Think about the last time you laughed so hard your stomach hurt.
Now think about the quiet warmth of holding someone you love, or the fascination of losing yourself in a good documentary.
All of these feel good, but in completely different ways. Yet for decades, psychology has often lumped them together under one umbrella: happiness.
A new data-driven study challenges that simplicity.
Using over 3,500 personal narratives of joyful moments, researchers applied computational text analysis to ask a deceptively simple question: what does it really mean to feel good?
Key Points
- Researchers analyzed over 3,500 real-life stories to map how people describe 22 different positive emotions.
- A machine learning model found four emotions that stood out as truly distinct: amusement, interest, lust, and tenderness.
- These emotions differed in arousal, involvement, and social connection—suggesting we “feel good” in more than one way.
- The study shows that language patterns can reveal the subtle architecture of our emotional lives.
- Understanding these nuances could help therapists and wellbeing researchers target different kinds of positive emotion in practice.
Listening to How People Talk About Joy
Instead of relying on rating scales or questionnaires, the research team asked 165 volunteers to describe real experiences of 22 different positive emotions, such as pride, awe, gratitude, and amusement.
Each story was recorded, transcribed, and fed into a machine learning model that searched for patterns in word use and meaning.
This “semantic space” approach looks for how words naturally cluster together in high-dimensional space—essentially letting the data show which emotions are most distinct in how people talk about them.
It’s a little like mapping constellations of feeling across language: where certain emotional stars – words like funny, adorable, or thrilling – cluster together, they reveal the hidden structure of our positive inner life.
Four Emotions That Stand Apart
When the algorithms grouped the narratives, a clear pattern emerged.
Among 22 candidate emotions, only four consistently formed distinct clusters:
- Amusement — joyful laughter and play.
- Interest — curiosity and mental engagement.
- Lust — high-arousal desire and attraction.
- Tenderness — gentle affection and warmth.
Each had a unique psychological signature.
Amusement was highly positive but moderately arousing—like a shared laugh that bonds people together.
Interest was low in arousal and personal involvement, suggesting a more detached curiosity—engaged but calm.
Lust, by contrast, showed the highest arousal and personal involvement, a full-body immersion of emotion.
Tenderness stood out for its low arousal but high intimacy, reflecting moments of closeness and care.
Together, they sketch a map of pleasure that ranges from electric energy to quiet connection.
The Language of Our Inner World
Language is more than a mirror of feeling, it actively shapes how we perceive emotion.
When people narrate experiences, they reveal the mental categories through which they understand life.
Traditional research often used coders to classify emotional text, but this brings bias, one person’s “joy” might be another’s “pride.”
By using unsupervised machine learning, this study bypassed those filters, allowing the patterns in language itself to tell the story.
This computational approach is part of a growing trend in affective science: letting words, not theories, reveal emotional structure.
The finding that only a handful of positive emotions emerge as clearly distinct suggests that our minds don’t always divide “good feelings” as finely as our vocabulary might imply.
A Map, Not a Final Answer
Interestingly, other emotions—like awe, gratitude, and inspiration—showed more blurry boundaries.
They often overlapped in meaning, hinting at complex blends rather than neatly separated feelings.
This fuzziness isn’t a flaw—it reflects how human emotion really works.
Our inner life doesn’t live in tidy boxes but in gradients. Feeling moved by a sunset and feeling awe at a concert may activate similar emotional networks, differing only in degree or context.
The study’s participants were young Dutch adults, so these emotional contours might look different in other cultures or age groups.
Still, the approach opens a powerful new window into how emotion language reflects lived experience.
Why It Matters
Understanding the architecture of positive emotion isn’t just academic—it’s deeply practical.
Different kinds of positive emotion may nourish mental health in different ways.
For example:
- Amusement can relieve tension and foster social bonding.
- Interest fuels learning, creativity, and engagement—key ingredients for resilience.
- Lust and physical desire connect us to bodily vitality and attachment.
- Tenderness promotes empathy and caregiving, building secure emotional ties.
In therapy and wellbeing programs, this means “feeling good” isn’t one-size-fits-all.
A mindfulness exercise might boost tenderness; a curiosity challenge might cultivate interest; humor could lift amusement.
Clinicians and coaches can draw on this emotional palette to help clients rediscover different shades of pleasure—especially after depression or burnout, when life’s colors often fade to gray.
The Takeaway
Happiness isn’t a single emotion—it’s a constellation of experiences.
By analyzing the language of real people’s memories, this study paints a richer picture of what “feeling good” really means.
Sometimes joy bursts out as laughter.
Sometimes it hums quietly as tenderness.
And sometimes it leans forward in curiosity.
Each is a different way our minds remind us: being alive can feel many kinds of good.
Reference
Kamiloğlu, R. G., Türkmen, İ. U., Sarnıç, T. E., Landman, D., & Sauter, D. A. (2025). What makes us feel good? A data-driven investigation of positive emotion experience. Emotion, 25(1), 271–276. https://doi.org/10.1037/emo0001417
