All species of fish are vertebrate animals, and as such are sentient beings, but can they feel emotions like us?
In the human supremacist and speciesist paradigm we currently live in, emotions are often treated as exclusively human, with perhaps a few favoured nonhuman mammals (like dogs or cats) granted that “privilege” because they live with us. Hardly ever are any of the species of the different fish classes (there is not a single biological class named “fish”) included. However, is this exclusion based on facts? It is acceptable to simply assume that the trillions of individuals of these classes are emotionless beings. When a fish flees in panic, waits anxiously in a corner of a barren tank, or plays with an object, how different is the cause and the brain process from a human exhibiting the same behaviour?
To explore this, it helps to separate “emotion” from “feeling”, to look at what brains and hormones actually do, and to examine the growing body of research with fish species. When we do, the idea that they are insentient, emotionless objects collapses, and it becomes clear that their emotional lives are not only real, but also older than ours, so the interesting question should be whether humans feel emotions like them, or we have not quite got there yet (as we just arrived, evolutionary speaking).
What emotions are
“Emotions” are complex physiological and neurological states with specific brain activity, hormone release, and behavioural patterns that prepare an individual to deal with challenges or opportunities in a coherent way that requires a coordinated response of several organs and body parts. An emotional state like fear, for instance, involves particular neural circuits firing, stress hormones being released, changes in heart rate and behaviour (such as freezing or fleeing), and shifts in attention and learning.
Feelings can lead to emotional states, and emotions often include several feelings, so these concepts are closely related, but not identical. Feelings are the individual sensory-driven subjective experiences of these states, and they can be short-lived and very localised (such as pain). However, emotions are more complex, can be externally or internally driven, affect more organs and parts of the body more systematically, and can be more lasting (even lingering beyond necessary). Another way to look at this is that emotions are largely complex, automatic physiological and neural responses, but feelings are the individual conscious subjective experience of those emotional responses.
Scientists often talk about “affective states” (patterns such as fear, anxiety, relief, or motivation) that can be detected by changes in physiology and decision‑making, even when the subjective experience cannot be verbally reported. Many of these states are associated with basic brain chemicals such as dopamine (linked with reward and motivation), serotonin (mood and impulse control), and stress hormones such as cortisol, as well as with brain regions like the amygdala and hippocampus in mammals. The key question, then, is not whether individuals of fish species have a neocortex like ours, but whether they have functionally similar systems that generate comparable states.
Teleosts (the huge group that includes sardines, salmons, and haddocks) do indeed have complex monoaminergic systems: dopamine, serotonin, and related neurotransmitters that regulate aversive stress, reward, and social behaviour in ways that parallel mammals. They also possess nociceptors (specialised nerve endings that detect potentially damaging stimuli), which send signals into the spinal cord and brain pathways capable of central processing, not just simple reflexes. This hardware is necessary if emotions are going to function as integrated “warning and guidance systems” in an organism’s life.
At the brain‑structure level, teleosts do not have a mammalian‑style layered cortex, but they do possess forebrain regions that developmental biologists and neuroanatomists identify as homologous to the mammalian amygdala, hippocampus, and related areas involved in emotion and memory. In zebrafishes, for example, subdivisions of the ventral telencephalon show the same molecular markers and connectivity patterns that define the extended amygdala in mammals, and damage to these regions alters social and reproductive behaviours in ways that mirror amygdala lesions in other vertebrates. These findings support the view that individuals of fish species have an “emotional brain” architecture shaped by the same broad evolutionary forces as ours.
We have significantly less data on cartilaginous species (sharks/rays) or jawless species (lampreys), though scientists generally apply the precautionary principle to them, so they assume that what we find in teleosts could also be found in others.
If emotions are biological control systems built from specific neurochemicals and circuitry, then these aquatic vertebrates are equipped with those systems.
Pain, suffering, and the evolutionary logic of emotions in fish species
One of the key ethical questions that has driven the discipline of animal welfare is whether individuals of fish species can suffer. Sentience is typically defined as the capacity to have positive and negative experiences (suffering being having a negative experience that lingers), and there is now substantial evidence that fish species meet this definition. They possess nociceptors responsive to mechanical, thermal, and chemical damage, and when these are activated, they show marked behavioural and physiological changes (i.e., they rub or guard the affected area, reduce feeding, and display altered breathing and stress‑hormone levels). These responses are modulated by analgesic drugs like morphine, which reduce both the abnormal behaviours and the physiological stress markers, just as in mammals.
These pain‑related responses can be costly. Rainbow trouts injected with a noxious substance, for instance, reduce their normal feeding and exploration even when they are hungry, and they will pay costs (such as swimming against a current) to avoid the context associated with the painful event. Such trade‑offs indicate central evaluation, such as individuals are weighing avoidance of discomfort against other needs, which is best explained if pain is more than a mere reflex. Reviews of this work conclude that while no experiment can directly read experience, the simplest explanation of the data is that at least some species feel pain.
Beyond pain, researchers are increasingly interested in how emotional states shape fish decision-making. Cognitive‑bias (or judgement‑bias) tasks are used across species to infer affect: animals in more positive states tend to interpret ambiguous signals “optimistically,” whereas those in negative states behave more “pessimistically.” These paradigms have been adapted to these aquatic vertebrates, who are trained that one cue predicts a strong reward (such as a richer food patch) and another predicts a weaker or aversive outcome, and then are presented with intermediate cues. Recent work shows that manipulations that increase fish stress (such as social isolation or unpredictable housing) shift their responses towards the pessimistic side, suggesting that their current emotional state biases their expectations of the future. This goes beyond mere sentience into mood‑like processes, where sustained emotional states colour how the world is judged.
From an evolutionary angle, this all makes sense. Emotions offer powerful advantages:
- Fear and anxiety promote avoidance of predators and risky situations, with stress systems and amygdalar circuits adjusting vigilance and learning.
- Positive states associated with reward systems support exploration, foraging, and social bonding, encouraging animals to seek out beneficial opportunities.
- Pain discourages the repetition of damaging behaviour and promotes protective care of injured tissues.
These functions are ancient solutions to the basic problem of surviving and reproducing in a complex, changing world. These aquatic animals live in environments dense with threats and opportunities, such as predators, social rivals, changing temperatures, fluctuating oxygen levels, currents, and complex social hierarchies. It would be odd, from an evolutionary point of view, if lineages that have been navigating such environments for hundreds of millions of years had evolved sophisticated sensory and learning systems but somehow skipped the equally adaptive systems that integrate bodily states, value, and behaviour, the very core of what neurobiologists call emotion.
If one grants that all aquatic vertebrates are sentient, the idea that they would have no emotional states becomes even more implausible. Sentient beings benefit from being able not only to detect threats and opportunities, but also to assign them an affective weight (to treat some outcomes as dangerous, others as promising, and to generalise these evaluations across similar situations). In fish species, the same monoamine systems and limbic‑like circuits that support such functions in mammals are present and active. Denying emotions to these vertebrates while accepting them in other vertebrates with comparable circuitry looks less like science and more like speciesist convenience.
Research that directly shows emotions in fish species
The most compelling evidence comes from studies designed not just to show that individuals of fish species are sentient, but to probe specific emotional processes such as fear, anxiety, mood, and even proto‑empathy and joy.
One influential study on “emotional fever” in zebrafishes is a good starting point. Stress‑induced hyperthermia (a small, transient rise in body temperature following a psychological stressor) has long been treated as a marker of emotional responses and, in some accounts, of a degree of consciousness in vertebrates. For a long time, no species of fish was thought to show this pattern. In an experiment using a tank with a thermal gradient (ethically questionable, as most experiments on sentient beings are), groups of zebrafishes were briefly confined in a net (a stressful event) and then allowed to swim freely. The stressed individuals actively moved into warmer compartments and stayed there, raising their body temperature by 2–4°C, whereas control individuals that had not been stressed remained near their usual temperature zone. This pattern is hard to explain as a simple mechanical response because it aligns closely with emotional fever seen in mammals and birds, indicating that these aquatic vertebrates undergo a global, stress‑linked physiological shift typical of an emotional response.
Another line of work uses judgment-bias tasks explicitly as measures of emotional state in individuals of fish species. In these experiments, individuals are first trained that one location or cue predicts a high‑value reward, and another predicts either a low reward or a mild aversive event. After establishing this, the fish is exposed to conditions designed to alter its affective state (for instance, being housed in more enriched versus more barren tanks, or undergoing a mild stressor). When later tested with ambiguous, intermediate cues, individuals in more negative conditions tend to behave as if expecting the worst outcome, while those in more positive conditions are more likely to behave “optimistically”. A recent study found that zebrafishes with higher trait sensitivity to stress showed stronger shifts in these cognitive‑bias measures, echoing patterns seen in mammals where anxiety‑prone individuals interpret ambiguity more negatively. This is targeted evidence about emotional states.
Perhaps most striking, from an ethical perspective, is recent work on emotional contagion and proto‑empathy in fish species, again using zebrafish as a model. Emotional contagion occurs when an individual who merely observes another in distress begins to display similar emotional and behavioural signs, not through simple mimicry, but through internalising the other’s state. In a series of experiments, observer zebrafishes watched demonstrator individuals exposed to an alarm substance that triggers fear‑like freezing and erratic swimming. Observers, without any direct contact with the alarm substance, began to show matching fearful behaviours. When the oxytocin system (a hormone and neuropeptide deeply implicated in social bonding and empathy in mammals) was genetically disrupted in the observers, this social fear contagion disappeared, but when oxytocin was restored, it reappeared. Brain imaging showed that the regions activated during this process in zebrafishes are homologous to those involved in emotional contagion in rodents, including parts of the ventral telencephalon and striatum. This combination of oxytocin dependence, matching distress behaviour, and activation of conserved social‑emotional circuits is precisely what one would expect for a basic, evolutionarily conserved form of empathy.
There is also accumulating evidence that some fish species also engage in behaviours consistent with positive affect, including play. Researchers have documented fish species repeatedly interacting with novel objects in ways that have no obvious immediate survival benefit (for instance, cichlids and other species repeatedly striking and riding a bottom‑weighted thermometer, or playing with water flows, returning again and again to “surf” jets or bubbles). Such patterns fit established criteria for play in non‑human animals and are best interpreted as associated with pleasure‑like states. While play alone does not prove joy, it strongly suggests that they, like other vertebrates, have positively balanced emotional systems motivating activity beyond mere survival maintenance.
Taking all this evidence together (emotional fever, judgement‑bias shifts, oxytocin‑mediated fear contagion, and play‑like behaviours) shows specific emotional mechanisms, using the same conceptual tools that researchers apply to mammals, including humans. So yes, a fish can feel emotions, and if they may not be identical to ours, they are likely to be similar.
