Empathy Is Not a Solo Act: Why Understanding Others Takes Two Minds

For decades, psychologists have measured empathy the way they measure height or intelligence: as a trait residing inside one person. Questionnaires, brain scans and behavioral tests have all probed the empathizer, the individual who perceives, understands and responds to another’s emotional state. But a sweeping new review published in Nature Reviews Psychology argues that this individualistic framing has quietly distorted the science. Empathy, according to Shir Genzer and Anat Perry of the Hebrew University of Jerusalem, is not something one person does to another. It is a dyadic process, jointly constructed by two partners whose cues, interpretations and responses continuously reshape one another across the course of an interaction. The shift in framing, the authors contend, has profound implications for how empathy is studied, measured and even engineered into artificial intelligence.

The review begins from a deceptively simple observation: empathy in the real world almost never unfolds in isolation. Everyday experience-sampling studies, such as work by Depow, Francis and Inzlicht published in Psychological Science in 2021, show that empathic moments punctuate daily life constantly, embedded in conversations, glances and shared silences. Yet the dominant laboratory paradigms strip away precisely what makes those moments work. Participants view static photographs of eyes or standardized facial expressions, or read vignettes about strangers, and their accuracy is scored against a fixed answer key. The target of empathy, in these designs, is a stimulus rather than a person, and the reciprocal feedback loop that defines real empathic exchange is severed before the experiment begins.

Genzer and Perry organize their dyadic account around three classic components of empathy: cognitive, affective and motivational. Each, they argue, looks fundamentally different when viewed from both sides of the exchange. On the cognitive side, empathic accuracy, the ability to infer another person’s thoughts and feelings, has traditionally been treated as a skill of the perceiver. But landmark work by Zaki, Bolger and Ochsner in 2008 demonstrated that accuracy depends on the target as well: people differ dramatically in how expressively they convey their inner states, and an expressive partner can make even a modest perceiver look sharp. The now-classic title of that paper, ‘It takes two,’ anticipated the review’s central thesis by nearly two decades, yet the field has been slow to follow through.

The affective component reveals the dyadic logic even more vividly. Physiological synchrony, the moment-to-moment coupling of heart rate, skin conductance and neural activity between interaction partners, has emerged as one of the most striking signatures of shared emotion. Hyperscanning studies, in which two brains are recorded simultaneously, show inter-brain correlations during cooperative tasks, emotional disclosure and storytelling. A 2026 fNIRS hyperscanning study by Yu and colleagues found that emotional disclosure enhances interpersonal closeness through coupled neural activity between speaker and listener. Crucially, the review notes, synchrony is not simply empathy read off one nervous system; it is an emergent property of the exchange, shaped by mood congruency, relationship intimacy and the sensory channels through which partners connect.

Motivation, the third component, may be the most consequential for the dyadic view. Empathy is not an automatic reflex but a costly, chosen activity: people frequently avoid it because of its cognitive demands, as Cameron and colleagues showed in 2019. Whether empathy is deployed depends heavily on the relationship context and on what the empathizer expects in return. Conversely, the recipient’s perception matters enormously. Studies distinguishing empathic accuracy from perceived empathic effort, including work by Cohen, Schulz, Weiss and Waldinger in 2012, suggest that feeling that a partner is genuinely trying to understand can matter as much as being accurately understood. Winczewski, Bowen and Collins extended this in 2016, showing that empathic accuracy facilitates responsive, caring behavior in couples only when it is paired with the motivation to use it.

This two-sided structure explains a puzzle that individual-difference approaches have struggled with: why empathic accuracy sometimes fails to predict relationship satisfaction. Meta-analytic work by Sened and colleagues in 2017 found only modest associations overall, with effects hinging on context, conflict and whether the perceiver’s understanding was actually communicated. The dyadic framework resolves this by treating empathy as a signal that must be sent, received and acknowledged. A perceiver who accurately decodes a partner’s distress but responds coldly has not completed the empathic transaction. Recent work on perceived empathic effort, such as a 2024 study by Le and Reis in Emotion, shows that people are more willing to express emotions when they believe their partner is trying hard to understand them, regardless of perfect accuracy.

The review also elevates a long-neglected dimension: positive empathy, the capacity to share in another’s joy rather than merely their pain. Research by Morelli, Lieberman and Zaki, and more recently by Depow, Gordon, Gu and Oveis in 2026, indicates that responding enthusiastically to good news, a process known as capitalization, builds bonds in ways that distress-focused empathy cannot. Andreychik’s 2019 findings suggest that empathy for a partner’s positive emotions independently strengthens relationship quality. Positivity resonance, a concept developed by Barbara Fredrickson, describes the momentary micro-experience of shared positive affect, connection and synchrony that emerges when two people genuinely co-experience an emotion. These shared uplifting moments, the review argues, are quintessentially dyadic: they cannot be generated by one person alone, no matter how empathic that person is.

The framework arrives at a politically charged moment, as artificial intelligence systems increasingly simulate empathic language. Chatbots now produce responses that third-party evaluators sometimes rate as more compassionate than those of expert humans, as Ovsyannikova, Oldemburgo de Mello and Inzlicht reported in 2025. Yet studies consistently show that people devalue identical supportive messages when labeled as AI-generated, a finding replicated by Yin, Jia and Wakslak in PNAS in 2024 and by Rubin and colleagues in Nature Human Behaviour in 2025. Perry’s own theoretical work frames empathy as a predictive signal: its value lies partly in knowing that another human mind has invested effort in understanding you. If empathy is inherently dyadic, a jointly constituted experience of togetherness, then a system that generates fluent empathic language without shared experience may be producing the form of empathy while missing its relational substance.

Genzer and Perry close with a research agenda that reads as a critique of the field’s methodological habits. They call for studies that move beyond stranger-based paradigms to examine dyadic empathy across relationship types, from romantic partners and friends to parent-child dyads and clinical encounters, and across emotional valences, including the positive emotions that have been chronically understudied. They advocate longitudinal, process-sensitive designs, such as intensive daily-diary and ecological momentary assessment methods, that can track how empathic interactions evolve as relationships develop, rather than snapshotting them at a single point. Techniques like hyperscanning, wearable physiology and second-person neuroscience, in which participants genuinely interact rather than observe, offer tools for capturing the reciprocal dynamics the framework describes. The deeper message is conceptual: empathy is not a possession but a performance, co-authored in real time by two nervous systems. Understanding it fully will require science to study the space between people, not just the minds on either side of it.

Subject of Research: Dyadic and interpersonal processes of empathy in social interaction

Article Title: Empathy as a dyadic process

Article References: Genzer, S., & Perry, A. (2026). Empathy as a dyadic process. Nature Reviews Psychology. https://doi.org/10.1038/s44159-026-00624-4

Image Credits: AI Generated

DOI: 10.1038/s44159-026-00624-4

Keywords: empathy, dyadic processes, empathic accuracy, social neuroscience, positivity resonance, interpersonal emotion regulation, hyperscanning, physiological synchrony, relationships, AI empathy, psychology, Nature Reviews Psychology