August 1, 2019

This piece is part of a short series on how our research team thinks about uncertainty. These posts represent the opinions of individual staff members, and not necessarily the organizational position of Wild Animal Initiative on uncertainty.

In this opinion piece, I explore an important area of uncertainty in wild-animal welfare research: determining who is a moral patient, using insects as a case study. A moral patient is one whose well-being deserves moral consideration. I take welfare to be intrinsically valuable; I also think that only (and all) sentient individuals have welfare. The definition of sentience I use is primary or phenomenal consciousness: the capacity to experience good or bad mental states. Sentience can occur in the absence of secondary (or access) consciousness: the ability to reflect upon or report those experiences. 

Determining whether or not an organism is capable of having negative mental experiences (suffering) is important for understanding where we should focus our welfare efforts. If we are uncertain about whether or not a being is sentient, we cannot be certain that being is a moral patient. I don’t think you can be truly certain of any being’s sentience besides your own. Yet, we make moral choices affecting others all the time, which means we are making assumptions about others’ moral patienthood. How should we deal with this uncertainty? 

Insects[1] are an interesting case study for handling uncertainty about moral patienthood. They lack some of the features traditionally important to our assumptions about what kinds of beings deserve moral consideration, but there is a good case to be made for some insects being sentient. Their behavior and anatomy are better-studied than many other invertebrate groups’, so we have some data to draw upon when developing an opinion about their moral patienthood. Because insects are more marginal in most of our moral circles than other wild animals, they are also a useful test case for our intuitions about moral patienthood and our duties as moral agents.

Informing our opinions

Many of your opinions about moral issues are based on intuition, and then refined or challenged through analysis. You can observe these opinions through thought experiments or practical experience. These observations are useful because they require you to arrive at an initial, intuitive response and then try to distill the factors that contributed to this response. Upon reflection, you may reject some of the premises of your intuitive response, such as a bias that you are trying to eradicate. Analysis of your intuitive response to a situation may also help you clarify the values you hold, and articulate a more internally consistent and reproducible moral framework. 

For example, despite their aversion to insects, many people avoid doing things that might hurt them if they are capable of suffering. This includes moving insects outside rather than killing them directly, avoiding stepping on insects on the sidewalk, and crushing insects as quickly as possible. These actions seem to suggest that many people intuitively believe insects are sentient[2], at least in the sense that they are able to experience pain. Intuitions like this one are formed unconsciously from personal experiences and social influences. When you state your intuitions explicitly, you can compare them to your consciously-formed ideas. Your past observations of an ant’s goal-directed behavior and your tendency to anthropomorphize may both contribute to your intuition that you should try not to step on ants. When you analyze this intuition, you may decide based on your consciously-formed idea that sentience is a biological phenomenon with observable correlates that your behavioral observations are useful data for forming this moral opinion, while your attribution of complex emotions is not. 

Intuitions are a good launching point for forming moral opinions, but it is important to determine if your intuitions are supported by reliable evidence. In order to ensure your opinions are not unduly influenced by unconscious bias or other features which you consciously disapprove of, they need to be reconsidered and updated when you encounter new information. Evidence relating to moral patienthood is especially tricky, because if we accept that sentience is necessary and sufficient for moral patienthood, and that sentience is a biological phenomenon, then we run into the “hard problem of consciousness”: how do subjective experiences arise from physical processes? Because there are no particular bits of neuroanatomy which we can point to as conclusive proof of sentience, developing best guesses as to which kinds of organisms are sentient requires a dual approach: “from the ground up,” where we make inferences based on neuroanatomy, and “from the top down,” where we infer based on behavior (Klein & Barron 2016, Mallatt & Feinberg 2016). 

Sentience may have emerged because it confers a selective advantage to motile organisms in changing environments. The idea is that they can navigate, survive, and reproduce more effectively and efficiently when they process the world in a more integrated, centralized way (Klein & Barron 2016). With this hypothesis in mind, we can identify a few features that are probably associated with or required for sentience. One is the presence of physiological characteristics related to affective states in humans, like an endogenous opioid system. It also seems important to display self-protective behaviors and the ability to learn and modify behavior to avoid harmful stimuli (Ray 2018). Behavioral studies can also provide evidence for an organism forming and referring to a mental map of their environment, and a sense of where the self is in relation to that map (Carpendale 2019, Klein & Barron 2016, Mallatt & Feinberg 2016). 

It is reasonable to assume that having neuroanatomy similar to that of organisms in whose sentience we are more confident, such as mammals, would be good evidence of sentience. The inverse, however, is not as clear: just because an organism has neuroanatomy that is significantly different from or simpler than mammals’ does not mean we should discount them from moral consideration. A frontal cortex was once considered requisite for sentience, but cognitive scientists are now moving towards a more functional model. It still seems to be true that an entirely decentralized nervous system is unlikely to support sentience, but simpler brains like insects’ have functional regions which are analogous to those of vertebrate brains (Klein & Barron 2016, Mallatt & Feinberg 2016). In addition, insect brains are capable of supporting complex behaviors with surprisingly few neurons; they are quite efficient, and have less redundancy than vertebrate brains (ibid.). 

Anatomical, behavioral, and physiological features which seem relevant to sentience are detailed by Sneddon et al. in “Defining and assessing animal pain” (2014)[3]. By their assessment, insects show many of these features, although not as many as mammals - this is unsurprising, as the current understanding and definition of sentience was expanded upon from humans and their close relatives. If an organism has some of these features, that is evidence to support that the claim that they are sentient, but it is not conclusive proof, just as the lack of one or more of these features is not proof that an organism is not sentient. Accumulating data such as this can help you reduce your uncertainty and allow you to form an evidence-based moral opinion about moral patienthood without requiring you to have a precise understanding of the nature of sentience. Sneddon et al. refer to this process as the Principle of Triangulation (2014). 

Applying our opinions

When our opinions about moral patienthood are based on uncertain evidence, it is important to consider that uncertainty when acting on our opinions. Expected value estimates are one method for approaching this problem. We obtain an expected value estimate for a moral choice by assigning a probability that a relevant premise is true and multiplying that probability by the magnitude we assign to the value or disvalue of a given outcome. The expected value of an intervention to improve insect welfare will depend on the confidence we have in insects being sentient (or how intense we think their experiences are) and the magnitude of the improvement to insect well-being we anticipate the intervention will provide. The usefulness of an expected value estimate depends on how well-informed our estimates for each component are. If we are only 10% confident that insects are sentient, but the sheer number of insects means that the value of their potential welfare is enormous, then we may find it worthwhile to address insect welfare issues, even though we are uncertain about whether or not insects have welfare. 

This idea is the Animal Sentience Precautionary Principle: “Where there are threats of serious, negative animal welfare outcomes, lack of full scientific certainty as to the sentience of the animals in question shall not be used as a reason for postponing cost-effective measures to prevent those outcomes” (Birch 2017). The application of this principle is difficult[4], as the animals whose sentience we are least sure of are also the most numerous. Evidence of sentience may also not be generalizable across taxa or even life stages of the same species (Adamo 2017). When used in conjunction with the Principle of Triangulation, however, the Animal Sentience Precautionary Principle is a pragmatic approach to handling uncertainty about moral patienthood.

We should maintain open minds and update our assumptions about sentience with new evidence. Further research into the origins and neurobehavioral correlates of sentience may eventually allow us greater certainty about who is a moral patient. In the meantime, we can do our best to minimize harm based on how confident we are that an organism is sentient, and the degree to which they seem to be sentient. In my opinion, the most parsimonious explanation for the complex and plastic behaviors insects show is that they are sentient, and we should err on the side of assuming they have interests and affective states. This opinion may not greatly affect your day-to-day interactions with insects. Like a lot of people, my intuition is that I should avoid directly causing harm to insects, and if I encounter an insect or other invertebrate in trouble I try to help - moving an earthworm from the sidewalk to grass, for example, to help them avoid death by desiccation or predation. It should, however, impact your consideration of insects at a larger scale, such as when applying agricultural insecticides or assessing the welfare impacts of switching from larger farmed animals to insect-based foods. Operating with some degree of uncertainty about sentience and moral patienthood is necessary in wild animal welfare work. If we are intentional about how we engage with our intuitions and the available evidence, however, we will be able to make better decisions about interventions regarding animals about whose sentience we are uncertain. 

Footnotes

[1] I will be using “insects” for simplicity, but I also have arachnids, isopods, centipedes, and other terrestrial arthropods in mind.

[2] An alternative explanation is that while people don’t think insects are capable of painful experience, they think that demonstrating consideration for insects as if they were sentient contributes to a more empathetic culture. A complementary concept is that treating insects in a way that is presumably painful indicates that someone is cruel and unempathetic in general, so that behavior is typically frowned upon.

[3] For a detailed report on features probably related to sentience, see Rethink Priorities’ piece “Invertebrate sentience: A useful empirical resource,” part of their series on invertebrate sentience (Schukraft 2019).

[4] Birch is careful to use “cost-effective” and “full scientific certainty” to qualify his principle; the ASPP would otherwise be quite impractical. Birch does not explore what full scientific certainty or cost-effectiveness would look like; see Adamo 2017 for these and other criticisms of the ASPP.

References

Adamo, S. (2017). The “Precautionary Principle” - A work in progress. Animal Sentience 16(4). Retrieved 29 July 2019. https://animalstudiesrepository.org/cgi/viewcontent.cgi?article=1222&context=animsent. 

Birch, J. (2017). Animal sentience and the precautionary principle. Animal Sentience 16(1). Retrieved 29 July 2019. https://animalstudiesrepository.org/cgi/viewcontent.cgi?referer=http://animalstudiesrepository.org/&httpsredir=1&article=1200&context=animsent. 

Carpendale, M. (2019). Interview with Jon Mallatt about invertebrate consciousness. Effective Altruism Forum. Retrieved 29 July 2019. https://forum.effectivealtruism.org/posts/5FesXkhArfEXF47mn/interview-with-jon-mallatt-about-invertebrate-consciousness. 

Klein, C. & Barron, A. B. (2016). Insects have the capacity for subjective experience. Animal Sentience. 9(1). Retrieved 19 May 2019. https://animalstudiesrepository.org/cgi/viewcontent.cgi?referer=http://animalstudiesrepository.org/&httpsredir=1&article=1113&context=animsent. 

Mallatt, J. & Feinberg, T. E. (2016). Insect consciousness: Fine-tuning the hypothesis. Animal Sentience. 9(10). Retrieved 29 July 2019. https://animalstudiesrepository.org/cgi/viewcontent.cgi?article=1141&context=animsent. 

Ray, G. (2018). Invertebrate sentience: Urgent but understudied. Wild Animal Suffering Research Archives. Retrieved 29 July 2019. https://was-research.org/paper/invertebrate-sentience-urgent-understudied/.

Schukraft, J. (2019). Invertebrate sentience: A useful empirical resource. Effective Altruism Forum. Retrieved 31 July 2019. https://forum.effectivealtruism.org/posts/AMFuSWrsutBFraZtE/invertebrate-sentience-a-useful-empirical-resource. 

Sneddon, L. U., Elwood, R. W., Adamo, S. A., & Leach, M. C. (2014). Defining and assessing animal pain. Animal Behaviour. 97: 201-212. Retrieved 31 July 2019. https://animalstudiesrepository.org/cgi/viewcontent.cgi?referer=http://animalstudiesrepository.org/&httpsredir=1&article=1068&context=acwp_arte