Hard to study: The most extensively used measure of emotions is applied by simply asking people questions about their feelings, but measures of subjective experience are limited to humans who can communicate and cannot be applied to nonhuman animals or young children.
Illustration by Yi-Hui Chang

Emotion research has a communication conundrum

In 2025, the words we use to describe emotions matter, but their definitions are controversial. Here, I unpack the different positions in this space and the rationales behind them—and I invite 13 experts to chime in.

By Nicole Rust
5 September 2025 | 11 min read

I remember a faculty meeting during my first year as an assistant professor in a psychology department, now 15 years ago, when a colleague proposed that we search for a faculty member in “affective science.” I was puzzled by what that meant, but afraid to ask for fear of looking foolish.

There are nontrivial reasons that some emotion researchers insist on calling brain functions that others label an “emotion” an “affective state” instead, and likewise, what others call a “mood” a “temporally extended affective state.” At the same time, these unfamiliar terms impede communication with non-experts, including members of the public. With the value of U.S. science being called into question and a growing need for scientists to communicate their research more clearly, this matters. Conflicting definitions of “emotion” also lead to confusion.

Controversy around how to characterize emotions dates back to ancient times; Aristotle preferred the passive term pathos to describe them. Today, researchers still don’t agree. Following the recent publication of the paper “Conserved brain-wide emergence of emotional response from sensory experience in humans and mice,” one researcher expressed concerns about the current state of the field to the press, explaining, “The underlying problem is that scientists still don’t agree on how to define an emotion.”

What’s behind all this disagreement? And what can be done to broaden communication about emotion research with the outside world? Here, I unpack the different positions in this space and the rationales behind them. I also invite several experts to chime in on the question, “What should be done about the communication conundrum in emotion research?”

What makes studying emotion so challenging?

 Understanding the debate around terms in emotion research requires appreciating why emotions are more challenging to study than other brain functions. To investigate memory, for example, researchers often design a memory test for which there are ground-truth correct and incorrect answers to the questions posed. In contrast, emotion is a subjective experience for which there is no ground truth—no one but you really knows how happy, anxious, scared or disgusted you are. This subjectivity makes emotions much more difficult to measure than other brain functions, especially in animals, because we cannot ask them how they feel.

The most extensively used measure of emotions is applied by simply asking people questions about their feelings, such as: “On a scale from 1 to 5, how upset are you?”. However, measures of subjective experience are limited to humans who can communicate and cannot be applied to nonhuman animals or young children. A second type of emotion measure probes biological reactions triggered by emotions, including facial expressions or galvanic skin responses; often, however, these responses do not map cleanly onto subjective reports. Other versions of physiological measures seek to quantify how emotions are reflected in dynamic patterns of brain activity (or their proxies, such as fMRI results). A final way to measure emotion is to induce an emotion and measure how it changes a person’s behavior on a well-defined task, such as a paradigm for risky decision-making. These different ways of measuring emotions set the stage for debates about what these measures should be called, centered around the evidence they provide.

To understand debates about emotion-related terms, it’s helpful to begin with textbook definitions of what those terms mean. According to one popular undergraduate psychology textbook, an emotion is an immediate, specific negative or positive response to environmental events or internal thoughts. The term is used interchangeably with affect. An emotion has three components: physiological changes (such as a galvanic skin response or pupil dilation), a behavioral response (such as a facial expression or running away) and a feeling based on cognitive appraisal of the situation and interpretation of body states. The word feeling is reserved for the subjective experience of emotion, which differs from the emotion itself. In contrast, moods are diffuse, long-lasting emotional states that do not have an identifiable trigger. Getting cut off in traffic can make a person angry (emotion), but for no apparent reason, a person can be irritable (mood). Definitions like these can help separate these concepts when studying them experimentally.

Though that backdrop might make this all sound simple and straightforward, hidden behind it is a lot of complexity. For instance, mental health researchers define mood differently, in ways that focus on what triggers its changes. In fact, so much controversy exists around terms in emotion research that more than 170 authors were inspired to unite to find the root causes of these definitional conundrums. They determined these to be differences in underlying goals and assumptions across theories of emotion, and they integrated those into a principled and cohesive framework called “The Human Affectome“—a terrific step. Still, controversy about what to call emotion-related phenomena remains. Here, I’ll describe two flavors of naming conventions in turn.

Debates about conscious experience and whether “emotion” can be studied in animals

The rationale behind the textbook definition of emotion described above follows from the idea that events can trigger “emotions” that, when sufficiently strong, can be experienced consciously as “feelings” like happiness or fear. The corollary is that emotions can exist subconsciously, an assertion that is supported by some evidence. In one study, for example, viewing emotional facial expressions subconsciously (briefly, followed by a perceptual mask) affected behavior—the amount of a pleasant beverage consumed.

According to this framework, because we cannot probe subjective experience in nonhuman animals, we can study emotions in animals—guided by well-defined criteria for what constitutes an emotion (described in more detail below)—but not feelings. One advantage of this framework is that the motivation behind studies of emotion is reflected in its description using words that anyone can relate to: analogous to the various terms for “memory” (such as “recollection memory”), emotions are described with terms such as “happiness,” “fear” and “mood.”

Some researchers, however, maintain that the term “emotion” and its flavors should be reserved for cases in which conscious experience is evident. They argue that because we cannot probe the subjective experience of animals, animals, by this definition, cannot have emotions but instead have “affective states.” One rationale behind this position is to prevent the downstream consequences that may follow from misattributing emotion-related behavior as emotional awareness. As some researchers argue, the subcortical neural circuits that support fear-related freezing behavior may be distinct from the cortical circuits that support the conscious experience of fear. As such, labeling those subcortical circuits “fear circuits” and targeting them for therapeutic interventions to relieve conditions such as post-traumatic stress disorder may be misguided.

By the same logic, the longer-lasting states called “moods” cannot be studied in animals, but researchers can study “temporally extended affective states.” One motivation behind this position is the desire to clarify what’s what, to ensure that no one gets confused, including the researchers themselves. Moreover, these terms are used in the service of an important end-goal: explaining what we consciously experience. Reserving the term “emotion” for such conscious experiences ensures that researchers keep their eye on that prize.

Notably, calls to reserve “emotion” for cases in which there is direct evidence of conscious awareness (unique to humans) set a higher bar for using the word “emotion” than researchers typically do for using terms for other brain functions. For instance, evidence for “memory” can come in all manner of forms, including how much longer a nonverbal human infant or a monkey will look at an image that they’ve never seen before versus one they have (a measure of “recognition memory”). Concerns about this higher bar for defining emotion include that it impedes communication with the public and that it disincentivizes nonhuman animal research, where arguably some of the biggest discoveries are being made. After all, you would have to be quite a neuroscience aficionado to be curious about a paper with “temporally extended affective state” (rather than “mood”) in its title.

Debates about the criteria for what counts as an “emotion” 

A second axis of the debate around what constitutes an emotion is the criteria for inclusion, even for humans (with whom subjective reports are available). Researchers generally agree that not all feelings are emotional: for instance, you can feel tired or hungry, but those are not typically considered emotions. The general idea is that emotions are triggered by external events that are real, remembered or imagined: for instance, fear is triggered by seeing a tiger. In contrast, tiredness and hunger are triggered by events that are internal to the body. Pain is considered an edge case.

When focusing on what an external event triggers, how do you decide if it’s an emotion? Does an obnoxious puff of air to your eye that triggers an unpleasant experience count? Under one set of criteria, it does. These criteria include that the experience falls on one side of a good/bad divide (valence), it’s surprising (arousal), it extends beyond the time the air puff is present (persistence) and it triggers multiple behaviors, such as eye closing and tearing (generalization).

However, not all researchers agree with those criteria. Some prefer to call the experience that follows an air puff to the eye an “affective state” and assert that the term “emotion” should be reserved for situations that involve a cognitive interpretation of external events (such as seeing a tiger) and internal body states (such as flushed cheeks). The idea, in part, is that emotions are not just triggered by stimuli, but also depend on context—seeing a tiger safely confined to its enclosure at the zoo or on a television won’t trigger fear. That said, what exactly counts as “context” is not well defined. For instance, the negative emotional experience following an air puff to the eye is not immutable, but disappears following a medical intervention (ketamine administration)—does that count as context? Either way, it’s worth noting that including context as part of the definition raises a different bar to using the word “emotion” than researchers typically apply for other brain functions, such as “perception.” In particular, some forms of visual perception can also be contextual, yet we don’t say that a bright flash of light is not perceptual merely because it always triggers the percept “bright flash of light.”

Stepping back, many debates about defining emotions boil down to different theoretical ideas about what drives them. One example is the tension between basic emotion theory and the theory of constructed emotions. In basic emotion theory, emotions (fear, disgust and so on) exist at their core as discrete categories designed to solve adaptive problems: for example, disgust follows from the fact that eating poisonous things is harmful. Proponents of basic emotion theory advocate for definitions of emotions that align with those discrete categories. In comparison, the theory of constructed emotion aligns with ideas about predictive processing. Inherent in this theory is the notion that any sensory signal can have more than one emotional meaning, depending on context. Proponents of constructed emotion theory highlight this as one of the components of emotion that must be considered when studying it. Categories are thought to be constructed ad hoc in a specific time and place (i.e., a specific context) for a particular function. In this view, the evolutionary adaptation is the ability to construct categories in the service of metabolic efficiency to control action, not the identity of the specific categories themselves (which are acquired via learning). In this view, there is not a single category for “anger,” but a distribution of spatiotemporal anger categories.

An important dimension of this conversation is thus the degree to which definitions of the phenomena of interest can and should be made independent of theories about how emotions come to be. Separation between the two has worked for some scientific problems (as when the thermometer developed in the 17th century, enabling the measurements required to converge on thermodynamics in the 19th century), but for emotion, progress may require intertwined definitions and theories, because it’s a different and more complex problem.

What do researchers think?

To better understand what researchers think, I asked various people in the field: “What should be done about the communication conundrum in emotion research to maximize communication with non-experts, minimize confusion both within and outside the field, and inspire progress?” Here’s what they had to say.

Responses

Get alerts for essays by Nicole Rust in your inbox.

Subscribe to get notified every time a new essay is published.

privacy consent banner

Privacy Preference

We use cookies to provide you with the best online experience. By clicking “Accept All,” you help us understand how our site is used and enhance its performance. You can change your choice at any time. To learn more, please visit our Privacy Policy.