The debate about fishes’ ability to feel pain focuses on sentience, i.e., to feel and be aware. According to many ethicists, sentience must be present to determine whether an animal is entitled to moral consideration. Fish sentience, however, is a highly polarized topic: at one extreme is the argument that fishes have no awareness of anything at all, including pain, while at the other extreme is the notion that fishes not only feel pain, but also other sensations like fear and joy. Skeptics claim that fish sentience is irrelevant, and that classifying fishes as sentient would be harmful to humans. 

The controversy rests on an important question: Why is sentience in fishes ethically relevant, and what are the consequences of recognizing it? Sentience implies that “there are limits to what humans can ethically do” to an animal. Currently, fishes experience practices that are likely to cause extreme suffering if they are sentient. Some imply that treating fishes as sentient (if they aren’t) would harm aquaculture, leaving society less able to feed a growing global population. In addition, many resources are devoted to legislation and policy that provide special status to sentient beings. If fish sentience is uncertain, then it’s important to balance the costs to industry, consumers, and research, while at the same time addressing the suffering caused by assuming fishes are not sentient when they are.

As non-verbal animals who can’t report how they feel it remains unclear how fish sentience can be determined. Complicating things is the lack of universally accepted measures of sentience for non-verbal species. The question is, are fishes aware of harmful stimuli, and can they feel true pain? This requires an understanding of how fishes react to tissue damage, and if their reactions indicate true pain or just an unconscious response.

In this review, the authors cover several topics in the fish sentience debate. The paper begins with a review of basic concepts of consciousness, including primary or “phenomenal” consciousness (P-consciousness), which is the ability to feel or be aware. This type of consciousness is most often associated with sentience. The other type is “access consciousness,” which can influence actions such as speech. Access consciousness allows for self-report, and through it, researchers can make inferences about P-consciousness. P-consciousness comprises multiple, higher-order dimensions (such as self-reflection, taste, pain, and sight). While some people use “sentience” to describe all aspects of P-consciousness, others use it to describe only one specific type — the ability to experience emotions. What’s more, evidence for one component of P-consciousness tells little or nothing about the existence of others. As such, testing a fish (or any animal) for sight and smell reveals nothing at all about their capacity to feel pain, which is at the core of sentience.

Two approaches are currently used to infer the experience of pain in non-verbal animals. The first is a deductive approach that relies on animals having specific structures of the brain, based on human anatomy. People who argue that fishes do not feel pain tend to use this approach because fishes lack the mammalian brain cortex typically associated with P-consciousness. The second approach is ‘inductive,’ meaning that researchers look for behaviors that align with sentience in the face of pain (e.g., rubbing, limping). The difficulty with this approach is distinguishing whether the observed behavior truly demonstrates an underlying experience of pain. This is important because there is a large body of research supporting many superficial behaviors that can occur without P-consciousness, or sentience. 

The authors then go on to review examples of specific responses that do not require sentience. They refer to these as “red herrings.” Research on subjects assumed not to have sentience or any kind of P-consciousness — plants, protozoa, spines, mammals without the cerebral part of their brains, and unaware humans — suggest that there are a range of learned and unlearned behavioral responses that do not seem to require P-consciousness. Examples include the ability to lean away from adverse stimuli (seen in plants) or the ability to avoid walking into an obstacle (seen in humans without sight due to brain damage).  Other behaviors appear emotionally-driven but do not, in fact, require sentience. As such, it’s difficult to be certain about the presence or absence of sentience, based on this type of evidence alone.  

There are other types of behavioral and cognitive evidence that could identify P-consciousness, and sentience more specifically. Examples include studying an animal’s working memory and operant conditioning, both of which are present in subjects assumed not to have sentience. Another indicator is self-report measures that gauge responses to certain stimuli but not others. Higher-order cognitive processes such as episodic memory and self-recognition are a fourth type of evidence that can yield insights into animal awareness.

Measuring P-consciousness, and therefore sentience, is currently impossible in non-humans. Even in humans, assessments based on self-reports can be inaccurate. The absence of scientifically proven evidence enables both sides of the fish sentience debate to continue. However, the authors argue that withholding fish protections  until it’s known if they’re sentient would be ethically wrong. Instead, a reasonable stance would be to support practical guidelines that choose to protect fishes while balancing the needs of other stakeholders. While much sentience research focuses on the question ‘is there evidence that this species is sentient?,’ the authors suggest that the question should be, ‘are we sure it is not?” Those interested in animal welfare should encourage researchers to focus more on understanding sentience, versus P-consciousness in general, if they want to get to the bottom of the debate.