How many emotions are there?

Six figures, six different questions

Verified sources support six basic families in Ekman's account, eight primary emotions in Plutchik's model, 27 varieties in Cowen and Keltner's study, and 27 emotion labels plus Neutral in GoEmotions. This site organizes 121 labels, while Russell's circumplex replaces a total with two dimensions.

The noun beside each figure matters more than the figure alone. Six refers to families, eight to primary states, and 27 to either study results or labels, depending on the source. Seventy-eight is the size of one practical vocabulary. Two is the number of axes in a different kind of model.

Why the counts differ

Emotion counts differ because researchers choose different evidence, boundaries, and jobs for their models. A theory may seek a small set of biologically basic families. An experiment may separate reported experiences statistically. A language dataset needs useful annotation labels, while a wheel needs enough words to help people choose.

Different units

Counting emotion families is like counting folders; counting emotion words is like counting the files inside them. A broad family can contain many experiences that vary in intensity, cause, bodily feeling, and social meaning. Treating a family and every narrower description as equivalent units makes a comparison look precise while changing what is being counted.

Different methods

A facial-expression theory, a self-report experiment, and a text-labeling project observe different evidence. Their results depend on the prompts, stimuli, response choices, language, and participants involved. A count derived from reactions to video clips does not automatically become the number of recognizable facial expressions or the number of words needed for written comments.

Different boundaries

Emotion categories do not always behave like boxes with hard sides. Cowen and Keltner found distinct varieties alongside continuous gradients between them. Plutchik explicitly allowed mixtures and differences in intensity. A researcher may split two nearby experiences for one purpose and group them for another without claiming that either description is the only possible one.

What six means in Ekman's model

Ekman's familiar figure is six basic emotion families, not six possible human feeling words. His 1992 paper argues that basic emotions have distinctive features and also share properties such as rapid onset and automatic appraisal. The small number identifies proposed families; it does not exhaust every variation within them.

This distinction is easy to lose when the model is reduced to a poster or search snippet. Ekman wrote about families because related emotional states can differ while retaining common features. A narrow label and its broader family are not two rival answers to the same level of classification.

Ekman discussed additional candidates in later work, but expanded lists circulated online do not all represent one equivalent, fixed taxonomy. This page therefore uses the six-family figure tied directly to the verified 1992 paper and does not assign a second total to Ekman without a comparably clear paper-based definition.

What eight means in Plutchik's model

Plutchik proposed eight primary emotions in a psychoevolutionary model, arranged as four pairs of opposites. The 1980 chapter also treats other emotions as mixtures or derivatives of primary states. Eight is therefore the model's starting structure, not a claim that only eight named emotional experiences can occur.

The circular arrangement carries information that a plain list omits. Neighboring primary states are similar enough to combine, opposites sit across from one another, and intensity can vary. Once combinations and intensity enter the model, the vocabulary extends well beyond its eight primary positions. Counting every possible derivative would answer a new question.

Plutchik's chapter is also explicit about the status of the primary emotions: they are theoretical constructs within a model of adaptive responses. That makes the figure useful for understanding the model, but it should not be presented as a census of everything a person can ever feel.

Where the 27 figure comes from

Cowen and Keltner found 27 distinct varieties of reported emotional experience in responses to 2,185 short videos. Their 2017 study used US participants and English emotion concepts, and the authors called the result an approximation rather than a definitive taxonomy. The figure belongs to that evidence and method.

Participants described their reactions through several self-report methods. Statistical analyses found 27 reliably separable patterns, but the study also found fuzzy boundaries and gradients between many categories. So the result is not a picture of 27 sealed compartments. It is a map with distinguishable regions and smooth transitions.

GoEmotions is often folded into this discussion, although it asked a different question. The 2020 dataset contains English Reddit comments labeled with 27 emotion categories or Neutral. Its paper says "27 + neutral," not 25 emotions. That taxonomy was designed for annotating language and training computational models, not for declaring a universal number of human emotions.

Why this feelings wheel has 121 labels

MyFeelings has 121 labels because its taxonomy contains 7 core, 39 secondary, and 75 specific entries across three rings. That number describes the vocabulary available in this particular interface. It is a design count, not a scientific verdict about how many emotions exist or how every wheel must be built.

Six of this wheel's seven cores are Ekman's basic six; the seventh, Bad, is not his and comes from the MyFeelings app's own vocabulary. Its three-ring shape — a small core set expanding outward to more specific words — follows the method of Gloria Willcox's 1982 Feeling Wheel, but not her vocabulary: her six are mad, sad, scared, joyful, powerful and peaceful, and her wheel holds 72 feelings. MyFeelings does not claim that Willcox published this exact 121-label taxonomy, or that she chose these cores. The DOI in the table verifies the source form; the 121 comes from the site's own imported taxonomy.

A wheel has a practical constraint that a theory does not: its labels must remain readable and navigable. Adding every emotion word would make selection harder, while using only broad families could leave useful distinctions unavailable. The core emotion page explains how the center works without repeating the taxonomy here.

You can browse the complete feelings wheel, use the words to describe feelings page when you want a list, or choose a printable feelings wheel when paper suits the moment better. The count stays attached to this tool, not to emotional life as a whole.

Russell's answer is a map, not a count

Russell's circumplex model does not need a final list of emotions. It places affective terms in a circular space defined by two bipolar dimensions: pleasure-displeasure and degree of arousal. New words can be located within that space, so the model organizes relationships without fixing how many labels are allowed.

The 1980 paper tested this structure using emotion-denoting adjectives and self-reported affective states. Its circle preserves both similarity and opposition: nearby terms occupy nearby positions, while contrasting terms fall farther apart. The model's key figure is two dimensions, but those dimensions are coordinates, not two emotions.

This approach answers the number question by challenging its premise. If affect can vary continuously along two axes, the number of named regions depends on how finely a language, study, or tool divides the space. Discrete categories can still be convenient; the circumplex simply does not require them to be the basic units.

Which number should you use?

Use the number that matches the task and name the model beside it. Six or eight can summarize a basic-emotion theory, 27 can report a specific study or dataset, and 121 can describe this site's wheel. For personal reflection, the best count is simply enough words to find an honest fit.

Precision comes from stating the unit, source, and scope, not from choosing the largest or smallest figure. "Cowen and Keltner identified 27 varieties in one study" is informative. "Science says there are 27 emotions" discards the study design and turns a bounded result into a universal claim the authors did not make.

More labels can help when they capture distinctions you genuinely notice. They can also create false precision when the experience is still unclear. The emotional granularity page covers specificity itself, while the feelings wheel for kids offers a vocabulary shaped for younger readers.

If your real question is how to record experiences over time, counts matter less than whether the method fits. The mood tracker alternativepage compares that choice directly. A model is most useful when its level of detail suits the decision in front of you.

Sources and limits

Every research citation on this page was checked against Crossref for DOI, title, first author, and year. The linked papers support the figures in context, but none establishes a final universal total. The comparison deliberately omits popular counts that could not be tied cleanly to a verified primary source.

1. Paul Ekman, "An argument for basic emotions," Cognition and Emotion 6, nos. 3–4 (1992): 169–200. doi:10.1080/02699939208411068
2. Robert Plutchik, "A General Psychoevolutionary Theory of Emotion," in Theories of Emotion (1980): 3–33. doi:10.1016/B978-0-12-558701-3.50007-7
3. Alan S. Cowen and Dacher Keltner, "Self-report captures 27 distinct categories of emotion bridged by continuous gradients," Proceedings of the National Academy of Sciences114, no. 38 (2017): E7900–E7909. doi:10.1073/pnas.1702247114
4. Dorottya Demszky et al., "GoEmotions: A Dataset of Fine-Grained Emotions," Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics(2020): 4040–4054. doi:10.18653/v1/2020.acl-main.372
5. Gloria Willcox, "The Feeling Wheel," Transactional Analysis Journal 12, no. 4 (1982): 274–276. doi:10.1177/036215378201200411
6. James A. Russell, "A circumplex model of affect," Journal of Personality and Social Psychology 39, no. 6 (1980): 1161–1178. doi:10.1037/h0077714