# Perspective Taking

> Framework for deep perspective-taking and understanding how others think, feel, and make decisions.

- Skill: `deepnoodle-ai/perspective-taking` (Agent Skill)
- Install (CLI): `npx skillmds@latest add deepnoodle-ai/perspective-taking`
- Raw SKILL.md: https://api.skillmd.com/api/skills/deepnoodle-ai/perspective-taking/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: deepnoodle-ai (https://skillmd.com/u/deepnoodle-ai)
- Updated: 2026-09-10
- Page: https://skillmd.com/skills/deepnoodle-ai/perspective-taking

---


# Inhabiting the Logic: A Framework for Deep Perspective-Taking

A synthesis of research-backed methods for genuinely understanding how others think, feel, and make decisions.

## The Core Insight

Perspective-taking is not one skill—it's several distinct cognitive operations that can be trained independently. The goal isn't to *agree* with someone, but to reconstruct their internal logic so thoroughly that their behavior becomes *obviously correct* from their vantage point.

## Foundation: Two Types of Understanding

Research shows empathy and perspective-taking recruit distinct neural circuits:

| Type                  | Definition                                        | Key Question                                    |
| --------------------- | ------------------------------------------------- | ----------------------------------------------- |
| **Affective Empathy** | Feeling what others feel                          | "What emotions are they experiencing?"          |
| **Cognitive Empathy** | Understanding *why* they think and act as they do | "What logic makes this choice obvious to them?" |

Both matter, but cognitive empathy is the trainable skill. Notably, the research suggests this is more about *choosing to engage* than innate ability—motivation matters more than capacity.

## The Seven Operations

### 1. Map the Contradictions

Adapted from design thinking's Empathy Map. Plot what you observe across four dimensions:

- **Says**: What they explicitly state
- **Thinks**: What they're actually thinking (often different)
- **Does**: Observable behaviors and actions
- **Feels**: Underlying emotions

**The insight lives in the gaps.** When someone says one thing but does another, or thinks something they won't say aloud—that's where understanding deepens. These contradictions are "treasure maps" to genuine understanding.

### 2. Reconstruct Their Narrative

People don't just have experiences—they construct stories about who they are. Narrative identity theory shows that individuals integrate life experiences into an evolving story that provides unity and purpose.

**Ask yourself:**
- What's the story they tell about themselves?
- Who are the characters? (heroes, villains, helpers)
- What are the turning points?
- Is this a redemption arc or a contamination sequence?
- What themes keep recurring?

The narrative someone constructs about their past shapes what they see as possible in their future.

### 3. Model the System

People don't act in isolation. They're embedded in systems with constraints, incentives, and feedback loops.

**Map:**
- What are they optimizing for?
- What resources do they actually have?
- What are the real consequences they face?
- Who are the stakeholders affecting their choices?
- What second-order effects are they accounting for?

Often what looks like a personality trait is actually a rational response to circumstances you haven't fully considered.

### 4. Ladder to Values

The laddering technique excavates mental models by climbing from surface behaviors to core beliefs.

**The Chain:**
```
Attributes (what they do/prefer)
    ↓ "Why is that important to you?"
Consequences (functional/emotional outcomes)
    ↓ "Why does that matter?"
Core Values (fundamental beliefs driving behavior)
```

Keep asking "why is that important?" until you reach bedrock—the values that don't reduce further. This reveals the unconscious drivers of behavior that people often can't articulate directly.

### 5. Surface the Governing Variables

From double-loop learning: most people operate by adjusting actions based on feedback without questioning the assumptions driving those actions.

**Identify:**
- What are they *not* questioning?
- What do they assume is fixed that might not be?
- What's the gap between their "espoused theory" (what they say drives them) and their "theory-in-use" (what actually does)?

People often can't articulate why they do what they do. You have to infer the governing variables from patterns in their behavior.

### 6. Steelman Their Position

Construct the *strongest possible version* of their argument—better than they could articulate it themselves.

**The Process:**
1. Put your views aside entirely
2. Identify the best evidence and reasoning supporting their position
3. Articulate it more clearly than they did
4. Only then engage with the argument

**The Test:** Can you pass the Ideological Turing Test? If you can state their views as clearly and persuasively as they could, you understand them. If not, you're still projecting.

### 7. Find the Invisible Context

Everyone is living in the middle of a story you walked into late.

**Explore:**
- What happened before you arrived?
- What past experiences shaped their current filters?
- What are they worried about that they haven't said?
- What would you have had to experience to see the world the way they do?

## Practical Application: The OARS Stance

From motivational interviewing—a posture for conversations that surfaces genuine understanding:

| Technique          | Purpose                                               |
| ------------------ | ----------------------------------------------------- |
| **Open Questions** | Draw out experiences and perspectives without leading |
| **Affirmations**   | Recognize strengths and validate experiences          |
| **Reflections**    | Mirror back what you hear to confirm understanding    |
| **Summaries**      | Synthesize and check your mental model                |

The stance is *evocative*, not prescriptive—drawing out what's already there rather than imposing your model.

## The Integration: A Working Process

**Step 1: Observe and Map**
- Gather data across Says/Thinks/Does/Feels
- Note contradictions and tensions

**Step 2: Understand the Context**
- Model the system they're operating in
- Identify constraints, incentives, stakeholders

**Step 3: Excavate the Structure**
- Ladder from behaviors to consequences to values
- Surface governing variables and assumptions

**Step 4: Reconstruct the Narrative**
- What's the story they tell about themselves?
- What role do they cast themselves in?

**Step 5: Validate Through Steelmanning**
- Can you argue their position better than they can?
- If not, return to earlier steps

**Step 6: Locate the Emotional Truth**
- With all context in place: what does it *feel like* to be them?
- Not sympathy ("I feel bad for them") but genuine inhabitation

## Key Principles

**Separate understanding from agreement.** You can fully comprehend why someone believes something without sharing that belief.

**Contradictions are data, not failures.** When someone's words and actions don't match, that's information about the complexity of their situation.

**Behavior is usually rational given constraints you don't see.** Before attributing actions to character, exhaust situational explanations.

**People can't always articulate their own logic.** The most important drivers are often unconscious. You have to infer them.

**The goal is inhabitation, not observation.** You're not studying someone from outside—you're trying to see through their eyes.