Recently, Rethink Priorities (a project of Rethink Charity) did an impressive review of our state of knowledge about invertebrate sentience. Considering the fact that 99,9998% of all animals are invertebrates[1], it is crucial to know whether they are sentient and can have positive and negative experiences.
Sentience implies two properties: having a consciousness (subjective experiences) and having a utility function (an internal goal function or reward system). A plant, a thermostat, an artificial intelligent machine or a robot are examples of things with a utility function but without consciousness. On the other hand, some experiences such as seeing a white wall or touching a table are conscious but neutral (no utility function): there is no reward or objective in seeing, touching, looking for or avoiding these things. Combining a consciousness with a utility function, we get valenced (positive or negative) mental states such as pain and pleasure.
Most of the invertebrates are very small roundworms and ringed worms, and it is not clear whether they are sentient. But as the Rethink Priority review shows, there is increasing evidence that arthropods are sentient and have valenced experiences. Arthropods are animals with an exoskeleton and a segmented body, such as insects, spiders and crustaceans. After reading their review, I have updated my personal guess about the probability of insects (in particular bees, fruit flies and ants) having valenced experiences to more than 50%.
Even if the probability of arthropods being sentient is considered much lower (e.g. less than 10%), the precautionary principle should be applied, because there are huge numbers of arthropods. At this moment, there are about 1018 terrestrial arthropods (mostly ants) and 1020 marine arthropods (mostly very small copepods as zooplankton). This can be compared with about 1010 humans. So if arthropods happen to be sentient and we erroneously believe they are not, we are neglecting huge amounts of welfare and suffering. When a lot is at stake, the precautionary principle is reasonable.
In this article I first present experimental results that are in my opinion the most convincing and amazing facts to demonstrate that insects (in particular bees, ants and flies) are sentient, i.e. they have a consciousness and a utility function. Next, I discuss the issue how to compare insect welfare with human welfare. Finally, I explore how we can avoid wild insect suffering (by bee protection) and cultured insect suffering (by avoiding the use of insects for food).
For more on insect suffering, see the work by Brian Tomasik here, here, here and here, and Wild Animal Initiative here and here.
Indications for a consciousness in insects
Although it is impossible to determine for sure whether an animal is conscious, clever experiments can indicate the presence of consciousness. We can use experiments to test for skills or behavioral responses that are never performed by humans who are believed to be in unconscious states, or that sometimes occur unconsciously in humans. For example there are unconscious learning processes[2], so learning does not necessarily require consciousness. Nociception is the ability to detect harmful stimuli, such as burning your finger. But pulling away your hand after touching a hot stove is an unconscious reflex. After a brief moment, you start to feel a burning pain in your finger. Pain is different from nociception, because pain involves a conscious state whereas nociception is an unconscious perception.
Integrated behavioral control system in the brain
Like the midbrain in vertebrates, insects also have a brain structure that integrates perceptual inputs (e.g. vision and touch) to create a neural simulation of the body position of an insect in space that allows for behavioral control. In other words: an insect can be conscious of its environment and its own body in that environment. This is demonstrated by looking at similarities between the functioning of brain structures between insects and vertebrates[3], and can be explained from an evolutionary perspective: those brain structures are an efficient solution to the basic problems of navigation.
Complex learning
Insects such as bees have complex learning skills that cannot be attributed to mere automatic reflexes or pre-programmed responses. They learn to use objects that they have never encountered before. For example bumblebees can learn to use small balls, rolling them to a target place to obtain a food reward. Bumblebees can learn this behavior from each other, and they can decide to use a ball with another color if that ball is closer to the target.[4]
Selective attention
A property of consciousness is selective attention: the ability to focus on one of several competing stimuli. This enables to ignore some stimuli and respond to the stimulus that is most relevant. Neural correlates of visual attention have been found in honeybees. When a bee has already seen something several times and it was not relevant for behavior, attention for that visual stimulus as well as activity in certain brain regions is reduced.[5]
Meta-cognition through uncertainty monitoring
Another indicator for consciousness is meta-cognition: being aware of one’s own mental states. One interesting mental state is the feeling of knowing something. There are experimental indications that insects such as bees and ants have an awareness of uncertainty. For example honeybees can learn to selectively avoid difficult choices that involve uncertainty.[6] They can learn to solve trials, a correct solution gives a reward (sugar), an incorrect solution gives a punishment (a bitter tasting chemical), but they can also decide to opt-out from the trial and receive neither reward nor punishment. When the difficulty of the trial increased, honeybees opted out more often because the risk of punishment increased. Using this option to opt out, honeybees improved their success-to-failure ratio. Hence, bees can perform rational behavior under uncertainty. In order to do that, bees need to monitor and evaluate their uncertainty. Also ants have the capacity for uncertainty monitoring.[7]
Indications for a utility function in insects
Pain can be distinguished from nociception because the former involves a consciousness. But merely having a consciousness is not enough for sentience, because conscious experiences and sensations can be neutral. For example, people with pain asymbolia or pain dissociation are able to consciously feel pain, but they do not have any negative evaluation or feeling of unpleasantness of that pain. For them, feeling a burned finger is like feeling a wet finger: the wetness of lukewarm water is neither positive nor negative. For positive and negative evaluations, a utility function is required. Some experiments indicate that insects have a utility function.
Making trade-offs
Animals can have a preference to avoid the unpleasantness of pain, but they can also prefer food (absence of hunger), safety (absence of predators) and so on. Sometimes they have to make trade-offs between different preferences. If they are able to make such trade-offs in a consistent way, they not only have a utility function, but their utility function is complex and integrated, involving several dimensions (preferences). Hence, being able to make trade-offs is an indicator for the presence of a utility function.
Fruit flies can make trade-offs. For example they choose to endure electric shocks in order to obtain access to alcohol. This is trading off a punishment for a drug.[8] They also trade-off safety for food: when fruit flies see an overhead shadow that resembles a predator, they disperse and hide to avoid potential predators. But if they are very hungry, they will continue eating from a food source.[9] Hence, fruit flies are able to weigh and evaluate benefits (food) and risks (predators), and that requires a unified utility function that measures both the preference for food and the preference for safety.
Communicating trade-offs
Even stronger evidence, is when insects are able to communicate their trade-offs to other animals. This is seen in the waggle dance of honeybees. When a bee has found a new food source, it faces a trade-off between abundance (how many flowers are there?), distance (how easy is it to get there?) and predation risk (how dangerous is it there?). This information is contained in the waggle dance to inform other bees.[10]
Self-administering drugs
Another behavior that is unlikely to be an unconscious process, is the self-administration of drugs. Insects like bees can get alcohol addictions, and they are able to actively look for alcohol.[11]
Sense of the future
Honeybees can not only make trade-offs between food and safety, but also between immediate versus deferred rewards. For example bees choose a larger droplet of sugar water received over five seconds above a smaller droplet received immediately.[12] These time trade-offs require self-control and a sense of the future.
Mood states (judgment bias)
There is evidence that bees and fruit flies have mood states. The most impressive evidence is probably the pessimism bias of agitated honeybees. When humans are anxious, depressed or stressed, they often become more pessimistic in uncertain situations, which means they have increased expectations of negative outcomes. The same can be seen in honeybees who are agitated by shaking them, simulating a predator attack on a bee hive. These bees are learned to associate a reward (water with sweet sugar) with one type of odor and a punishment (water with bitter quinine) with another odor. When bees face a new, third type of odor, they can either expect a reward or a punishment, so they can choose to approach or withdraw from the water. This is a situation of uncertainty. The bees that were shaken became more pessimistic: they more often withdrew from the water, expecting a punishment. And these shaken bees had lower levels of happiness hormones such as dopamine and serotonine.[13] This cognitive bias is a measure of negative emotional states. In another experiment, honeybees became more aggressive when they were in social isolation, as if being in isolation increased their levels of frustration.[14]
Also fruit flies seem to experience mood states such as depression. Due to enduring, uncontrollable stress, they experience anhedonia, a loss of appetite.[15] Antidepressants that work in humans also work in these fruit flies.[16]
Next to anhedonia, learned helplessness is also associated with depression in humans and can be a marker for mood states in insects. Researchers compared fruit flies that were exposed to shaking in an inescapable maze versus flies that were shaken in an escapable maze. Both groups of flies were shaken equally hard. But when these flies entered another, escapable box several hours later, the flies who were shaken in the inescapable maze took much more time to escape from the box.[17] They were less willing to search for an escape route, as if they accepted their bad situation. They learned to be helpless, as if they had pessimistic judgments about their efforts to escape. Also honeybees show evidence of learned helplessness.[18]
Next to depression, fruit flies also appear to have anxiety states, with the physiological and behavioral responses that are similar to anxiety-like states in mammals such as rats.[19] Anxiolytics that work in humans also work in these fruit flies.
Relief learning
Fruit flies can display relief learning. When the flies first experienced an electric shock and then smelled an odor after the shock, the flies learned that the odor predicted relief from the painful stimulus, so they started to like that odor (i.e. they approached it).[20]
Chronic pain
According to a recent study, fruit flies can experience chronic pain. When an injury damages a nerve in one leg of a fruit fly, the fly’s other legs had become more sensitive, even after the wound in the first leg healed.
Comparison of welfare
With the above evidence, we can estimate the likelihood that insects are sentient. So we can say that our confidence level for bees being sentient is say 60%. But even if we take the precautionary principle and assume they are sentient, we have to compare the welfare of a bee with the welfare of other animals such as humans. Interpersonal comparison of well-being is very difficult (I have made some attempts elsewhere).
Welfare has several dimensions, but when it comes to painful and pleasurable experiences, intensity and duration are two important dimensions (other dimensions are the certainty and the order of experiences, for example choosing between first pleasure and then pain or vice versa; see the felicific calculus). How can intensity and duration of for example pain be compared?
Concerning the intensity of a feeling, there are two extremal options. On the one side, we can equate minimum levels of perception (or minimum differences in utility) between individuals. There are indications that sense perception is discrete. For example a change of a subjective experience always requires a minimum (not infinitesimally small) amount of a change of an external stimulus. This is the just-noticeable difference or JND. Therefore, a painful sensation can be decomposed as a sum of just-noticeable differences in pain. The number of JNDs required to go from zero pain to the actual feeling of pain, can be considered as a measure of the level or intensity of the pain. More generally, the utility function can be represented as a multidimensional staircase with discrete steps in several directions. Moving from one situation (e.g. a reference point without pain and hunger) to another situation (e.g. with a certain level of pain and hunger), requires a number of steps of the utility function, and this number measures the overall preference of the new situation compared to the old situation.
Assume, as an example, that a bee is a sentient being and experiences burning pain from hot water. How bad are the burns for the bee? We can count the just-noticeable differences of pain from the burns. Going from zero burns to a just-noticeable burn decreases the utility of the bee with one unit. An extra burning sensation decreases the utility with another unit, and so on. Suppose the bee experiences 100 negative utility units, which means 100 just-noticeable differences of pain are exceeded. Now I want to compare this with my painful experience of hot water. Perhaps, hypothetically, when I put a fingertip (the size of a bee) in hot water, I also experience 100 negative utility units, as much as when the whole bee is submerged in hot water. So what the bee experiences is what I would experience when I burn my fingertip, and is much less painful than what I experience if I’m completely submerged in burning hot water. The underlying reason is the smaller size of the brain of a bee: a bee has fewer pain neurons and a smaller brain processing capacity to produce a feeling of pain. That means a bee would be less sensitive: relatively more external stimulus is required in order to generate a JND. If one JND of a bee is comparable to one JND for me, and if I know the number of JNDs of a bee, I can imagine how painful an experience is for a bee.
However, the second extremal option equates the maximum levels of experience between individuals. The brains of a sentient being are finite in size, so it is unlikely that they can generate infinite levels of pain and pleasure. If there exist a maximum level of pain, we can for example consider a bee who experiences maximum pain from burning in hot water. This can be equated to my maximum pain level when I feel burning hot water all over my body. If we take this comparison, a bee would be highly sensitive, like a human, and there will be huge amounts of insect suffering in the world.
To make it more quantitative: suppose there are 1018 insects and 1010 humans. Suppose an average insect has a brain with 105 to 106 neurons (bees have relatively large brains for insects, with roughly 1 million (106) neurons). A human has roughly 1011 neurons. Suppose that the number of JNDs of an experience with a given perceptual input is proportional to brain size, measured as the number of neurons (other options for brain size are the number of neuronal connections or synapses). Hence, a human is more than 105 times as sensitive as an insect. Suppose all humans and insects experience a similar situation, such as dying (from disease, injuries, coldness, starvation or predation). According to the first approach, equating the minimum levels of utility, this experience generates more than 105 times as much discomfort in a human than in an insect. So we have to discount the experiences of insects with a factor 105 or 106. As there are 108 more insects than humans, and their suffering counts 105 to 106 less, total insect suffering from dying is 100 to 1000 times higher than total human suffering from dying. However, if we take the second approach, equating the maximum levels of utility, insect experiences are not discounted and total insect suffering is 108 times higher than human suffering from dying. Furthermore, the rate of dying of insects is much higher, because their lifespans are much shorter (ranging from a few days to a few years). If an average insect lives for a few weeks, it’s mortality rate is 1000 times higher than the human mortality rate, which means insect suffering from dying can be 106 or 1011 times higher than human suffering.
Matters could be even worse for insect suffering when we account for the subjective duration of an experience. Insects such as flies have faster brain processing speeds. Consider vision: humans can see at most 60 flashes of light per second. Showing flashes at a higher frequency results in seeing a continuous light. The flicker fusion rate measures how fast a light has to be switched on and off before one sees it as a continuous light. A fly has a flicker fusion rate four times higher than a human, which means a fly can see 250 images or flashes per second. This explains why it is so difficult to swat a fly: a fly sees everything in slow motion, four times slower than we do.
Perhaps not only vision, but also conscious experiences have a maximum frequency. What is the smallest time interval that we can experience? Suppose an experience of pain is turned on and off. Suppose at this moment you do not feel pain, a second later you feel pain, another second later the pain is gone. That means every second you can have a different conscious experience. But what if we increase the frequency? At this moment you do not feel pain, a millisecond later there is a pinprick. Another millisecond later the needle is removed, and so on. Now you might feel a slight, continuous pain instead of different pain pulses, which means you cannot consciously distinguish milliseconds.
Suppose the flicker fusion rate of your consciousness is 60 experiences per second, as with vision. This is as if you have an internal clock that has a moving hand rotating full circle in 60 steps per second. Every position of the moving hand corresponds with a different conscious state. You can have at most 60 different conscious experiences per second. But some insects may have faster internal clocks. In one real second, they can have 250 different conscious experiences. If you experience pain for one second, you actually have 60 conscious states of pain. But if insects can feel pain and if they feel pain for one second at a higher brain speed, that corresponds with 250 conscious states of pain. It is as if you would experience 4 seconds of pain.
Perhaps the tiny brains of insects indicate that the intensity of their pain experience is lower than the intensity of pain experienced by animals with larger brains. But if their brains are faster, they experience pain in slow motion, meaning that a second of pain appears to last longer for insects. That means one second of pain for a human should be discounted compared to one second for an insect: one second of insect pain counts a few times more than one second of human pain.
One further complication is our sensitivity for time intervals. We have a decreasing marginal sensitivity for time: the longer the time interval, the less important an extra second becomes. The difference in the preference for 0 seconds of extreme pain above 1 seconds might be bigger than the difference in the preferences for 1000 versus 1001 seconds of extreme pain. In the latter case, the one extra second is less important (you probably won’t notice the extra second). The rate of decreasing marginal sensitivity for time intervals might also depend on brain complexity and size. The bigger the brain, the more memory capacity it has and the more previous seconds are remembered and taken into account. This means that insects with smaller brains could have an almost constant marginal sensitivity for time: no matter how much time they experienced pain, an extra second of pain remains equally bad for them. On the other hand, it is possible that insects are not capable to perceive and judge long time intervals, which means they don’t have a preference between 100 seconds of pain and 1000 seconds of pain.
Avoiding insect suffering: prioritization
Humans are harming insects, by accidentally killing them (when running around, driving cars,…), intentionally killing them (using insects for food and clothing, using insecticides, insect traps, …) or indirectly killing them (by competing for resources, natural habitat destruction, pollution…). So one could argue we need less cars, less agriculture, less pollution, less concrete, less insecticides, less walking on the grass,… But things are not so simple when it comes to wild animal suffering.
First, a lot of insects are parasites or predators that kill other insects. So if we (accidentally, intentionally or indirectly) kill some insects, especially predators, we might save the lives of many other insects. Or stated differently: saving one ladybird might mean killing hundreds of aphids. Second, insects in the wild can have net-negative lives, i.e. short lives with more negative than positive experiences. These are lives not worth living. This is due to their reproductive strategy: a fertile adult insect can lay thousands of eggs. If the insect population does not explode at an extreme exponential rate, it is logically required that almost all of the newborn insects will have to die prematurely. The ways of dying are often extremely negative experiences: coldness, starvation, predation, parasitism,…. If an insect is killed, it prevents the birth of many insects with net-negative lives. So, if most insects face very short lives anyway and die horrible deaths anyway, it is far from clear whether killing insects increases overall future insect suffering. We need much more scientific research to estimate the overall effect of killing insects on global welfare.
Prioritization research involves looking for the most effective methods and interventions to improve insect welfare and reduce insect suffering. For example if we come to the conclusion that we should minimize killing insects, we can look for effective means to reduce the killing of insects. One interesting opportunity is the use of non-lethal methods for insect pest control. There are also many methods to avoid and remove insects in your house, for example catching flies with a transparent glass and releasing them outside.
If killing insects is inevitable, we should look for humane killing methods. Switching to more humane insecticides in agriculture could be a very effective way to minimize suffering. Using natural predators to combat insect pests on the other hand might be as bad as using inhumane insecticides.
In our prioritization research we should not only consider harm to insects caused by humans. Harm caused by nature (e.g. by other animals) counts equally. We can investigate safe and effective methods to intervene in wild nature to improve insect welfare. For example new technologies such as gene editing and gene drives could help reduce insect suffering by controlling insect populations in order to limit predation, parasitism, starvation and other causes of suffering.
Avoiding wild insect suffering: bee protection
Our prioritization research should not only consider different methods or interventions, but also consider which populations or species to target. Probably the clearest case can be made for bee protection. One three levels, bees are special.
First, as we have seen above, of all the studied insects, bees (the clade of antophyla) show probably the most scientific evidence for having consciousness and sentience. If they are not conscious, then other insects are probably not conscious either.
Second, unlike most other insects, bees have no negative externalities. A lot of insects are predators or parasites, which means they harm other animals. Herbivorous insects do not kill others, but they can be a pest in agriculture. Beas are the best: they do not harm other animals (except in self-defense) and they do not destroy food (e.g. crops) for other animals.
Third, bees have huge positive externalities: they improve crop yields by pollination. So bees even help to provide food for other animals (humans, birds) who like to eat fruits.
The bad news is: bees face difficult times. The colony collapse disorder kills many bees. Some neonicotinoid insecticides can make the bees more vulnerable to diseases and parasites, so an effective intervention is to replace those insecticides. However, those insecticides are not the only culprit of the colony collapse disorder. Pests such as pathogens and parasites are the biggest threat. Effective solutions for these threats are not yet known, so research is important.
Other ways to help bees, especially in western Europe, is to combat the Asian hornet. This is an invasive exotic species, and European bees do not recognize this insect as a predator. As a consequence, Asian hornets kill many bees (as well as many other insects). Eradicating a nest of Asian hornets could save many bees and other insects. This measure also finds support amongst environmentalists who are against invasive exotic species.
We also have to be careful not to take ineffective or counterproductive measures to protect bees and other insects. One example of an ineffective measure is the ban on GMOs in Europe. Bt-crops are GMOs that produce an insecticide (Bt) that is normally found in soil bacteria. According to a meta-analysis, Bt-crops are not harmful for honeybees.[21] Another meta-analysis shows that fields with Bt-crops have higher biodiversity levels of nontarget invertebrates (beetles, butterflies, spiders,…) compared to non-GMO fields where Bt-insecticide is sprayed (including organic fields, because Bt-insecticides are allowed in organic farming).[22] Spraying of Bt-insecticides not only kills the pest insects but also many other nontarget insects. Bt-crops can reduce the spraying of Bt-insecticides, and hence reduce the overall killing of insects. Also organic farming could be a counterproductive measure for bee protection: organic farming allows the use of insecticides that are harmful for bees, and the lower crop yields in organic farming means that more agricultural land is required. Hence, land that could serve as flower meadow for bees is sacrificed. We have to be careful however: as mentioned above, reducing insect killing or increasing natural habitat might increase insect suffering. So we first need more scientific research to estimate the overall effects of interventions.
The same goes for the most obvious animal rights issue related to bees: the consumption of honey. The production of honey is in many ways harmful to bees: they are often killed accidentally and in many cases also intentionally by the bee keepers (culling less productive hives, making the bees vulnerable to diseases by taking away their nutritious honey, clipping the queen bees’ wings,…). However, boycotting honey most likely means a replacement by other sweeteners, and most of those sweeteners come from agruculture that harms wild insects. Sugar involves the accidental and intentional killing of insects (using insecticides, machines,…). Perhaps very sweet stevia or some artificial sweeteners such as aspartame are better for the insects because they involve less agriculture. But a cookie or breakfast cereals with a strong sweetener contains less volume of sugar and hence more weight in grains and fats, which means more agriculture. On the other hand, more agriculture means less natural habitat and hence less suffering of insects in wild nature. The situation is very complex; we really don’t know the overall effects of honey. The situation is comparable to fishing: there is direct harm to used animal, the captured fish, but all the indirect effects on aggregate welfare of aquatic animals are unknown. As a rule of thumb one could use a provisional deontological principle and abstain from the consumption of fish and honey, avoiding direct harm associated with the use of animals. The idea is: if the presence of the body (of the bee or fish) is required to obtain your objective (consuming honey or fish), and if that animal is harmed (i.e. treated against his/her will), that harm counts more than all the unknown indirect effects. In the meantime, much more research about those indirect effects is required. We should not underestimate the value of information about indirect consequences of our choices. If the picture about indirect effects become clear, we might come to the conclusion that fishing or honey production are overall not negative for aggregate welfare. The known indirect effects gain more weight in the overall evaluation.
Avoiding cultured insect suffering: insects for food
Another important measure, is the decrease of insect farming. Insects are used for food and clothing (silk). An example is cochineal, a scale insect that produces a red dye carmine that is used as a colorant in food. It takes almost 100.000 insects to produce one kilogram of dye. Replacing cochineal dye with synthetic dyes is very feasible, because synthetic dyes are about four times less expensive.
Another worrying trend is the rise of insect meat consumption. Worms and crickets are used for insect burgers and sausages. The problem is that a lot of insects are required for insect meat. If one insect sausage requires more than 100 insects, whereas one beef sausage requires less than one thousandth of a cow, the number of animals used and killed for insect meat could be almost a million times higher than beef meat. Even discounting for brain size or pain sensitivity, insect meat can involve a lot of suffering.
Insect meat is often promoted as a more sustainable option than livestock meat. Chicken meat has the lowest environmental footprint of all livestock meats (lower than pork and beef), and the footprint of insect meat is about half that of chicken meat. However, the environmental impact (in terms of agricultural land use and greenhouse gas emissions) of insect meat is still 10%-50% higher than plant-based protein sources.[23] Eating insects is usually less efficient than eating plants. In other words, insect meat not only requires the intentional direct killing of insects for meat, but also more indirect and accidental killing of insects in agriculture to produce insect feed, compared to vegan alternatives. As insect meat is on the rise in Western countries but still far from being established, it is still possible to halt insect meat. Hence, campaigning against insect meat is very feasible. And equally feasible is campaigning against livestock farming, because a lot of insects are fed to pigs and chickens.
[1] 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.
[2] Kuldas, S., Ismail, H. N., Hashim, S., & Bakar, Z. A. (2013). Unconscious learning processes: Mental integration of verbal and pictorial instructional materials. SpringerPlus, 2(1), 105.
[3] Barron, A. B., & Klein, C. (2016). What insects can tell us about the origins of consciousness. Proceedings of the National Academy of Sciences, 113(18), 4900-4908.
[4] 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.
[5] Paulk, A. C., Stacey, J. A., Pearson, T. W., Taylor, G. J., Moore, R. J., Srinivasan, M. V., & Van Swinderen, B. (2014). Selective attention in the honeybee optic lobes precedes behavioral choices. Proceedings of the National Academy of Sciences, 111(13), 5006-5011.
[6] Perry, C. J., & Barron, A. B. (2013). Honey bees selectively avoid difficult choices. Proceedings of the National Academy of Sciences, 110(47), 19155-19159.
[7] 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.
[8] Kaun, K. R., Devineni, A. V., & Heberlein, U. (2012). Drosophila melanogaster as a model to study drug addiction. Human genetics, 131(6), 959-975.
[9] Gibson, W. T., Gonzalez, C. R., Fernandez, C., Ramasamy, L., Tabachnik, T., Du, R. R., … & Anderson, D. J. (2015). Behavioral responses to a repetitive visual threat stimulus express a persistent state of defensive arousal in Drosophila. Current Biology, 25(11), 1401-1415.
[10] Abbott, K. R., & Dukas, R. (2009). Honeybees consider flower danger in their waggle dance. Animal Behaviour, 78(3), 633-635.
[11] Maze, I. S., Wright, G. A., & Mustard, J. A. (2006). Acute ethanol ingestion produces dose-dependent effects on motor behavior in the honey bee (Apis mellifera). Journal of insect physiology, 52(11-12), 1243-1253.
[12] Cheng, K., Peña, J., Porter, M. A., & Irwin, J. D. (2002). Self-control in honeybees. Psychonomic bulletin & review, 9(2), 259-263.
[13] Bateson, M., Desire, S., Gartside, S. E., & Wright, G. A. (2011). Agitated honeybees exhibit pessimistic cognitive biases. Current Biology, 21, 1070e1073.
[14] 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.
[15] Abelaira, H. M., Reus, G. Z., & Quevedo, J. (2013). Animal models as tools to study the pathophysiology of depression. Brazilian Journal of Psychiatry, 35, S112-S120.
[16] Ries, A. S., Hermanns, T., Poeck, B., & Strauss, R. (2017). Serotonin modulates a depression-like state in Drosophila responsive to lithium treatment. Nature communications, 8, 15738.
[17] Brown, G. E., Mitchell, A. L., Peercy, A. M., & Robertson, C. L. (1996). Learned helplessness in Drosophila melanogaster?. Psychological reports, 78(3), 962-962.
[18] Dinges, C. W., Varnon, C. A., Cota, L. D., Slykerman, S., & Abramson, C. I. (2017). Studies of learned helplessness in honey bees (Apis mellifera ligustica). Journal of Experimental Psychology: Animal Learning and Cognition, 43(2), 147.
[19] Mohammad, F., Aryal, S., Ho, J., Stewart, J. C., Norman, N. A., Tan, T. L., … & Claridge-Chang, A. (2016). Ancient anxiety pathways influence Drosophila defense behaviors. Current Biology, 26(7), 981-986.
[20] Yarali, A., Niewalda, T., Chen, Y. C., Tanimoto, H., Duerrnagel, S., & Gerber, B. (2008). ‘Pain relief’learning in fruit flies. Animal Behaviour, 76(4), 1173-1185.
[21] Duan, J. J., Marvier, M., Huesing, J., Dively, G., & Huang, Z. Y. (2008). A meta-analysis of effects of Bt crops on honey bees (Hymenoptera: Apidae). PloS one, 3(1), e1415.
[22] Marvier, M., McCreedy, C., Regetz, J., & Kareiva, P. (2007). A meta-analysis of effects of Bt cotton and maize on nontarget invertebrates. Science, 316(5830), 1475-1477.
[23] Smetana, S., Mathys, A., Knoch, A., & Heinz, V. (2015). Meat alternatives: life cycle assessment of most known meat substitutes. The International Journal of Life Cycle Assessment, 20(9), 1254-1267.
Van Diepen J. e.a. (2018). Eiwit-transitie Vlaanderen. Studie naar de status en het potentieel van (hoog-) technologische oplossingen om vleeseiwitten te vervangen in het dagelijks dieet. Blonk Consultants. Gouda, Nederland.
Broekema R. & van Paassen M. (2017). Milieueffecten van vlees en vleesvervangers. Blonk Consultants. Gouda, Nederland.
Stijn Bruers, the rational ethicist 