How Scientists Study Animal Emotions Without Assuming Human Feelings

Why animal emotions require indirect evidence

When a dog changes posture or a cat responds differently to a situation, it is natural to ask what the animal is feeling. But dogs and cats cannot describe their inner experience through language-based self-report. Researchers therefore cannot simply ask an animal to name an emotion and record the answer.

How animal emotions are studied scientifically is therefore an indirect evidence question. Researchers measure changes in behaviour, body processes, cognition and, where possible, neural or surface-temperature responses. They then ask whether several findings fit together across controlled contexts. This approach can provide evidence consistent with an affective state, while stopping short of claiming direct access to a dog’s or cat’s private experience.

What exactly is being measured?

Researchers use terms such as emotional valence and arousal to describe dimensions of affect. Valence refers broadly to whether a state is more positive or negative. Arousal refers to the level of activation or intensity associated with a state. These dimensions are useful because they do not require researchers to assume that an animal’s experience matches a human label such as happiness, fear or sadness.

Animal emotion may involve behavioural, physiological, neural, cognitive and potentially conscious components. Some of these components can be measured; others are not directly observable. In particular, whether a measured response involves conscious subjective experience remains controversial. Definitions and terminology also remain contested, so studies may use different frameworks for describing the same broad research problem.

Behavioural indicators: useful, but not self-explanatory

Behaviour is often the most visible starting point. Studies of dogs and cats may record posture, facial expression, vocalization, attention and how an animal responds to a particular situation. Researchers may also examine learning, memory and other changes in performance.

Context is essential. A posture or vocalization can have different meanings in different circumstances, and animals do not all respond identically. Breed, age, individual history, environment and learning experiences may influence behaviour. For that reason, a familiar action is not treated as a direct translation of a human emotion. The question is whether the response changes systematically under controlled conditions and whether other indicators show a compatible pattern.

• Body posture and movement
• Facial expression
• Vocalization
• Attention and responses to surroundings
• Learning and memory
• Responses in a specific context

Physiological indicators: signals, not emotional labels

Physiological measures add information that may not be visible in behaviour. Heart rate and stress-related physiology can show that an animal’s regulatory systems have changed between conditions. Such measurements may help researchers compare responses and examine how bodily processes accompany behaviour.

They do not, however, uniquely identify one emotion. A physiological response can be compatible with more than one affective state, and its interpretation depends on the experimental context. This is why a change in heart rate or another stress-related measure should not be presented as proof of a particular feeling in a dog or cat.

• Heart rate and related autonomic responses
• Stress-related physiology
• Other measurable changes in body regulation

Cognitive methods: studying attention, memory and judgement

Cognitive methods provide another way to investigate affect without asking an animal to speak. Researchers can study attention, memory and judgement-related responses. For example, a task may examine how an animal responds to information that is ambiguous or how an affective condition relates to learning and memory.

These approaches are valuable because they can explore possible differences in valence, distinguish affective states that may share a similar valence and investigate links between cognition and emotion. They remain indirect measures: a performance difference does not automatically reveal a conscious human-like feeling. Its interpretation depends on the task, comparison conditions and other evidence.

Neural, facial and thermal approaches

Research may also include neural measures, facial analysis and infrared thermography. Infrared thermography is a non-invasive method that records body-surface temperature. Changes in particular body regions can reflect autonomic processes, including altered blood flow or stress-induced hyperthermia.

Thermal images can therefore contribute to a broader picture of an animal’s response. They cannot independently identify one emotion, and the method has limitations that researchers must consider. The same principle applies to facial or neural measures: they can provide additional evidence, but they do not create a direct window into subjective experience.

Why scientists combine methods

The strongest inferences generally come from convergence. If behaviour, physiology and a cognitive response all change in a consistent pattern across controlled contexts, the combined evidence may support a cautious interpretation of an affective state. No single measurement has to carry the whole conclusion.

Researchers also need to separate temporary emotional states from relatively stable personality traits. An individual dog or cat may show a consistent tendency across time, while a temporary response reflects a particular situation. Confusing these levels can make a stable characteristic appear to be a momentary feeling, or the reverse.

Evidence versus anthropomorphism

Anthropomorphism is the interpretation of animal behaviour through familiar human concepts. It is understandable for a pet owner to describe an animal using words such as love, jealousy or sadness. Scientifically, however, the label is an interpretation, not evidence by itself.

Researchers can investigate whether such an interpretation is supported by repeatable changes across behaviour, physiology, cognition and other measures. They must also consider alternative explanations and avoid assuming that similar outward behaviour means identical inner experience. Human-like labels can serve as hypotheses, but they need validation rather than automatic acceptance.

Limits and open questions

Important questions remain open. The field has not reached complete agreement on how animal emotion should be defined or which combination of measurements is best. Positive, low-arousal states are particularly difficult to identify reliably. Quiet or subtle responses may be harder to distinguish than highly activated responses, and results may not apply equally to every dog or cat, breed, age, environment or learning history.

Most importantly, researchers cannot directly confirm an animal’s subjective conscious experience through language-based self-report. Scientific studies can document patterns consistent with affective states and can test competing interpretations. They cannot turn an observable response into certainty about a human-like inner feeling.

What the evidence can—and cannot—say

The scientific study of animal emotions is therefore a process of careful comparison rather than emotional mind-reading. Researchers observe dogs and cats in context, measure multiple components and look for convergence across indicators. Behaviour, physiology, cognition, neural approaches and infrared thermography each contribute part of the picture, with specific limitations.

This evidence can improve understanding of how animals respond to their surroundings without assuming that their experiences are identical to ours. The most defensible conclusion is often precise and modest: a pattern of findings is consistent with a particular affective dimension or state under studied conditions, while the animal’s private conscious experience cannot be directly confirmed.