Invertebrate Welfare Cause Profile

Cross-posted on the EA Forum

Executive Summary

More than 99.9% of animals are invertebrates. There is modest evidence that some large groups of invertebrates, especially cephalopods and arthropods, are sentient. The effective animal activism community currently allocates less than 1% of total spending to invertebrate welfare. That share should rise so that we can better understand invertebrate sentience and investigate the tractability of improving invertebrate welfare.

Introduction and Context

This post is the tenth in Rethink Priorities’ series on invertebrate welfare. In the first post we examine some philosophical difficulties inherent in the detection of morally significant pain and pleasure in nonhumans. In the second post we discuss our survey and compilation of the extant scientific literature relevant to invertebrate sentience, as well as the strengths and weaknesses of our approach to the subject. In the third post we explain some anatomical, evolutionary, and behavioral features potentially indicative of the capacity for conscious experience in invertebrates. In the fourth post we explain some drug responses, motivational tradeoffs, and feats of cognitive sophistication potentially indicative of the capacity for conscious experience in invertebrates. In the fifth post we explain some learning indicators, navigational skills, and mood state behaviors potentially indicative of the capacity for conscious experience in invertebrates. The sixth post announces our Invertebrate Sentience Table. In the seventh and eighth posts, we present our summary of findings by feature and by taxa. The ninth post asks what we can learn about sentience from examining process that operate unconsciously in humans.

In this post we apply the standard importance-neglectedness-tractability framework to invertebrate welfare to determine, as best we can, whether this is a cause area that is worth prioritizing. We conclude that it is. In a separate post, slated to be published next month, we present and examine the best arguments against our analysis.

What Is Invertebrate Welfare?

Invertebrates1 comprise an enormous and diverse array of animals, from nematodes and earthworms to jumping spiders and jellyfish, crabs and krill to cuttlefish and cockroaches. Because invertebrates are a large and heterogeneous class of animals, there are few things that can be said about invertebrates in general (other than that they are animals that lack a backbone). Moreover, it is uncertain which (if any) invertebrates have the capacity for valenced experience,2 so it is unclear whether these animals have a welfare. Thus, the term ‘invertebrate welfare’ is inevitably misleading.3 Using the term to denote a cause area is something of a terminological simplification.

If invertebrate welfare were a mature field, it would encompass perhaps dozens of distinct and unrelated interventions.4 A campaign to promote the use of humane insecticides is quite different in kind from a campaign to promote strict laboratory standards for the treatment of octopuses.5 What the interventions have in common is that they concern a group of animals whose welfare has historically been ignored.

When we consider the arguments for and against prioritizing invertebrate welfare as a cause area, what we are considering is whether we should prioritize learning more about this group of animals. At this early stage, supporting the cause of invertebrate welfare means supporting additional research on invertebrate sentience, advocacy strategies, and cost-effective interventions. Opposing the cause means de-prioritizing this research. It’s possible to support the cause now, and, as the results of the additional research come in, later oppose the cause.

Complicating matters further, the moral status of invertebrates is almost certainly not uniform across the category. Supporting invertebrate welfare as a cause area does not mean championing all invertebrates equally. Caring about highly intelligent cephalopods doesn’t entail caring about brainless cnidarians. Part of the invertebrate welfare project is determining which invertebrates (if any) matter and to what degree.

Scale and Limiting Factors

The (Rough) Number of Invertebrates

First, a caveat: estimating the number of invertebrates is hard. Most invertebrates are small, wild, short-lived, geographically dispersed, and, as a result, difficult to systematically collect and count. Moreover, there are few academic or economic motivations to undertake the arduous task of estimating global population sizes for most invertebrates. As a consequence, many of the estimates presented in this section could easily be wrong, in either direction, by one or more orders of magnitude. Nonetheless, one fact is clear: there are many, many more invertebrates than vertebrates.

Invertebrates account for approximately 85% of all animal biomass. However, because the mass of individual animals varies by many orders of magnitude across taxa, this figure vastly understates the extent to which individual invertebrates outnumber individual vertebrates. More than 98% of all extant animal species are invertebrates. Again, though, because the number of animals per species is not uniform across taxa, this figure also understates the extent to which individual invertebrates outnumber individual vertebrates. Recent estimates put the total number of animals on earth at around a sextillion. Of that sextillion, approximately 99.9998% are invertebrates.6 From the point of view of total individuals, ‘animals’ is basically synonymous with ‘invertebrates.’

Of course, from a welfare perspective, the total number of invertebrates is irrelevant if only a handful of invertebrate species are sentient. Annelids and nematodes are extremely numerous,7 but they are less plausibly moral patients than, for example, coleoid cephalopods (such as octopuses and cuttlefish), decapod crustaceans (such as crabs, lobsters, and crayfish), and eusocial insects (such as ants, bees, and termites). It will be helpful, then, to look at more fine-grained population estimates.

Total insects number somewhere between 1 and 10 quintillion,8 of which ants are by far the most numerous. (Ants alone account for roughly 15% of terrestrial animal biomass.) Antarctic krill (Euphausia superba) are the most numerous marine arthropod (and, by weight, the largest single animal species on the planet). Recent studies put the total number of Antarctic krill at any given moment in time at 800 trillion.9 (That’s just shy of the total number of individual fish across all fish species.) estimates that between 220 billion and 526 billion decapod crustaceans were slaughtered in aquaculture production in 2015 alone.10 (This figure doesn’t include decapods that died before slaughter.) Even the lower estimate is almost three times the number of land vertebrates slaughtered for food each year. Based on an analysis of cephalopod prey populations, it has been estimated that the total biomass of cephalopods is approximately 0.05 gigatonnes carbon.11 (Individual cephalopods range in mass from less than a gram to almost 200 kilograms for the giant squid.) That number compares favorably to the total biomass of humans at 0.06 gigatonnes carbon, and far outstrips the biomass of wild mammals (0.007 Gt C) and wild birds (0.002 Gt C).12

These numbers alone don’t tell us that invertebrate welfare is an important cause area. Even if most invertebrates are sentient, it might be the case that there is little we can do to improve their welfare, either due to a lack of cost-effective interventions or perhaps because invertebrate lives are already mostly positive. We know precious little about the life history of most invertebrates and even less about the interventions that might improve those lives. The point of this section is to show that if there is a non-negligible chance that invertebrates have the capacity for morally significant valenced experience, then it follows that there is a huge pot of expected moral worth out there that has hitherto been almost completely neglected. To reiterate a point above, supporting invertebrate welfare at this stage of our understanding merely entails supporting additional research into this group of animals.

Moral Weight of Invertebrates

One subject that looms large here and must be addressed directly is the relative moral weight of (various groups of) invertebrates. Even if it’s true that there are quintillions of invertebrates with the capacity for valenced experience, if the moral significance of these experiences is negligible, then the total moral value of all invertebrates may not amount to much.

Questions of moral weight are notoriously difficult to adjudicate. To simplify a bit, we can assume that, for a fixed lifespan, if creature x has less moral weight than creature y, it’s because x has fewer intrinsic morally significant properties and/or x has those properties to a lesser degree.13 Of course, we don’t know what all the morally significant properties are, and even if we did, we probably don’t have epistemic access to all of them. (For example, even if we were certain that general intelligence were the only morally significant trait, we would still have a hard time determining relative moral weight with much precision because general intelligence is extraordinarily difficult to measure across dissimilar taxa.) So discussions of moral weight inevitably rely on imperfect proxies and even more imperfect intuitions.

Range of Phenomenal Intensity and Related Issues

Suppose that the only intrinsic morally significant feature of an entity is its phenomenally conscious experiential states. For creatures with the capacity for valenced experience, these states can come with a positive or negative affect. But of course, the distinction is not binary. Some positive states are much better than other positive states, and some negative states are much worse than other negative states. Call this the phenomenal intensity of valenced experience. It might be the case that, ceteris paribus, the greater the range of phenomenal intensity a creature is capable of experiencing, the more morally valuable that creature is. (That is to say, creatures capable of experiencing higher highs and lower lows are worth more, morally, than creatures whose experience tends to be neutral.)

Some people have the intuition that if invertebrates have the capacity for valenced experience, their phenomenal range would be much narrower than the phenomenal range of vertebrates. For instance, Matt Ball writes, “So even if insects can have any subjective experience, their most intense sensation would be the palest hint of a feeling—a tiny fraction of the worst suffering we can experience.” It’s not clear what exactly motivates this view. It’s plausible that there are differences of phenomenal intensity across phylogenetically distant taxa, but it’s certainly not obvious that the difference would be as stark as Ball suggests. It's unclear what fitness advantage the palest hint of a feeling could convey. Pain motivates animals to do things like avoid bodily damage; pleasure motivates animals to do things like reproduce. Subjective experiences so faint as to barely register would do a poor job motivating anything.

Perhaps the reason some people think invertebrates have a diminished range of phenomenal intensity is that they think invertebrates are, in some sense, less conscious than vertebrates. Here we must tread carefully because we are on shaky theoretical ground. According to our definition, an entity is conscious if and only if there is something it feels like to be that entity, no matter how strange or faint that phenomenology might be. So in our sense, consciousness is binary: either an entity is conscious or it is not. Nonetheless, there are certain aspects of consciousness that admit of gradations, and variations of these gradations might be morally relevant.

There are at least three mundane ways in which consciousness, loosely speaking, might come in degrees. An entity that is conscious might be conscious all the time or only part of the time. (Animal sleep cycles range from 2 to 20 hours a day.14) For an entity that is currently conscious, consciousness might span many or few modalities. (Some creatures are sensitive to differences in light, sound, temperature, pressure, odor, bodily orientation, and magnetic field. Other creatures are sensitive to far fewer sensory modalities.) For an entity that is currently conscious of a given sensory modality, that modality might be coarse-grained or fine-grained. (Within the vision modality, some creatures are only sensitive to differences in brightness, while other creatures are also sensitive to hue and saturation.)

In this loose sense, it appears to be the case that if invertebrates like nematodes (roundworms), cnidarians (coral/jellyfish/anemones), poriferans (sponges), and annelids (earthworms/leeches) are conscious, they are conscious to a lesser degree than vertebrates. But for precisely this reason, there is little evidence that these creatures are conscious at all. According to many researchers, the role of consciousness is to integrate diverse streams of information into one global workspace.15 If an animal only receives coarse-grained information from a single sensory modality, there is no need to develop a global workspace. The invertebrates for which we have the best evidence of sentience, arthropods and coleoid cephalopods, do appear to process distinct sensory modalities into a unified whole.16 It’s still possible that these creatures are less conscious (in the loose sense) than the average vertebrate, but it seems unjustified to conclude that they are drastically less conscious than the average vertebrate.

Another way to potentially get a handle on the phenomenal intensity of nonhuman experience is to consider again the evolutionary role that pain plays. Pain teaches us which stimuli are noxious, how to avoid those stimuli, and what we ought to do to recover from injury. Because intense pain can be distracting, animals in intense pain seem to be at a selective disadvantage compared to conspecifics not in intense pain. Thus, we might expect evolution to select for creatures with pains just phenomenally intense enough (on average) to play the primary instructive role of pain. Humans are among the most cognitively sophisticated animals on the planet, plausibly the animals most likely to pick up on patterns in signals only weakly conveyed. In general, less cognitively sophisticated animals probably require stronger signals for pattern-learning. If pain is the signal, then we might reasonably expect the phenomenal intensity of pain to correlate inversely with cognitive sophistication.17 If that’s the case, humans might experience (on average) the least intense pain in all the animal kingdom.18

A final consideration involves not the phenomenal intensity of pain but its phenomenal extension (that is, its felt duration). Due to neurological differences, phenomenal extension might not be uniform across dissimilar taxa. Consider brain-processing speed and rates of subjective experience, both loosely defined. Animals with faster metabolisms and smaller body sizes tend, according to some metrics, to process information faster. Thus, there is some reason to think that smaller animals have, in general, faster subjective experiences. So a honey bee might experience one minute of objective time as longer, in some robust sense of the term, than a human would. If that’s true, then a given honey bee and a given human experiencing a pain of the same phenomenal intensity over the same objective duration of time would not, ceteris paribus, suffer equally. The honey bee would suffer more. Hence, we should not naively equate the phenomenal extension of pain with its duration expressed in objective time. The takeaway here is that the moral significance of pains and pleasures might be related in important ways to an entity’s processing speed. As with other areas, more research is needed.

Other Morally Significant Properties

Range of phenomenal intensity is probably not the only intrinsic property relevant to moral weight. When philosophers discuss gradations of moral status, they typically invoke a wider set of features. For instance, David Degrazia, an ethicist at George Washington University, describes “a sliding-scale model, according to which there are any number of degrees of moral status. On this view, the degree of consideration to which you are entitled—that is, the degree of moral weight your interests are to receive in comparison with others’ comparable interests—depends on the degree of your cognitive, affective, and social complexity” (emphasis added).19 Later in the same paper he writes, “Personhood... is a cluster concept that serves as a summary placeholder for other concepts such as moral agency, autonomy, the capacity for intentional action, rationality, self-awareness, sociability, and linguistic ability. But most... of these properties can be reasonably understood as coming in degrees; and many of them... are found to some degree in nonhuman animals. Now, if these morally relevant properties come in degrees and cross species boundaries, it is natural to judge that the moral status based on these properties also comes in degrees while extending beyond our species—supporting the sliding-scale model of moral status” (emphasis added).20

How do invertebrates stack up against vertebrates according to these criteria? It depends on the invertebrate. Moon jellyfish don’t display much cognitive, affective, or social complexity, and there’s little reason to suspect they possess moral agency, autonomy, the capacity for intentional action, rationality, self-awareness, or linguistic ability. Other invertebrates are more impressive. Consider the large order Hymenoptera (which includes wasps, bees, and ants). One literature review documented 59 distinct behavior types in honey bees; that number compares favorably against many mammalian species, such as the North American moose (at 22), De Brazza monkeys (at 44), and bottlenose dolphins (at 123). A separate study shows that honey bees exhibit more self-control (defined as the tendency to choose large delayed rewards over small immediate rewards) than rats and pigeons. Another recent study demonstrates that wasps are capable of transitive inference. Many of these animals display extraordinary social complexity. Some ants live in supercolonies containing millions of individuals. The effects of social deprivation can be drastic. Social isolation increases aggression in honeybees, reduces lifespan in ants, and destroys wasps’ ability to recognize faces. There is even evidence of affective complexity in order Hymenoptera. Agitated honey bees exhibit the same sorts of pessimistic cognitive biases that anxious mammals do.

None of this is to say that ants, wasps, and bees enjoy the same moral status as cows and pigs. It’s still plausible that insects are worth less, morally, than mammals. But the case that invertebrates are, as a rule, worth far less than vertebrates is quite weak. The invertebrate taxa for which we have the best evidence of sentience, coleoid cephalopods and arthropods, contain members capable of cognitive, social, and affective complexity that rivals many vertebrate species. And if these invertebrates can suffer, there is so far little reason to think that they suffer exponentially less than typical vertebrates.

Limiting Factors

When considering the scale of a problem, it is also prudent to examine the limiting factors that constrain the degree to which interventions in the near-term are able to tackle the problem. For example, there may be tremendous, easily preventable suffering occurring on a far-off planet in our galaxy. We might come to detect this suffering with next-generation astronomical equipment. But if the technology to actually reduce the far-off suffering (e.g., by sending spacecraft to the planet) is still centuries away, then for purposes of near-term interventions, the scale of the suffering is irrelevant. Of course, in the long-term, nearly all limiting factors are mutable, so the cost of eliminating a limiting factor is just one more input into a traditional cost-effectiveness analysis.21 Still, the general lesson should be recognizable: there’s no point buying bednets today if the ship to deliver them isn’t going to be ready for a hundred years.

This lesson is particularly relevant for invertebrate welfare. Take some extremely numerous group of potentially sentient invertebrates, say arthropods. Even assigning an egregiously low credence to the claim that arthropods have the capacity for morally significant valenced experience and adjusting downward for moral weight, one should still believe that the expected moral worth of the phylum Arthropoda is comparable to the expected moral worth of many classes of vertebrates.22 Taking a more reasonable (though still low) view of the possibility of morally significant arthropod experience results in an expected moral worth higher than any class of vertebrates. Does that necessarily mean that we should start pouring tens of millions of dollars into invertebrate welfare? No.

Put simply, there are a variety of limiting factors that will constrain the usefulness of large amounts of money in this area for many years to come. The biggest limiting factor is uncertainty. We don’t know enough about invertebrates to be able to judge with much confidence which invertebrates (if any) are sentient. We don’t know enough about the life history of most invertebrates to know which interventions would be most beneficial. Of the interventions that do seem potentially promising (e.g., humane insecticides) we don’t know enough about the flow-through effects to recommend the interventions at scale.23 Even if we were in a position to recommend large-scale interventions, we don’t know enough about public attitudes toward invertebrates to know whether or not advocating for invertebrate welfare would be net-negative for the effective altruism movement as a whole. And so on.

All this uncertainty translates to few actionable interventions other than additional investigation. Moreover, there are very few researchers and even fewer organizations explicitly thinking about invertebrate welfare.24 Thus, in the short-term, the invertebrate welfare cause area will only be able to absorb a modest amount of funding.25

Although in the long-term we should work to dismantle these limiting factors, in the near-term they may contribute to, rather than detract from, the case for invertebrate welfare. Because the capacity for invertebrate welfare to absorb additional funding is so limited, it’s possible to fully fund (for the near-term) a potentially important cause area for a relatively small sum of money. In that sense, invertebrate welfare is a steal.

The Evidence for Invertebrate Sentience

Some Caveats

First, the goal of this section is not to definitively convince anyone that invertebrates are sentient. The aim is merely to make the idea of invertebrate sentience somewhat plausible, where ‘somewhat plausible’ is compatible with credences well below 50%. Because there are so many invertebrates, even a relatively low credence in the proposition that invertebrates are sentient will, when coupled with a few basic assumptions discussed elsewhere, generate a high expected moral value.

Second, limited evidence of sentience should not be confused with limited sentience. The scientific literature on invertebrate sentience is young and incomplete and does not yet represent our best understanding of every group of invertebrates. More research is needed.

Third, this overview is intentionally cursory. A comprehensive overview is simply not feasible in a length suitable for the EA Forum. Whole books have been written about much smaller groups of animals. Due to space constraints, we have condensed the results of dozens of studies into a few pages worth of material. For a fuller account, see our summary of findings (part 1, part 2) or our Invertebrate Sentience Table.

Fourth, the type of evidence that one considers relevant to invertebrate sentience depends on the methodology one adopts for investigating the question. We endorse the methodology employed by philosopher Michael Tye in his 2016 book Tense Bees and Shell-Shocked Crabs. Tye uses an argument form called inference to the best explanation. The strategy is simple. Consider the behavior of some group of animals. If the best explanation for that behavior is that those animals are conscious, then, in the absence of defeaters, one is licensed to prefer the explanation that those animals are conscious to alternative explanations that don’t invoke consciousness.26 For more on inference to the best explanation, see our first and second posts.

Fifth, the evidence presented here is suggestive, not definitive. There are in fact many alternative explanations for the behaviors described below that do not invoke consciousness. We discuss these alternatives, as well as other potential defeaters, in our forthcoming counterarguments post.

Finally, due to the extraordinary diversity of invertebrates, there are few pieces of information that might qualify as evidence for invertebrate sentience in general.27 Nevertheless, it is perhaps worth noting that nearly all invertebrates withdraw from potentially harmful stimuli, and nociception, the neural process of encoding and processing potentially harmful stimuli, is widely conserved throughout the animal kingdom, including in invertebrates.28 Nearly all invertebrates are capable of some form of learning, with many invertebrates displaying long-term behavior alteration in response to harmful stimuli. Still, there is relatively little to be gained from investigating sentience at such a high level of generality. Better to analyze specific groups of invertebrates. We turn then to three case studies: coleoid cephalopods, decapod crustaceans, and eusocial insects.

Case Studies

Coleoid Cephalopods

Coleoid cephalopods encompass squid, cuttlefish, and octopuses. Coleoid cephalopods are soft-bodied, exclusively marine animals characterized by bilateral body symmetry, a prominent head, and eight or ten highly developed tentacles. Coleoid cephalopods are molluscs, and, like all molluscs, they evolved from creatures with shells. Unlike their extant sister group, the nautiloid cephalopods, which retain an outer shell for protection and to which they are only distantly related, coleoid cephalopods have either internalized or eliminated their shells. Coleoid cephalopods first appear in the fossil record a little more than 300 million years ago, although given the fact that they are soft-bodied, it’s possible they split from other cephalopods somewhat earlier. Coleoid cephalopods are divided into two major superorders, the ten-tentacle decapodiformes (cuttlefish and most squid) and the eight-tentacle octopodiformes (octopuses and vampire squid).29

Coleoid cephalopods are universally recognized as the most intelligent invertebrates. Octopuses have roughly five hundred million neurons, and although an astonishing 60% of these neurons are located in the tentacles, octopuses still have a recognizable central brain that processes and integrates exteroceptive and interoceptive sensory information.30 Octopuses are notoriously clever tool users. In addition to countless laboratory anecdotes, octopuses in the wild have been observed assembling coconuts into portable defensive shelters.31 Octopuses also have a keen sense of object permanence: they will make detours to get at prey seen through a transparent barrier, even if the detour takes them out of sight of their target.32

Cuttlefish deploy a wide variety of chromatic, textural, postural, and locomotor elements to communicate with predators, prey, and conspecifics, and there is evidence that they possess a sophisticated theory of mind.33 During courtship male cuttlefish engage in a remarkable form of tactical deception with rival males: simultaneous dual gender signalling. A male will position itself between a rival male and a potential mate. On the side of its mantle facing the potential mate, the deceptive male will produce typical chromatic courtship patterns. But on the side of its mantle facing the rival male, it will mimic typical female displays, thus confusing the rival and significantly reducing the odds that the rival male will attempt to disrupt copulation.

Jennifer Mather, a biologist and psychologist at the University of Lethbridge and a pioneer in cephalopod research, argues that coleoid cephalopods plausibly satisfy the ‘global workspace’ criterion on consciousness. According to this popular theory of consciousness, what’s required (and sufficient) for consciousness is the integrated representation of various sensory inputs competing for an organism’s attention. Cephalopods’ “rich discriminatory behavior” and “domain generality of learning” are, in Mather’s words, evidence of “primary consciousness.”

In 2005 the Scientific Panel on Animal Health and Welfare of the European Union concluded that cephalopods “have a nervous system and relatively complex brain similar to many vertebrates, and sufficient in structure and functioning for them to experience pain.”34 As a result of this conclusion, EU member states opted to give cephalopods used for scientific research the same legal protection that was previously afforded only to vertebrates (Directive 2010/63/EU).

Decapod Crustaceans

Decapod crustaceans encompass prawns, shrimp, crayfish, crabs, and lobsters. Decapod crustaceans are ten-footed arthropods characterized by the carapace extending from their thorax to their head.35 They can be found both on land and at sea in a wide range of habitats worldwide. Crustaceans are a commercially important group of animals. As noted above, hundreds of billions of crustaceans are slaughtered for food every year. (Most crustaceans consumed by humans are decapods.36) Decapod crustaceans have a long evolutionary history. Prawns first appear in the fossil record in the Triassic Period, shrimp and crabs first appear in the Jurassic, and lobsters first appear in the Cretaceous.

Decapod crustaceans engage in complex motivational tradeoffs that demonstrate the sort of behavioral plasticity one would expect from creatures with the capacity for valenced experience (and that one would not expect from creatures without the capacity for valenced experience). For example, in a laboratory setting, hermit crabs will abandon their shells if they are subjected to a mild shock. Initially, such behavior was thought to be purely reflexive. However, recent experiments show the crabs are significantly less likely to abandon their shells after shock if the odor of a predator is present. The fact that the crabs remain in their shells when the odor of a predator is present suggests that the behavior is not reflexive. A natural explanation is that the crabs weigh the pain of the shock against the fear of a predator, thus incorporating different interests and demands into a unified utility function. Such integration is often considered a distinctive attribute of consciousness.

Another example: shore crabs generally avoid well-lit areas, preferring to hide in dark environments such as those found under rocks. Given the choice between two chambers in a laboratory setting, one brightly lit and the other dark, the crabs will universally choose the dark chamber. However, if the dark chamber is rigged to deliver a mild shock, the crabs will begin to opt for the normally avoided well-lit chamber. The crabs do so in increasing numbers (and increasingly quickly) as the number of trials increases. A compelling explanation of this behavior is that the crabs feel pain, then learn to avoid the pain by choosing the opposite, otherwise undesirable chamber.

Decapod crustaceans also appear to lead complex emotional lives. For instance, stress-induced avoidance behavior in crayfish bears a striking resemblance to mammalian anxiety. A 2014 study demonstrated that shocked crayfish develop an extended, context-independent aversion to light.37 (The shocks were not associated with levels of illumination.) In contrast, unshocked crayfish, though preferring the dark, were happy to explore both illuminated and unilluminated areas of their environment. Most importantly, injecting the shocked crayfish with the anxiolytic drug chlordiazepoxide (used to treat anxiety in humans) eliminated the aversion to light.38 In humans, anxiety is often associated with danger that is perceived to be unavoidable39 or situations in which the threat is ambiguous or unknown.40 The electric shocks applied to the crayfish fit this description. In humans, anxiety is associated with generalized fear, that is, increased fear of unrelated stimuli. The shocked crayfish appeared to exhibit increased fear of light that is unrelated to the source of stress. In humans, anxiety is reduced by anxiolytic drugs. In crayfish, anxiety-like behavior is reduced by anxiolytic drugs. The most natural explanation of this phenomenon is that crayfish, like humans, are capable of experiencing negatively valenced emotional states.

Eusocial Insects

Eusocial insects are group-living animals characterized by reproductive division of labor, cooperative care of young, and overlapping adult generations. All ants and termites are eusocial, as are many bees, wasps, and aphids.41 Because the genetic fitness of sterile eusocial workers is determined by the success of the colony and because hives and colonies are capable of complex group decision-making, eusocial insects are sometimes thought to constitute superorganisms. Although eusociality is rare, eusocial insects are among the most successful animals on the planet, dominating many terrestrial habitats and often occupying cornerstone ecological niches. The comparison to human societies is inescapable. Eusocial insects typically live in densely populated, intricately tunneled hives or mounds. Colonies often employ specialized labor castes and depend on extensive social communication, including relatively complex forms of observational learning. The result is a remarkable array of multifaceted colony-level activity, such as aphid ‘farming’ and corpse management.

In addition to impressive colony-level behaviors, individual eusocial insects are surprisingly intelligent. Ants learn quickly and do not forget easily. Honey bees are able to communicate the distance, direction, and relative reward-to-danger ratio of nearby flower patches to hivemates using their famous waggle dance. Ants and bees are both capable of using tools flexibly. Funnel ants use artificial debris to transport liquid food to the nest, adopting different materials for different types of liquid in order to optimize handling and soaking properties. In another recent study researchers trained bumble bees to see that a ball could be used to dispense a reward. In subsequent iterations of the experiment, the bees independently learned to solve the task more efficiently by using a ball positioned more closely to the target, even though the ball was a different color.42 Both ants and bees engage in contextual learning. For instance, ants can be trained to associate the scent of an ant from another colony with a food reward and once trained will approach the odor in an appetitive context. However, the conditioned ants are not easily fooled. When a non-nestmate ant is present alongside the odor, the conditioned ants revert to their aggressive behavior.43 Similarly, honey bees can come to understand that when a certain color is presented to them, odor A predicts a sucrose reward and odor B does not, but when a different color is presented, the relationship between the odors is reversed.

There is even evidence of metacognitive abilities in eusocial insects. Using the “uncertain response” paradigm, one study tasked honey bees with discriminating between two stimuli. A correct answer earned a reward (sucrose) and an incorrect answer earned a punishment (quinine). When given the choice to opt out, honey bees were found to opt out more often when the trial was difficult (when the honey bee was proportionately more likely to receive a punishment than a reward). The honey bees on average improved their success-to-failure ratio when given the option to opt out of trials.44 When searching for a new nest site, individual ants adjust their behavior according to their relative confidence. They employ a “highly sophisticated ‘copy-when-uncertain’ social learning strategy similar to that observed in a few vertebrate species:” when their confidence is high, ants tend to act on their own initiative, but when they are uncertain, they tend to copy the actions of their nestmates.45 Another study found that ants upregulate pheromone trail deposition in response to changes in the location of food. In an experiment with a T-maze, researchers trained ants to a feeder location, then altered the environment by changing the feeder location to the other arm of the T-maze. After finding the new food source, ants upregulated pheromone deposition if they had made a wrong choice. Additionally, the researchers found that outgoing ants that went on to make an error deposited less pheromone. This seems to imply that the ants can measure the reliability of their own memories and respond accordingly by depositing more or less pheromones.46


EA Space

Invertebrate welfare is mostly ignored within the effective altruism community. Despite the fact that invertebrates comprise more than 99.9% of all animals, there is no single organization in the effective animal activism movement exclusively working to promote invertebrate well-being. In the effective altruism community there are only two organizations that devote a significant portion of their (modest) budgets to invertebrate welfare. Wild Animal Initiative is currently conducting a research project to investigate the feasibility of a humane insecticide program. The other organization is Rethink Priorities, the group that produced this report. There are also a handful of independent EA researchers promoting the cause area. Brian Tomasik has written extensively on invertebrate suffering. Max Carpendale, an independent researcher at the EA hotel and past contributor to Rethink Priorities’ work on invertebrate welfare, has also written about invertebrate sentience.

Several other organizations in EA space are aware of the potential magnitude of invertebrate welfare. Animal Ethics has written about animal sentience, including invertebrate sentience, as well as animal exploitation, including the exploitation of invertebrates. Foundational Research Institute and Effective Altruism Foundation are closely aligned with Tomasik’s views on invertebrate sentience. Both Animal Charity Evaluators and the Open Philanthropy Project have previously expressed interest in research that investigates the importance of invertebrate welfare. Faunalytics and Animal Equality have also occasionally written about invertebrates.

Based on the above, I estimate that the effective altruism community as a whole currently commits no less than $150,000 per year and no more than $300,000 per year to invertebrate welfare.47 Because there are no organizations exclusively working on invertebrate welfare, an exact estimate depends on the way the proportion of various research budgets is allocated to invertebrate welfare. Additionally, the exact number appears prone to pretty severe year-on-year fluctuations. Given certain developments, the number could plausibly fall to ~$0 in 2020. By way of comparison, Open Phil recommended nearly $28 million in animal welfare grants in 2018.48 If we assume that Open Phil accounts for one-half to two-thirds of total effective animal activism spending, that puts the percentage of money spent on invertebrates in effective animal activism between 0.27% and 0.71%.

Non-EA Space

Invertebrate welfare is also generally neglected outside the effective altruism community. Excluding EA organizations, there are perhaps fewer than half a dozen animal rights groups in the anglophone world that take invertebrate welfare seriously. PETA is one of them. PETA campaigns against eating crustaceans, wearing silk, owning hermit crabs, and killing household pest insects. Another important, albeit limited, organization is Crustacean Compassion. Crustacean Compassion is a UK-based charity campaigning to have crabs, lobsters and other decapods included in animal welfare legislation on the grounds that crustaceans feel pain. Among other goals, Crustacean Compassion hopes to replicate the live boil ban recently enacted in Switzerland. FishCount, a UK-based group promoting more humane commercial fishing, supports these goals. Viva!, another UK-based animal rights organization, also campaigns on behalf of lobsters. Organizations that promote strict veganism, including abstention from honey and silk, such as The Vegan Society, also take the interests of invertebrates into account. Many Dharmic religions, but especially Jainism, advise adherents to minimize harm to all living things, including invertebrates.

Conservation Groups

In addition to animal welfare and animal rights organizations, there are also dozens of conservation groups that target invertebrates. It is difficult to know how to assess these bodies. For one, it’s not clear that any of the organizations covered in this section believe that invertebrates possess intrinsic moral value or that they have the capacity for valenced experience. Most of the organizations emphasize the instrumental value that many invertebrates (e.g., insect pollinators) provide to ecosystem stability or the aesthetic value that certain iconic invertebrate species (e.g., monarch butterflies) provide to humans. Moreover, it’s unclear if the campaigns in which these organizations engage are net-positive for invertebrates as a whole. Nonetheless, these groups do possess various forms of expertise relating to invertebrates and could potentially make for good partners for advocacy, research, or on-the-ground interventions.

There are two general purpose invertebrate conservation groups in the anglophone world: the Xerces Society in North America and Buglife in Europe. However, there are dozens, perhaps as many as a hundred, conservation charities that specialize in particular groups of invertebrates. For example (and just to name a few), there are dragonfly charities: The Dragonfly Society of the Americas and the British Dragonfly Society. There are bee charities: The Honeybee Conservancy, Planet Bee Foundation, Pollinator Partnership, and the Bumblebee Conservation Trust. And there are butterfly charities: Save Our Monarchs, North American Butterfly Association, Butterfly Conservation, and the Lepidopterists’ Society.


Although it is difficult to estimate what percentage of academics conduct research on invertebrate welfare relative to other groups of animals, it is clear that invertebrate welfare is largely neglected in academia. That’s not to say that invertebrates themselves are neglected. Many model organisms are invertebrates. The fruit fly Drosophila melanogaster, the sea slug Aplysia californica, and the nematode Caenorhabditis elegans, are among the best understood animals on the planet. One reason these creatures are so widely studied is the belief that fruit flies, sea slugs, and nematodes are incapable of experiencing pain. Many experimental designs that would never be sanctioned for use on mammalian subjects are routinely conducted on invertebrates without ethical oversight.

A handful of academic researchers do take the question of invertebrate welfare seriously, and an even smaller group is actively investigating invertebrate sentience. Robert Elwood, professor emeritus in the school of biological sciences at Queen’s University Belfast, has been advancing the notion that crustaceans experience pain and stress for over a decade. Lynne Sneddon, director of bioveterinary science at the University of Liverpool, also promotes the idea that a wide variety of aquatic animals, including crustaceans, demonstrate the potential for pain perception. Anthony Rowe, an animal welfare officer at CSIRO, believes use of decapod crustaceans in research should require ethical review. Jennifer Mather, an expert in cephalopod cognition at the University of Lethbridge, has long argued that octopuses, squid, and cuttlefish appear to be conscious. Robyn J. Crook, an evolutionary biologist and behavioral neuroscientist at San Francisco State University, runs a lab studying sensation, emotion, and cognition in cephalopods, with an aim to provide empirical support for regulation governing the use of complex invertebrates in scientific research. Shelley Adamo, a neurobiologist at Dalhousie University, has reviewed the evidence for insect sentience and was asked to testify before a Canadian senate committee in 2003 as to whether invertebrates feel pain. Andrew Barron, a neuroethologist and ARC Future Fellow at Macquarie University, argues that insects have the capacity for subjective experience.49

Barriers to Attention

Human psychological heuristics conspire to make empathizing with invertebrates extremely difficult. Compared to household pets, farmed animals, and iconic wild species, most invertebrates are small and alien-looking. Many invertebrates are associated in the popular mind with disease or uncleanliness. Encounters with invertebrate often produce feelings of disgust, anxiety, or fear. These reactions can alter the contours of scientific investigation. Taxonomic bias is rampant in conservation biology and biodiversity research. Large, charismatic species are wildly over-represented in scientific studies, and this skew is driven largely by societal preferences. (Insects are particularly under-represented.) Even when invertebrates are not ignored, their welfare typically is. Our psychological heuristics explain some of this neglect, but it is also due to the nature of the subject. Invertebrate sentience research lies at the intersection of philosophy, biology, and neuroscience. As Max Carpendale puts it, “Many invertebrate biologists who might otherwise have a lot to contribute in the area are not philosophically inclined, and have not thought about the ethical implications of their knowledge, and so become confused about the question of invertebrate sentience.” With these barriers to attention in place, it is no surprise that invertebrate welfare is a highly neglected cause area.


How Should We Evaluate the Tractability of Invertebrate Welfare?

Improving invertebrate welfare will undoubtedly be difficult. How difficult? No one knows. Very few organizations and researchers have seriously investigated cost-effective interventions to improve the lives of invertebrates, and those that have seriously investigated the issue have been doing so for less than a decade, constrained by modest budgets and small staffs. Much more research is needed to ascertain the difficulty of improving invertebrate lives. Even if large-scale interventions are intractable, there may be low-hanging fruit that has been missed simply because the field is so neglected. Additional investigation may yet reveal that invertebrate welfare is an intractable cause area; however, we are not yet in a position to know this.

In this section we offer brief examples to illustrate the tractability of improving invertebrate welfare along two timeframes. The first subsection concerns the tractability of improving invertebrate welfare now. The second subsection concerns the tractability of improving our chances to locate cost-effective interventions in the medium-term. (These representative example interventions are not intended to be exhaustive. For more information on our recommended directions for future work, see our forthcoming “Next Steps” post.) Because the invertebrate welfare movement is in its infancy and supporting the cause area at this point simply means supporting additional research, only the second subsection is relevant to the importance-neglectedness-tractability framework. Nonetheless, for those readers with the strong pre-theoretic belief that effectively helping invertebrates would be practically impossible, the first subsection serves to show that this belief is probably mistaken.

Although the prospects of helping invertebrates at scale in the very near future are dim, there are several concrete, tractable steps we can take now to better understand invertebrate sentience and better position ourselves to help invertebrates in the future. The most important intervention at this stage is plausibly raising awareness about the cause area.50 The examples that follow may be most usefully evaluated according to that criterion.

Helping Invertebrates Now

Humane Insecticides

Wild Animal Initiative is currently conducting a research project to investigate the feasibility of a humane insecticide program. The goal of the project is to identify insecticides that kill faster, less painfully, or both, while minimizing downstream ecological effects and impacts on non-target species. According to their four month update, WAI has built a database of over 250 insecticides, documenting their modes of action, producers, and other relevant information. They have also contacted outside domain specialists to discuss the viability of the project. The hope is that relatively small financial incentives could motivate many users to adopt faster, less painful methods for controlling insect populations, thereby reducing a significant amount of invertebrate suffering.

Mesh to Reduce Bycatch in Sticky Traps

Sticky traps are adhesive-coated cards with an alluring odor and/or color deployed in large numbers by pest managers to monitor for flying insect pests in commercial orchards. They produce a significant amount of bycatch, particularly ladybugs and lacewings but also lizards and even some small birds. Research shows that simple, inexpensive mesh attached to the traps significantly reduces bycatch. Death by sticky trap is probably especially slow and unpleasant, so reducing the amount of bycatch could plausibly reduce a significant amount of invertebrate suffering (and also a small amount of vertebrate suffering).

Crustacean Live Boil Bans

It’s been illegal to boil crustaceans alive in New Zealand since 1999. Switzerland enacted a similar ban in 2018. Crustacean Compassion is actively campaigning for a similar ban (along with other protections) in the UK. If such a ban could be replicated across the whole of the EU, it would spare billions of animals from what appears to be an extremely painful death.

Campaigns Against Carmine

Carmine, also known as cochineal extract, is a red dye widely used in yogurts, ice creams, sodas, cupcakes, donuts, and lipstick. It is made from the female cochineal insect. According to Wikipedia, it takes roughly 80,000 insects to make a kilogram, and more than 220 tons of the stuff is produced worldwide, yielding an estimate of roughly 16 billion insects killed per year.51 The insects seem to be either boiled or baked alive. The general public appears not to know that carmine is widely used in food products and cosmetics. The general public also appears to harbor aversive attitudes towards both consuming ground-up bug paste and smearing it on faces. For those reasons, public awareness campaigns against carmine might be quite effective in reducing its use. It has worked at least once in the past: on March 14, 2012, reported that Starbucks uses carmine in its strawberry Frappuccinos. An avalanche of negative press followed. On April 19, a mere 36 days later, the chain announced it would stop using carmine in its food and beverage products.

Thwarting Entomophagy

Entomophagy refers to the practice of eating insects, especially by humans. Although people have been eating insects for most of human history and many cultures continuing to eat insects, the practice has not yet been adopted on a large scale in the industrialized world. If certain futurists are to be believed, that is set to change. Interest in entomophagy appears to be surging. In 2013, the FAO released a comprehensive report on edible insects. Entomophagy increasingly is marketed as a solution to food insecurity and a remedy to global warming. (A quick Google search reveals headlines like “Entomophagy: How giving up meat and eating bugs can help save the planet.”) As of last year, when the relevant regulations came into force, edible insects can be found in shops and supermarkets across Europe. However, if insects are sentient, there are some pretty serious welfare concerns at stake. So we seem to be at a key moment in history regarding this issue. Sustained pressure may be able to derail the entomophagy movement now, thus sparing trillions of insects unnecessary suffering. If the movement is allowed to gain even more momentum, it may be significantly harder to kill later.

Helping Invertebrates Later

Understanding Invertebrate Life History

Before we can hope to improve the lives of invertebrates, we must first know what those lives are like. Unfortunately, we know precious little about the lives of most invertebrates. As noted above, pervasive taxonomic bias in the life sciences ensures that invertebrates are systematically understudied. Even in the absence of taxonomic bias, there are many more invertebrate species than vertebrate species, and they are in many respects comparatively more difficult to study, so it’s no surprise we know so little. And among the invertebrates that have been thoroughly studied, hardly any have been studied with their welfare in mind. Thankfully, some EA organizations are trying to close this knowledge gap. For instance, Rethink Priorities recently released a life history report (written by a tenured professor of ecology) on herbivorous insects. The report documents various data regarding lifespans, fecundity, mortality rates, and mortality causes. These data will help us better understand which interventions might most effectively improve the welfare of herbivorous insects. Similar reports should be commissioned for other large groups of invertebrates.

Polling Public Attitudes

In the long-term improving invertebrate welfare may well depend on actively changing attitudes towards invertebrates, especially among effective altruists, animal activists, academics, and, ultimately, among legislators and lobbyists. But an advocacy campaign cannot operate in an information vacuum. One tangible and tractable next step is to poll public attitudes about invertebrates. This polling should not only investigate current attitudes towards invertebrates but also how malleable those attitudes are. So, for example, an experiment might compare differences in assigned probability of consciousness or moral weight in: i) people asked about a given taxon (e.g., ladybugs), ii) people asked about “an animal” that has a list of capacities, iii) people asked about ladybugs and told they have certain capacities. With modest funding, Rethink Priorities is in a position to conduct such surveys.

Compiling Extant Scientific Research on More Species

Although we have already compiled the extant scientific research relevant to invertebrate sentience for 18 biological taxa, our efforts could have been more comprehensive. There are at least a dozen more taxa that we would like to see added to this database. To strengthen analogical and similarity-driven arguments, it would be helpful to have a more complete understanding of vertebrate sentience. Hence we would like to add a representative amphibian (class Amphibia), a representative reptile (class Reptilia), a representative bony fish (superclass Osteichthyes), and a representative cartilaginous fish (class Chondrichthyes) to the database. On the invertebrate side, there are several additions that could potentially be helpful. Phylum Mollusca contains coleoid cephalopods (represented by family Octopodidae on our table), and of all invertebrates, we are most confident that these creatures are conscious. But Mollusca also contains creatures like oysters and mussels, and there is a strong initial case to be made that these animals are not conscious. Thus, it would be good to investigate class Bivalvia (of which oysters and mussels are members) to better understand the distribution of sentience within Mollusca.52 Staying within the phylum, snails are consumed by humans in many cultures53 and have attracted some attention as an edge case of consciousness in philosophical circles. A representative from class Gastropoda would therefore be useful. And if some interventions are going to target human consumption of invertebrates, more widely consumed invertebrates, like shrimp (suborder Dendrobranchiata), should also be added to the table. Other invertebrates exploited by humans include silkworms (genus Bombyx), cochineal (genus Dactylopius), mealworms (genus Tenebrio), house crickets (genus Acheta), and lac scales (family Kerriidae). Before campaigning against the exploitation of these animals, we should have a firm grasp of the evidence that they are sentient. Finally, as noted above, the Antarctic krill (genus Euphausia) is among the most numerous individual species on the planet. Any animal species numbering in the hundreds of trillions is worth understanding better.

Although much of the research we compile can be used by animal advocates, they are not the only target audience. Another goal of our project is to synthesize and disseminate information back to invertebrate researchers. A well-placed review article or conference presentation could spark discussions that ultimately help motivate researchers to work more directly on this topic.

Better Population Estimates

If possible, it could be extremely useful to fund additional research to better estimate invertebrate populations. For example, the most widely cited global insect population estimate is nearly 60 years old. It’s an extrapolation based on the number of insects found on the Rothamsted Farm Site in the United Kingdom before the widespread use of pesticides. Without more studies (and, where feasible, a more rigorous methodology), our estimates are likely to be off, in one direction or the other, by several orders of magnitude. There is an almost incomprehensible difference between there being 100 quadrillion insects and there being 100 quintillion insects.

New Scientific Research

Academic outreach is likely to be a high-value activity in the near-term. Additional scientific research will help us better understand invertebrate sentience. For instance, the effects of analgesics (pain-killing drugs), anxiolytics (anti-anxiety drugs), and various recreational drugs have been well-documented in many invertebrates. However, it’s difficult to judge whether the behavioral changes these drugs induce are merely a product of physiological responses or instead reflect a change in the valence of experience. Self-administration studies can help us tease the two apart. So, for example, if a purportedly stressed animal consistently favored a food source laced with an anxiolytic, that would be mild evidence that the stress causes a negatively valenced emotional state for which the animal seeks relief. To our knowledge, there has only been a single analgesic self-administration study on an invertebrate species and no anxiolytic or antidepressant self-administration studies on invertebrate species.

Preliminary Conclusion

Our approach to invertebrate sentience and invertebrate welfare is essentially comparative. The effective animal activism community already devotes considerable resources to helping mammals, birds, and (to a lesser extent) fish.54 Animal advocates think that mammals, birds, and fish are sentient on the basis of well-established behavioral and neurobiological facts. There is no reason in principle why this approach cannot be extended to invertebrates. Although the underlying science is more uncertain, the same sort of behavioral and neurobiological evidence that leads us to attribute conscious states to mammals, birds, and fish is also available for some invertebrates, notably cephalopods and many arthropods. So if one thinks mammals, birds, and fish are conscious, one should take the idea of invertebrate consciousness seriously. Moreover, there are far more arthropods (to take just one invertebrate phylum) than there are mammals, birds, and fish. So even if the case for arthropod consciousness is weaker than the case for mammal, bird, and fish consciousness, the expected value of helping arthropods might be higher—potentially much higher—than the expected value of helping mammals, birds, and fish.

Of course, tremendous uncertainty remains. As our scientific and philosophical understanding of consciousness continues to improve, we may discover definitive reasons to think consciousness is restricted to vertebrates. Even if we think that invertebrates are conscious, we might come to justifiably believe that the moral value of their experiences is negligible. And even if invertebrates do have morally valuable experiences, it might turn out that there is no cost-effective way to help them. But we aren’t yet in a position to know any of these things. We don’t know enough about consciousness to be certain that arthropods and cephalopods aren’t conscious. We don’t know enough about normative ethics to be certain we don’t have moral obligations to these creatures. We don’t know enough about the tractability of improving invertebrate welfare to be certain we can’t help these animals. And because there are a truly mind-boggling number of invertebrates, the cause area has the potential to be extremely high-value. Thus, we think that at this stage invertebrate welfare is a cause area that ought to be prioritized.


Rethink Priorities Logo

This essay is a project of Rethink Priorities. It was written by Jason Schukraft. Thanks to Marcus A. Davis, Jamie Gittins, Michelle Graham, Kieran Greig, Sam Fox Krauss, Peter Hurford, David Moss, Abraham Rowe, Jacob Schmiess, Gavin Taylor, Daniela R. Waldhorn, and Rachael Woodard for helpful feedback. If you like our work, please consider subscribing to our newsletter. You can see all our work to date here.

Works Cited

Abbott, K. R., & Dukas, R. (2009). Honeybees consider flower danger in their waggle dance. Animal Behaviour, 78(3), 633-635.

Atkinson, A., Siegel, V., Pakhomov, E. A., Jessopp, M. J., & Loeb, V. (2009). A re-appraisal of the total biomass and annual production of Antarctic krill. Deep Sea Research Part I: Oceanographic Research Papers, 56(5), 727-740.

Baars, B. J. (1988). A Cognitive Theory of Consciousness. Cambridge: Cambridge University Press.

Bar-On, Y. M., Phillips, R., & Milo, R. (2018). The biomass distribution on Earth. Proceedings of the National Academy of Sciences, 115(25), 6506-6511.

Bateson, M., Desire, S., Gartside, S. E., & Wright, G. A. (2011). Agitated honeybees exhibit pessimistic cognitive biases. Current biology, 21(12), 1070-1073.

Belzung, C., & Philippot, P. (2007). Anxiety from a phylogenetic perspective: is there a qualitative difference between human and animal anxiety?. Neural plasticity, 2007.

Bos, N., Guerrieri, F. J., & d’Ettorre, P. (2010). Significance of chemical recognition cues is context dependent in ants. Animal Behaviour, 80(5), 839-844.

Boulay, R., Quagebeur, M., Godzinska, E. J., & Lenoir, A. (1999). Social isolation in ants: evidence of its impact on survivorship and behavior in Camponotus fellah (Hymenoptera: Formicidae). Sociobiology, 33(2), 111-124.

Breed, M. D. (1983). Correlations between aggressiveness and corpora allata volume, social isolation, age and dietary protein in worker honeybees. Insectes Sociaux, 30(4), 482-495.

Brown, C., Garwood, M. P., & Williamson, J. E. (2012). It pays to cheat: tactical deception in a cephalopod social signalling system. Biology letters, 8(5), 729-732.

Chapman, A. D., & Chapman, A. D. (2009). Numbers of living species in Australia and the world.

Cheng, K., Peña, J., Porter, M. A., & Irwin, J. D. (2002). Self-control in honeybees. Psychonomic bulletin & review, 9(2), 259-263.

Chittka, L., & Niven, J. (2009). Are bigger brains better?. Current Biology, 19(21), R995-R1008.

Czaczkes, T. J., & Heinze, J. (2015). Ants adjust their pheromone deposition to a changing environment and their probability of making errors. Proceedings of the Royal Society B: Biological Sciences, 282(1810), 20150679.

DeGrazia, D. (2008). Moral status as a matter of degree?. The Southern Journal of Philosophy, 46(2), 181-198.

Finn, J. K., Tregenza, T., & Norman, M. D. (2009). Defensive tool use in a coconut-carrying octopus. Current Biology, 19(23), R1069-R1070.

Fossat, P., Bacqué-Cazenave, J., De Deurwaerdère, P., Delbecque, J. P., & Cattaert, D. (2014). Anxiety-like behavior in crayfish is controlled by serotonin. Science, 344(6189), 1293-1297.

Gutnick, T., Byrne, R. A., Hochner, B., & Kuba, M. (2011). Octopus vulgaris uses visual information to determine the location of its arm. Current biology, 21(6), 460-462.

Healy, K., McNally, L., Ruxton, G. D., Cooper, N., & Jackson, A. L. (2013). Metabolic rate and body size are linked with perception of temporal information. Animal Behaviour, 86(4), 685-696.

Hochner, B., Shomrat, T., & Fiorito, G. (2006). The octopus: a model for a comparative analysis of the evolution of learning and memory mechanisms. The Biological Bulletin, 210(3), 308-317.

Kellert, S. R. (1993). Values and perceptions of invertebrates. Conservation biology, 7(4), 845-855.

Loukola, O. J., Perry, C. J., Coscos, L., & Chittka, L. (2017). Bumblebees show cognitive flexibility by improving on an observed complex behavior. Science, 355(6327), 833-836.

Maák, I., Lőrinczi, G., Le Quinquis, P., Módra, G., Bovet, D., Call, J., & d'Ettorre, P. (2017). Tool selection during foraging in two species of funnel ants. Animal behaviour, 123, 207-216.

Magee, B., & Elwood, R. W. (2013). Shock avoidance by discrimination learning in the shore crab (Carcinus maenas) is consistent with a key criterion for pain. Journal of Experimental Biology, 216(3), 353-358.

Magee, B., & Elwood, R. W. (2016). Trade-offs between predator avoidance and electric shock avoidance in hermit crabs demonstrate a non-reflexive response to noxious stimuli consistent with prediction of pain. Behavioural processes, 130, 31-35.

Mather, J. A. (2008). Cephalopod consciousness: behavioural evidence. Consciousness and cognition, 17(1), 37-48.

Mota, T., Giurfa, M., & Sandoz, J. C. (2011). Color modulates olfactory learning in honeybees by an occasion-setting mechanism. Learning & Memory, 18(3), 144-155.

Öhman, A. (2000). Fear and Anxiety: Evolutionary, Cognitive, and Clinical perspectives. in M. Lewis & J.M. Haviland-Jones (eds.). Handbook of Emotions. pp. 573–93. New York: The Guilford Press.

Oizumi, M., Albantakis, L., & Tononi, G. (2014). From the phenomenology to the mechanisms of consciousness: integrated information theory 3.0. PLoS computational biology, 10(5), e1003588.

Perry, C. J., Barron, A. B., & Cheng, K. (2013). Invertebrate learning and cognition: relating phenomena to neural substrate. Wiley Interdisciplinary Reviews: Cognitive Science, 4(5), 561-582.

Perry, C. J., & Barron, A. B. (2013). Honey bees selectively avoid difficult choices. Proceedings of the National Academy of Sciences, 110(47), 19155-19159.

Piqueret, B., Sandoz, J. C., & d'Ettorre, P. (2019). Ants learn fast and do not forget: associative olfactory learning, memory and extinction in Formica fusca. Royal Society Open Science, 6(6), 190778.

Plowes, N. (2010). An introduction to eusociality. Nature Education Knowledge, 3(10), 7.

Rodhouse, P. G., & Nigmatullin, C. M. (1996). Role as consumers. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences, 351(1343), 1003-1022.

Rosenthal, M. F., Gertler, M., Hamilton, A. D., Prasad, S., & Andrade, M. C. (2017). Taxonomic bias in animal behaviour publications. Animal Behaviour, 127, 83-89.

Rowe, A. (2018). Should Scientific Research Involving Decapod Crustaceans Require Ethical Review?. Journal of Agricultural and Environmental Ethics, 31(5), 625-634.

Schultz, T. R. (2000). In search of ant ancestors. Proceedings of the National Academy of Sciences, 97(26), 14028-14029.

Smith, E. S. J., & Lewin, G. R. (2009). Nociceptors: a phylogenetic view. Journal of Comparative Physiology A, 195(12), 1089-1106.

Sneddon, L. U. (2017). Comparative physiology of nociception and pain. Physiology, 33(1), 63-73.

Stroeymeyt, N., Giurfa, M., & Franks, N. R. (2017). Information certainty determines social and private information use in ants. Scientific reports, 7, 43607.

Sunamura, E., Espadaler, X., Sakamoto, H., Suzuki, S., Terayama, M., & Tatsuki, S. (2009). Intercontinental union of Argentine ants: behavioral relationships among introduced populations in Europe, North America, and Asia. Insectes Sociaux, 56(2), 143-147.

Thomas, A., & MacDonald, C. (2016). Investigating body patterning in aquarium-raised flamboyant cuttlefish (Metasepia pfefferi). PeerJ, 4, e2035.

Tibbetts, E. A., Agudelo, J., Pandit, S., & Riojas, J. (2019). Transitive inference in Polistes paper wasps. Biology letters, 15(5), 20190015.

Tibbetts, E. A., Desjardins, E., Kou, N., & Wellman, L. (2019). Social isolation prevents the development of individual face recognition in paper wasps. Animal Behaviour, 152, 71-77.

Troudet, J., Grandcolas, P., Blin, A., Vignes-Lebbe, R., & Legendre, F. (2017). Taxonomic bias in biodiversity data and societal preferences. Scientific Reports, 7(1), 9132.

Wells, M. J. (1964). Detour experiments with octopuses. Journal of Experimental Biology, 41(3), 621-642.

Williams, C. B. (1960). The range and pattern of insect abundance. The American Naturalist, 94(875), 137-151.

Young, R. E., Vecchione, M., & Donovan, D. T. (1998). The evolution of coleoid cephalopods and their present biodiversity and ecology. African Journal of Marine Science, 20.


  1. Vertebrates constitute a subphylum in the phylum Chordata. Cladistically, it would be more precise to speak of ‘chordates’ and ‘non-chordates.’ In using the terms ‘vertebrates’ and ‘invertebrates’ we defer to common usage. However the number of invertebrates in the phylum Chordata is trivial compared to the number of invertebrates outside Chordata, so common usage is not wholly inaccurate. 

  2. We use the terms ‘sentience,’ ‘phenomenal consciousness,’ and ‘subjective experience’ interchangeably. An organism is sentient just in case there is something it is like to be that organism. ‘Valenced experience’ denotes a proper subset of conscious experience in which experiences take on a positive or negative affect. All creatures with the capacity for valenced experience are necessarily sentient, but not all sentient creatures necessarily have the capacity for valenced experience. 

  3. ‘Invertebrate sentience’ is similarly misleading for the same reason. Additionally, some species of invertebrates are much more studied than others. Thus, given the current scientific evidence, we have different degrees of confidence about which of them are sentient. It isn’t appropriate to generalize based on evidence that is only valid for specific groups of invertebrates. 

  4. Some of the interventions would target invertebrates subjected to human exploitation, while other interventions might target invertebrates suffering due to natural causes. 

  5. Because invertebrate welfare is not yet a mature field, it’s impossible to know in advance which interventions will be high-value. These examples are used for illustrative purposes only. 

  6. See Table S1 in the Supplementary Information Appendix of Bar-On, Y. M., Phillips, R., & Milo, R. (2018). The biomass distribution on Earth. Proceedings of the National Academy of Sciences, 115(25), 6506-6511. 

  7. Axel Rossberg of Queen Mary’s University estimates that at least 600 quintillion nematodes are born every day

  8. Williams, C. B. (1960). The range and pattern of insect abundance. The American Naturalist, 94(875), 137-151. This study is old, and its methodology fairly simplistic. Nonetheless, it is the most commonly cited global insect population estimate in the literature. Given the way our understanding of taxonomy has improved since this study was conducted, this figure might better represent the total number of terrestrial arthropods. 

  9. Atkinson, A., Siegel, V., Pakhomov, E. A., Jessopp, M. J., & Loeb, V. (2009). A re-appraisal of the total biomass and annual production of Antarctic krill. Deep Sea Research Part I: Oceanographic Research Papers, 56(5), 727-740. This figure appears to be for postlarval krill. If most krill die young, this figure could be a big underestimate. 

  10. The number of wild-caught crustaceans is potentially even higher. According to FAO data, more than six million metric tons of crustaceans were captured in 2017. 

  11. See page 44 of the appendix of Bar-On, Y. M., Phillips, R., & Milo, R. (2018). The biomass distribution on Earth. Proceedings of the National Academy of Sciences, 115(25), 6506-6511. 

  12. See Figure 1 in Bar-On, Y. M., Phillips, R., & Milo, R. (2018). The biomass distribution on Earth. Proceedings of the National Academy of Sciences, 115(25), 6506-6511. 

  13. In this section I focus solely on healthy adults for all species. This simplification ignores the moral significance of the potential for developing morally significant properties. It also ignores the moral significance (if there is any) of belonging to a group whose average or normal members have a certain moral status. 

  14. Only dreamless sleep lacks a phenomenology so only dreamless sleep counts as unconscious according to our definition. 

  15. See, inter alia, Baars, B. J. (1988). A Cognitive Theory of Consciousness. Cambridge: Cambridge University Press. 

  16. For example, honey bees are capable of multimodal associative learning. They can recognize that when one color is presented to them, odor A predicts a sucrose reward and odor B does not, but when a different color is presented, odor B predicts a sucrose reward and odor A does not. This finding reflects the fact that when foraging, a single modality by itself can’t parse beneficial flower species from non-beneficial species. Bees rely on a combination of color, shape, and odor to distinguish good flowers from useless ones. Mota, T., Giurfa, M., & Sandoz, J. C. (2011). Color modulates olfactory learning in honeybees by an occasion-setting mechanism. Learning & Memory, 18(3), 144-155. 

  17. Alternatively, it might be the the statistical regularity of the pattern rather than the phenomenal intensity of the pattern that would be assisted by cognitive sophistication. Thanks to Gavin Taylor for this point. 

  18. Even more so than the rest of this post, the foregoing paragraph is highly speculative. The aim of the argument therein is simply to push back against the intuition that phenomenal intensity declines as cognitive sophistication declines. 

  19. DeGrazia, D. (2008). Moral status as a matter of degree?. The Southern Journal of Philosophy, 46(2), 181-198. He adds, “Thus persons have the highest moral status, Great Apes and dolphins a bit less, elephants and monkeys somewhat less than apes and dolphins, middling mammals still less, rodents less, and so on down through the phylogenetic scale (to the extent that it tracks complexity of the relevant sorts) from birds to reptiles to amphibians to any other animals who are sentient.” 

  20. DeGrazia 2008: 193. 

  21. So, for instance, the scale of far-off suffering certainly gives us a reason to speed up research into interstellar travel. 

  22. There are, conservatively, 10 quintillion arthropods, give or take an order of magnitude. Say you only have a 1% credence that arthropods have morally relevant experiences. That knocks you down to 100 quadrillion expected arthropods. Say you think moral weight is determined by neuron count. Take a conservative average neuron count for arthropods (100,000) and use 100 million (putting you roughly in the range of rodents) as an example of a morally valuable vertebrate. That means, even with 1% credence and adjusting downward for moral weight, arthropods are worth roughly 100 trillion morally valuable vertebrates. (For comparison, there are probably no more than a trillion birds.) 

  23. Other intervention targets are less ambiguous. For instance, we can say with reasonable certainty that lobsters should not be boiled alive. 

  24. This is not to say that there aren’t neuroscientists, ecologists, ethologists and the like doing research that is relevant to invertebrate sentience and welfare. (See our Invertebrate Sentience Table for a collation of such knowledge.) The point is that these scientists aren’t framing their research in terms of figuring out whether invertebrates are sentient and if so how to improve their welfare. 

  25. On the academic side of things, there is potentially more ability to absorb funding, if we can get the right people on board. The infrastructure for scientific research on questions relevant to invertebrate sentience is already largely in place. To utilize this infrastructure we need to build interest in the topic, identify promising projects, and coordinate the relevant researchers. 

  26. It’s important to note that one can prefer an explanation without fully believing the explanation. If there are numerous plausible explanations, the best explanation might only warrant a credence of 20%. For example, it’s consistent to have a fairly low credence in the claim that invertebrates feel pain and yet think that that explanation of their behavior is more likely than any other explanation of their behavior. 

  27. Evidence for panpsychism would trivially qualify as evidence for invertebrate sentience. Similarly, evidence for the view that all animals or all living things are conscious would also qualify. 

  28. Smith, E. S. J., & Lewin, G. R. (2009). Nociceptors: a phylogenetic view. Journal of Comparative Physiology A, 195(12), 1089-1106. Nociceptors, the specialized peripheral sensory cells that detect potentially harmful stimuli, have been identified in fruit flies, sea slugs, and the nematode C. elegans. They are probably ubiquitous. See Sneddon, L. U. (2017). Comparative physiology of nociception and pain. Physiology, 33(1), 63-73. 

  29. Young, R. E., Vecchione, M., & Donovan, D. T. (1998). The evolution of coleoid cephalopods and their present biodiversity and ecology. African Journal of Marine Science, 20

  30. Hochner, B., Shomrat, T., & Fiorito, G. (2006). The octopus: a model for a comparative analysis of the evolution of learning and memory mechanisms. The Biological Bulletin, 210(3), 308-317. 

  31. Finn, J. K., Tregenza, T., & Norman, M. D. (2009). Defensive tool use in a coconut-carrying octopus. Current Biology, 19(23), R1069-R1070. Importantly, the only known source of these clean and lightweight shells is the coastal human communities, and thus the octopuses have not interacted with these items on an evolutionary timescale. 

  32. Wells, M. J. (1964). Detour experiments with octopuses. Journal of Experimental Biology, 41(3), 621-642. See also Gutnick, T., Byrne, R. A., Hochner, B., & Kuba, M. (2011). Octopus vulgaris uses visual information to determine the location of its arm. Current biology, 21(6), 460-462. 

  33. Thomas, A., & MacDonald, C. (2016). Investigating body patterning in aquarium-raised flamboyant cuttlefish (Metasepia pfefferi). PeerJ, 4, e2035. 

  34. European Food Safety Authority (EFSA). (2005). Opinion of the Scientific Panel on Animal Health and Welfare (AHAW) on a request from the Commission related to the aspects of the biology and welfare of animals used for experimental and other scientific purposes. EFSA Journal, 3(12), 292. 

  35. Rowe, A. (2018). Should Scientific Research Involving Decapod Crustaceans Require Ethical Review?. Journal of Agricultural and Environmental Ethics, 31(5), 625-634. 

  36. According to FAO data, approximately 90% of wild-caught crustaceans in 2017 were decapods. 

  37. Fossat, P., Bacqué-Cazenave, J., De Deurwaerdère, P., Delbecque, J. P., & Cattaert, D. (2014). Anxiety-like behavior in crayfish is controlled by serotonin. Science, 344(6189), 1293-1297. Note that the popular introduction to the article mistakenly states that the crayfish were afraid of the dark, not the light. A corrected title for the popular introduction is available here

  38. Separately, the injection of serotonin in unshocked crayfish induced light-aversion behavior that was also eliminated by chlordiazepoxide. In humans, elevated levels of serotonin are associated with anxiety. 

  39. Öhman, A. (2000). Fear and Anxiety: Evolutionary, Cognitive, and Clinical perspectives. in M. Lewis & J.M. Haviland-Jones (eds.). Handbook of Emotions. pp. 573–93. New York: The Guilford Press. 

  40. Belzung, C., & Philippot, P. (2007). Anxiety from a phylogenetic perspective: is there a qualitative difference between human and animal anxiety?. Neural plasticity, 2007

  41. Plowes, N. (2010). An introduction to eusociality. Nature Education Knowledge, 3(10), 7. 

  42. Loukola, O. J., Perry, C. J., Coscos, L., & Chittka, L. (2017). Bumblebees show cognitive flexibility by improving on an observed complex behavior. Science, 355(6327), 833-836. See here for an inexplicably adorable demonstration. 

  43. Bos, N., Guerrieri, F. J., & d’Ettorre, P. (2010). Significance of chemical recognition cues is context dependent in ants. Animal Behaviour, 80(5), 839-844. 

  44. Perry, C. J., & Barron, A. B. (2013). Honey bees selectively avoid difficult choices. Proceedings of the National Academy of Sciences, 110(47), 19155-19159. 

  45. Stroeymeyt, N., Giurfa, M., & Franks, N. R. (2017). Information certainty determines social and private information use in ants. Scientific reports, 7, 43607. 

  46. Czaczkes, T. J., & Heinze, J. (2015). Ants adjust their pheromone deposition to a changing environment and their probability of making errors. Proceedings of the Royal Society B: Biological Sciences, 282(1810), 20150679. It should be noted that the authors are uncertain about what their findings represent. They write that “it's hard to believe that such tiny-brained animals are capable of such an advanced cognitive feat” and “one could conceive of several alternative explanations for our findings, which do not invoke metacognition.” At the same time, they argue that their findings, “alongside similar results from honeybees (Perry & Barron, 2013), are suggestive of metacognitive abilities in social insects.” 

  47. Rethink Priorities accounts for roughly $70,000 of this figure. Wild Animal Initiative projects $353,710 in 2019 expenses, of which roughly $85,000 will go towards invertebrate specific research and $50,000 to $70,000 will go towards things potentially relevant to future invertebrate research. 

  48. As another comparison, in the 2019 first quarter grant round, CEA’s Animal Welfare Fund disbursed $445,000, CEA’s Meta Fund disbursed $512,000, CEA’s Long-Term Future Fund disbursed $923,150, and CEA’s Global Health and Development Fund disbursed $1,705,000

  49. There’s also a smattering of philosophers investigating invertebrate sentience. Colin Klein at ANU co-authored the Barron paper arguing that insects have the capacity for subjective experience. Michael Tye at UT Austin argues that we are licensed to prefer explanations of insect and crustacean behavior that invoke phenomenal states to explanations that do not. Peter Godfrey-Smith at the University of Sydney has long maintained that cephalopods are conscious. In most cases, philosophical investigations into invertebrate sentience depend heavily on empirical work conducted by researchers of the sort cited in the main text. 

  50. Of course, it’s important to proceed cautiously with any outreach campaigns. A poorly planned or executed campaign could backfire and lead not only to reduced support for invertebrate welfare but reduced support for effective altruism as a whole. There is also the worry that rushing into an advocacy campaign could create hard-to-reverse lock-in effects. If the initial message is suboptimal, these lock-in effects could impose substantial costs. 

  51. Brian Tomasik estimates a much higher figure: 120 billion. 

  52. Not all bivalves are as dull as oysters and mussels. Clams and scallops have eyes and move around a bit more. Thanks to Gavin Taylor for this clarification. 

  53. Global snail production amounted to 43,000 metric tons in 2016. 

  54. The vast majority of these resources go to helping farmed mammals, birds, and fish. However, the wild animal welfare crowd also targets these groups of animals.