Perspective-Taking Versus Perspective-Getting — The Difference Matters

The distinction between perspective-taking and perspective-getting was crystallized in a 2018 paper by Nicholas Epley and Tal Eyal published in the Journal of Experimental Social Psychology, though the research tradition behind it extends further.

Epley and Eyal's central finding was counterintuitive: instructing people to take another person's perspective — to imagine the world from the other's viewpoint — reliably increased their confidence in their understanding of the other person but did not reliably increase their accuracy. In some experimental conditions, perspective-taking actually decreased accuracy relative to a control condition. People told to imagine another person's experience became more certain they knew what that person was experiencing, while the gap between their imagined version and the other person's actual reported experience remained constant or widened.

The mechanism Epley and Eyal identified: when you imagine another person's perspective, you anchor to your own current state and adjust insufficiently. This is the standard finding from judgment and decision-making research — people adjust from anchors, but rarely far enough. If you are in a comfortable, warm environment, and you try to imagine how someone in a cold environment feels, your imagination generates a cold-adjusted version of your current experience — which is systematically less cold than the actual cold the other person is experiencing.

This "insufficient adjustment" problem is compounded in social contexts by a specific bias: we assume other people's inner states are more visible to us than they actually are. Nicholas Epley's broader research program on "mind reading" — summarized in his book Mindwise — documents the consistent finding that people overestimate their accuracy in understanding others' thoughts, feelings, and motivations. The confidence with which we believe we understand other people is substantially higher than the accuracy with which we actually understand them. And perspective-taking — because it generates the feeling of empathy without requiring contact with the actual other person — amplifies this overconfidence.

Perspective-getting, as a practice, has a specific structure. It is not simply asking questions. It is asking questions in a mode that prioritizes their answer over your prediction.

This is harder than it sounds because the standard conversational mode is not actually inquiring — it is confirming. People in conversation constantly ask questions whose primary function is not to learn the other person's view but to see whether the other person shares their own. "Don't you think...?" "Wasn't that terrible...?" "Can you believe...?" These questions have the grammatical form of inquiries and the actual function of invitations to agree.

Real inquiring questions — questions in the perspective-getting mode — have a different quality. They are genuinely open. The asker does not have a preferred answer. They are designed to generate information the asker does not already have, rather than to confirm information the asker already believes they have. They follow the other person's lead rather than the asker's agenda. And they remain open to answers that disconfirm the asker's existing model.

Sara Lawrence-Lightfoot, in her book Respect, describes what she calls "thick description" of other persons — the careful, patient, observational and dialogic practice of actually capturing another person's reality in their own terms, without immediately translating it into familiar categories. She developed this practice in the context of qualitative research, but her description of what it requires — suspension of the researcher's framework, genuine curiosity about the other's terms, willingness to be surprised — is a precise description of what perspective-getting demands at the level of ordinary conversation.

The structural requirements of perspective-getting, operationalized:

First: ask open questions before drawing conclusions. When you notice yourself forming a judgment or inference about another person's experience, convert it to a question before acting on it. "I assumed you were angry — were you?" "I thought this was hard for you — what was it actually like?"

Second: treat their answer as more reliable than your prediction. This seems obvious. It is violated constantly. People who ask "how are you actually feeling about this?" and receive an answer that conflicts with their expectation frequently treat their expectation as more reliable. "I know you say you're fine, but I can tell you're upset." Sometimes they're right — people don't always accurately report their own states. But the default should be that their account takes precedence.

Third: update your model. After receiving the other person's account, consciously revise the model you had of their experience before you asked. This is the step most consistently skipped. People hear the answer, nod, and continue operating from the pre-asking model. The update must be deliberate.

The gap between imagined and actual perspective is not random. Research on this gap identifies consistent systematic errors — directions in which perspective-taking reliably diverges from the other person's actual experience.

The "egocentric anchor" bias is the most foundational: we start from our own current experience and adjust, and the adjustment is insufficient. The findings replicate across domains. In negotiations, people who are told to imagine their counterpart's perspective do not improve their ability to predict their counterpart's actual preferences and priorities — but they feel more confident that they understand them. In medical contexts, physicians who try to imagine patient distress consistently underestimate it — a robust finding across multiple studies comparing physician assessments of patient-reported quality of life.

The "visible features" bias compounds the egocentric anchor. When people try to imagine another's perspective, they give disproportionate weight to features of the other person's situation that are visible and salient, and underweight features that are internal or invisible. The obvious difficulties — the visible disability, the obvious loss — are easier to imagine than the invisible ones: the chronic low-grade humiliation, the effort that's invisible because it's been normalized, the history that gives current events their meaning.

Jacqueline Mattis and colleagues' research on empathy across cultural and racial lines adds a further complication: the egocentric anchor is worse when the imaginer and the imagined come from different groups. The imaginer's baseline is less applicable, the visible features are more likely to be misread, and the invisible features are more likely to remain invisible. Perspective-taking across significant social distance doesn't just fail to help — it can actively generate confident misunderstanding.

This is where perspective-taking becomes a political problem, not just a personal one. Well-meaning people in privileged positions imagining the experiences of people in marginalized positions and concluding that they understand — while those whose experiences have been "understood" are reporting that they feel unseen — is a recognizable and common dynamic. The problem is not lack of good intent. The problem is the epistemic move: projecting and concluding rather than asking and updating.

The case against perspective-taking is not absolute. There are conditions under which it improves understanding, and it is worth being precise about them.

Perspective-taking works reasonably well when the imaginer and the imagined are very similar — when the egocentric anchor is actually close to the other person's baseline. It also works better for predicting behavior than for understanding inner states: imagining what someone in a given situation might do can produce useful predictions even when the imagined feeling-states are inaccurate.

Perspective-taking also serves functions other than accuracy. It activates prosocial motivation — the tendency to care about and act in the interest of another person. Research by Adam Galinsky and colleagues has shown that perspective-taking increases helping behavior, reduces in-group favoritism, and shifts resource allocation toward the imagined other. Even if the imagined perspective is not perfectly accurate, the act of imagining someone else's experience produces a motivational shift that is itself valuable.

The integration of these findings suggests a practice that is neither pure perspective-taking nor pure perspective-getting, but a sequence: perspective-taking first (to activate concern and motivation), followed by perspective-getting (to calibrate the imagined picture against actual reality). The concern activated by imagination is the motivation to ask. The asking then provides the accuracy.

Daniel Ames's research on negotiation adds a useful nuance. He found that perspective-taking advantages show up most in initial encounters, where there is no relationship history to draw on and imagining the other person's likely position can help calibrate opening moves. Perspective-getting advantages show up most in ongoing relationships, where repeated interaction provides opportunities to update and where the cost of confident misunderstanding accumulates over time. The lesson: use imagination to get started, use inquiry to get accurate.

Underlying the shift from perspective-taking to perspective-getting is an epistemological commitment: the acknowledgment that other people's inner lives are not directly accessible to you.

This sounds obvious. It is regularly violated.

The feeling of knowing what someone else is experiencing is phenomenologically indistinguishable from the experience of actually knowing it. When your imagination generates a picture of another person's inner state — when you "see" that they're hurt, or angry, or lonely — it doesn't come labeled as a projection. It comes with the experiential quality of direct perception. The philosopher Charles Taylor calls this "reading" other persons — and he notes that the reading can be so fluent and so automatic that it is experienced as transparent perception rather than interpretation.

The move that perspective-getting requires is treating your reading of another person as a hypothesis rather than a conclusion. To say, at least to yourself: this is my best current guess, not a known fact. And to act accordingly — which means being genuinely willing to discover you're wrong.

Timothy Wilson's research on introspection, summarized in Strangers to Ourselves, adds a complicating layer. People are not always accurate reporters of their own mental states. Introspective reports are constructed rather than read off directly from an inner display. This means that perspective-getting, while generally superior to perspective-taking, is not simply a matter of asking and receiving transparent truth. The other person's account is more reliable than your imagination, but it is itself subject to the limits of self-knowledge. The skilled practitioner of perspective-getting holds this complexity: taking the other person's account seriously as the most reliable available data while remaining open to the possibility that the account itself is incomplete.

This is not an argument for dismissing what people report. It is an argument for listening carefully enough to hear the layers — including the parts of the experience that the person hasn't fully articulated to themselves, which sometimes surface through good listening precisely because they hadn't been formulated until there was someone capable of receiving them.

The application of perspective-getting varies significantly by context. Several contexts deserve specific treatment.

In personal relationships, perspective-getting is disrupted by the assumption of familiarity. Long-term partners, close friends, and family members frequently stop asking because they believe they already know. This is one of the more reliable predictors of relational deterioration: not conflict, but the slow ossification of each person's model of the other into something that no longer tracks who the other person actually is. The person continues to evolve; the partner's model of them does not. Couples who maintain the habit of genuine inquiry — not performing interest but actually treating each other as sources of information about themselves — maintain a quality of contact that protects against this.

In professional contexts, perspective-getting is particularly valuable across hierarchy. Research on "the knowledge curse" (also called the "curse of knowledge") shows that people with more expertise in a domain have increasing difficulty imagining what it is like not to have that expertise. Managers who rose through technical ranks consistently overestimate how much their less-experienced team members understand; physicians overestimate how much their patients follow medical reasoning; teachers overestimate how accessible their explanations are to students who lack their conceptual scaffolding. Perspective-getting — actually asking "what's confusing about this?" rather than imagining what might be confusing — provides the information that imagination cannot.

In cross-cultural contexts, perspective-getting is not just preferable to perspective-taking; it is the only responsible practice. The alternative — members of one culture imagining the inner lives of members of another culture and treating their imagination as understanding — has a well-documented history of producing harmful confident errors: colonial missionaries certain they understood the spiritual needs of the people they were converting, development workers certain they understood the economic priorities of the communities they were serving, researchers certain they understood the psychological makeup of groups whose baseline assumptions differed radically from their own. The record here is not ambiguous.

Richard Shweder's work on cultural psychology demonstrates that the most fundamental categories through which experience is organized — self, agency, cause, emotion, time, the body — are culturally variable. You cannot accurately imagine the experience of someone operating with fundamentally different organizational categories by working from your own. The structure within which their experience is constructed is different from the structure within which yours is constructed. This is not an argument against cross-cultural empathy. It is an argument that cross-cultural empathy requires actual engagement — perspective-getting — rather than the extension of your own categories to cover their situation.

The Law 1 claim — that shared humanity, fully recognized, is the substrate for resolving the problems we call intractable — depends on people actually understanding each other. Not imagining they understand each other. Actually understanding.

The history of failed peace processes, failed negotiations, failed reconciliation efforts is in part a history of perspective-taking deployed in place of perspective-getting. Parties who believed they understood the other side's position because they had imagined it — and who had never actually asked in a mode capable of receiving the answer — negotiating past each other, sometimes for decades.

Stuart Eizenstat's account of the Dayton Accords negotiations emphasizes the turning point at which mediators stopped theorizing about the parties' interests and created conditions in which the parties themselves could articulate what they needed directly. The actual peace agreement turned on specifics that none of the imaginative modeling had captured. The only way to get those specifics was to ask, and to create conditions in which the asking could receive honest answers.

At the personal scale, the aggregate of millions of perspective-gettings — millions of conversations in which people treated each other as the authoritative source of information about their own experience — produces a very different texture of social reality than the aggregate of millions of perspective-takings. In one world, people feel known. In the other, they feel categorized. Feeling known changes how people engage with their communities, their institutions, and people different from themselves. It's not guaranteed to produce cooperation, but it creates the conditions in which cooperation becomes possible.

The central premise of Law 1 is not sentiment. It's structural. If every person treated every other person as the authoritative source of information about their own inner life — if perspective-getting replaced perspective-taking as the default mode of human social interaction — the information necessary to actually solve collective problems would become available. We would stop misallocating resources toward imagined needs and start allocating them toward stated ones. We would stop designing institutions for the people we imagine and start designing them for the people who are actually there.

That's not a utopia. That's an upgrade. And it starts with one conversation in which someone decides to ask rather than assume.

The Prediction-Then-Ask Practice: Before a significant conversation, write down your best prediction of what the other person is thinking, feeling, or wanting from the conversation. Then have the conversation, and ask directly enough to actually find out. After, compare. This surfaces the gap between your imagined version and their actual account, and trains the habit of holding predictions as predictions rather than conclusions.

The Confirming-Question Audit: For one week, notice every question you ask in conversation. Classify each question: Is it an open question designed to learn something I don't know? Or is it a confirming question designed to see if the other person agrees with me? The ratio will be informative.

The Surprise Tracker: After any significant conversation, ask yourself: did I learn anything that surprised me — that contradicted or complicated my prior model of this person? If the answer is consistently no, you're probably confirming rather than inquiring. Real conversations with real people consistently produce surprises. If you're not getting surprises, you're probably not actually getting their perspective.

The "What Was That Like?" Question: Practice asking this single question more often. Not "how are you?" (which receives a social script) and not "that must have been hard" (which is your interpretation offered for them to confirm). "What was that like for you?" returns the authority over the experience to the person who had it.

The Update Ritual: After any conversation in which you received significant information about another person's experience, explicitly update your internal model of them. Articulate to yourself what you believed before and what you now know is different. This makes the update deliberate rather than passive, and it reinforces the habit of holding models as provisional rather than settled.

The distinction between perspective-taking and perspective-getting is not about good intentions. The people who project their imagination onto others and treat it as understanding are not bad people. They are people using an available tool — a tool that feels like empathy, that produces the emotional warmth of empathy — without having checked whether it's producing the accuracy of empathy.

The upgrade is simple to describe and genuinely difficult to practice: put the imagination in service of inquiry rather than in place of it. Use what you imagine to generate the question. Use the question to generate the actual data. Use the actual data to see the actual person.

That's the practice. The rest follows.