviewed in this section, it is clear that the neocortical structures commonly thought to be necessary for pain affect are not required in any case (cf., Merker, 2007; Ginsburg & Jablonka, 2019) Surely the other similarities are sufficient to make reasonable the inference to the presence of consciously felt fish pain (cf., Tye, 2017, 91ff.). For mammals, as we have seen, all of these similarities are in place.

That “phrenological” approach doesn’t work well for fish, and even less well for cephalopods which have virtually no homology with our brains. Those seven criteria are a useful rubric for figuring out if a given brain can be aware or feel pain, and like the authors say fish meet all of the criteria except #2, having homologous pain-related brain structures. But we also just saw that Patient R fails to have homologous pain related structures. It would be strange to then assert that an organism that has the other six features would have failed, in its evolutionary history, to have incorporated them into an integrated pain awareness system.

Segner codifies the basic analogy argument for the presence of negative affect in animals that goes beyond the one we all spontaneously draw from our admittedly fallible raw intuitions. On our view, this basic analogy argument coupled with the considerations about neurofunctional resilience (and evolutionary analogies) we have adduced yield a reasonably high probability for the claim that negative pain affect is present in mammals, avians, fish, and cephalopods. Even if we are mistaken about the latter three, however, these considerations make the claim that it is present in mammals so probable that the more ambitious NC thesis of M. Murray and W.L. Craig is cast into nearly insurmountable doubt.

The paper goes on to discuss in detail the specific question of whether fish feel pain (short answer: yes), cetaceans, the issue of blindsight, and much more briefly, consciousness, which would require a stack of books to consider. I’ll stop here, though, disappointed that nowhere does the paper discuss spider pain, or any invertebrates other than cephalopods. Invertebrates are so alien and distantly removed from us that it is nearly impossible to discern a pain affect in them, but they also meet 6 of the 7 Segner criteria. The pain pathways in vertebrate and invertebrate systems also show homology at the molecular level, and it’s possible to see similarites across many phyla.

Maybe that’s for a different day. I’ve been having a fun time lately diving into an area I’ve neglected for a while, developmental neuroscience, and maybe I’ll be motivated to tell you all why you should be kind to bugs, because they have feelings, too, and maybe even experience suffering.