I’ve been thinking about the name “Artificial Intelligence,” and the more I sit with it, the less it holds up.

We named this technology after what we think it is — intelligent. But almost every conversation I hear about it describes what it does: it writes drafts, it builds websites, it analyzes data, it generates images, it autocompletes your sentences when you’re too tired to finish them yourself.

If we named it by what it actually does, we’d call it something like Automated Labor. Or Digital Task Completion. Or, if we’re being honest about most people’s Monday morning use case, Fancy Autocomplete That Sometimes Lies.

Drop the “Artificial” entirely, and you could call it Digital Assistance. Automated Pattern Matching. Computational Draft Generation. Each of those names is more accurate than “intelligence” — but none of them would have attracted $100 billion in venture capital, so here we are.

The name matters because it front-loads an answer to a question nobody has agreed on. Call it “intelligence,” and every conversation that follows inherits an assumption: this thing thinks. Now half the room is excited, and the other half is terrified, and both reactions are responses to the name, not the tool. The person who built a website with it in seventeen days and the manager who thinks it’s a crutch for weak employees are both reacting to the word “intelligence” — and they’re arriving at opposite conclusions from the same two syllables. Neither of them is wrong based on the information that they have.

That gap — between what something does and what we’ve decided it is — runs deeper than AI. It shows up every time a demonstration gets mistaken for a definition. And it happens in almost every room I study.

The conflation

There’s a move that shows up in conversations about technology, medicine, policy, leadership, really any domain where a concept carries weight and a demonstration carries force. Someone asks, “what is this thing?” and someone else answers by showing what it can do. A product demo replaces a definition. A capability display substitutes for an explanation. A personal success story stands in for a general claim.

The audience nods. The question feels answered. And nobody notices that the actual question — what is this thing, what does it mean, who gets to define its role — was never addressed.

This conflation — treating what something does as identical to what something is — operates quietly in almost every high-stakes conversation I study.

A surgeon demonstrates a new technique that cuts procedure time in half. Impressive. But “what can this technique do in one surgeon’s hands?” is a different question than “what is this technique’s role in standard care?” The demonstration doesn’t answer the second question. Adoption protocols, failure modes, training requirements, and patient selection criteria do.

A team lead shows a quarterly dashboard where every metric is green. The room concludes the team is healthy. But green metrics are what the team produced. Whether the team is healthy depends on questions the dashboard can’t answer: how sustainable is the pace, who is burning out, and what conversations are being avoided to keep the numbers clean.

An employee shows five ways they used AI to finish a project faster. The manager concludes the employee is dependent on a crutch. A different manager concludes the employee is a visionary. Same demonstration, opposite definitions — because neither manager is responding to the demo. Both are responding to what “AI” already means in their heads, and the demo just gave each of them permission to feel more certain.

Why this happens

Alfred Korzybski identified the root of this problem nearly a century ago. In Science and Sanity (1933), he warned against what he called the “is of identity” — the tendency to collapse a thing with its description, a map with its territory, a word with the object it points to. When someone says “AI is a game-changer,” Korzybski would flag that sentence. The word “is” performs an act of identification: it treats the label and the thing as the same. But “AI” is a label covering thousands of different tools, methods, capabilities, and contexts. Saying “AI is a game-changer” skips every question that matters: which AI, for whom, under what conditions, by what measure.

Korzybski’s broader point was that humans routinely confuse levels of abstraction. A demonstration lives at one level — concrete, specific, bounded by context. A definition lives at a higher level — abstract, general, meant to travel across contexts. When a demonstration gets treated as a definition, the concrete swallows the abstract. The specific case becomes the general rule. And the room moves forward on a foundation that feels solid but isn’t.

Psychologist Edward Thorndike identified the same pattern from a different angle. In 1920, he documented what he called the halo effect: the tendency for a strong impression in one area to color judgment in unrelated areas. Military officers who were tall and attractive were also rated as more intelligent and better leaders by evaluators who had never spoken to them. One visible trait radiated outward and shaped the assessment of everything else.

Thorndike’s finding has been replicated across domains for over a century. A 1977 study by Nisbett and Wilson showed that college students who watched a warm, friendly lecturer rated him higher on physical appearance and accent — traits that had nothing to do with his warmth. The initial impression didn’t just influence related judgments. It rewired unrelated ones.

The same pattern appears in technology. A study from the Nielsen Norman Group found that websites with high visual appeal received high satisfaction ratings from users even when the task-failure rate on those same sites exceeded 50%. Users liked how the site looked, and that impression bled into their assessment of how well it worked — even when it demonstrably didn’t work. “Beautiful” became a stand-in for “usable.” The demo replaced the definition.

Where it shows up in teams

In the conversations I study — teams under pressure, organizations making high-stakes decisions, groups trying to coordinate across different assumptions — the is/does conflation creates a specific and recurring failure pattern.

Someone shows impressive results. The room treats the results as proof of a larger claim. The larger claim goes unexamined because the demonstration felt like enough. Decisions get made. And when those decisions break down later, nobody can trace the failure back to the moment where a definition was needed, and a demo was offered instead.

Here’s how it sounds in practice:

“We rolled out the new system and productivity jumped 15% in the first quarter.” That’s what the system did. What the system is — its actual role, its fit with existing workflows, its long-term maintenance burden, its impact on the people using it — requires a different conversation.

“Our culture is strong. Look at our engagement scores.” That’s what the survey produced. What “strong culture” means on this team, in this building, under these specific conditions — that question is still open.

“AI is transforming everything. Look what I built in two weeks.” That’s what happened in one person’s hands, with one set of skills, on one project. What “AI” is — as a category, as a policy question, as a set of decisions your organization needs to make — remains undefined. And every person in the room is filling in that definition with their own assumptions, silently, while nodding at the same screenshots.

The repair

The fix is simple to describe and hard to practice, which is the definition of infrastructure.

When someone shows you an impressive result, train yourself to notice the moment your brain wants to leap from “that’s what it did” to “that’s what it is.” That leap feels natural. It feels like a conclusion. In reality, it’s a shortcut — and the gap it jumps over is where most coordination failures begin.

The question that closes the gap: “That’s a strong result. Now — what does this mean for us, in our context, with our constraints?”

That question moves the conversation from demonstration to definition. It honors the evidence without letting the evidence do work it can’t do. It creates space for the room to build a shared understanding instead of leaving with five separate private interpretations of the same demo.

A few other versions of the same move, depending on the context:

“That shows what it can do. What do we think it is — for this team, right now?”

“Impressive demo. What would need to be true for that result to hold across our full operation?”

“I can see the capability. What’s our definition of success for this, and does this demo match it?”

Each of these separates the is from the does. Each one costs about ten seconds. Each one prevents a room full of people from walking away with a shared experience and no shared meaning.

Why it matters now

We are living through a period where demonstrations are abundant and definitions are scarce. New tools produce visible, shareable, impressive outputs at unprecedented speed. Screenshots travel faster than analysis. A build log can go viral while the question “what does this tool mean for how we work?” remains unasked.

That asymmetry — demonstrations outpacing definitions — is the engine of most organizational confusion around technology. Teams adopt tools before they agree on what the tools are for. Leaders see a capability demo and assume alignment that doesn’t exist. Individuals have transformative personal experiences with a tool and can’t understand why others don’t share their certainty.

The certainty is real. The experience is real. The demonstration is real. What’s missing is the shared definition — the agreement about what this thing means in this room, for these people, under these conditions.

Until that definition exists, every person in the room is watching the same demo and seeing a different thing.

This essay is part of the What Do You Mean? series on What Time Binds, where I study what happens when people use the same words and mean different things — and what to do about it.

I’m building a 10-module course called Meaning Repair for High-Stakes Teams on this Substack. Module 1 is completely free. If the pattern in this essay felt familiar, that’s where to start: what-time-binds.com

I first watched the Richard Feynman magnet interview in 2015, late at night, the way people fall into YouTube rabbit holes when their brain refuses to shut off. I’ve always been drawn to physics, especially theoretical physics, because it demands a kind of honesty that most public discourse avoids. You don’t get to win with your feelings. You don’t get to hide behind a slogan. You have to build a coherent model.

Subscribe now

That’s why the magnet clip stuck with me.

The interviewer holds up two magnets and asks what most of us have wondered at some point: What is that “feeling” between them? The push. The pull. The invisible insistence.

And Feynman doesn’t answer the “why” first.

He answers the meaning first.

He pauses and says, in effect: What do you mean by “the feeling between the magnets”?
Not to be clever. Not to stall. To do the work that makes an answer possible.

That’s the part people miss. The clip is shared with the caption “Richard Feynman. Why.” I think the heavy lifting happens earlier. The moment he refuses to accept a fuzzy question and treats clarity as a form of respect.

Because if you let the question drift, the conversation drifts with it.

And that is the story of our current information environment: same words, different payloads, everyone convinced the other person is being ridiculous.

The moment he changes the room

Here’s the move.

The interviewer says, basically: There’s something there when you push magnets together. What is it?

Feynman replies: What do you mean by that?

He’s not arguing with the curiosity. He’s diagnosing the problem. “Feeling” can mean a tactile sensation in your hands. It can mean a mysterious substance in the air. It can mean a mechanism. It can mean a metaphysical “reason.”

If you don’t decide which one you mean, you don’t get an explanation. You get two people talking in different frames, passing sound through each other.

Clarity is the price of admission for “why.”

This is exactly what the What Do You Mean? series is about: the invisible infrastructure that lets people share a world long enough to think together.

And Feynman shows something else that matters right now: the emotional temperature of a conversation often rises because the question was vague, not because anyone was malicious.

One person thinks they’re asking for mechanism.
The other hears an existential demand.
Then we get heat, posture, performance.

Feynman defuses that by repairing the question before he touches the answer.

“Why” questions have floors

Once Feynman has the question stabilized, he gives a second lesson that belongs in every civic conversation: every “why” rests on assumptions you agreed to stop questioning.

He tells a story about “Aunt Minnie” in the hospital. People accept a certain level of “why” as satisfying because they share background knowledge about hips, hospitals, ice, gravity, and social roles. Change the listener (he imagines someone from another planet) and suddenly your explanation collapses. The “why” depends on a shared framework.

That’s a brutal truth about public discourse:

We often argue about “why” while hiding the fact that we don’t share a floor.

One person’s floor is “markets.”
Another person’s floor is “justice.”
Another person’s floor is “God.”
Another person’s floor is “history.”
Another person’s floor is “my lived experience.”

If you don’t surface the floor, the conversation becomes a spiraling ladder of “why” that ends in contempt.

Feynman doesn’t pretend he can solve that.

He says, plainly, that explanations bottom out. At some level, you hit fundamentals, things you take as elements of the world, at least until you have a deeper theory. In magnet talk, you run into electromagnetic force. You can go deeper if you’re trained. You still hit bedrock eventually.

It’s called epistemic hygiene.

It’s also a direct challenge to the culture of instant certainty.

The “cheating” principle

This is my favorite part of the interview, and it’s a warning label for the whole internet.

Feynman says he can’t explain magnet attraction by comparing it to rubber bands, because that would be cheating. It would sound familiar. It would feel satisfying. It would smuggle in a mechanism you haven’t earned. Then, the moment you push the analogy, you end up back at the same forces you were trying to explain.

That’s the whole modern persuasion economy.

“Like a rubber band.”

“Like a virus.”

“Like a parasite.”

“Like a war.”

“Like slavery.”

“Like Nazis.”

“Like communism.”

A metaphor that wins the room can still be cheating.

The point is not “never use metaphor.” The point is: don’t use a metaphor to avoid the hard work of mechanism. Don’t use familiarity to replace understanding.

Feynman is modeling a different ethic: if an explanation makes you feel smart while making you less accurate, it’s not an explanation. It’s a performance.

60-second tool: The Feynman Reset

Use this when a conversation starts getting heated quickly, at work, in civic meetings, in comment threads, and in family group texts.

Step 1: Pin the word.
What term is doing the most work right now?

Step 2: Name the meaning layer.
Are we stuck on definition, mechanism, scope, evidence, implication, or values?

Step 3: Ask one clarifying question and stop talking.
Let the other person repair their meaning.

Steal-this scripts:

• “When you say X, what do you mean by that in this context?”

• “What would X look like in practice?”

• “Walk me through the mechanism—how does X lead to Y?”

• “Which cases count, and which cases don’t?”

• “What would we need to see to agree we’re right or wrong?”

This is repair-by-design. You’re changing the conditions so shared reality can re-form.

Why it matters

Feynman’s magnet clip keeps circulating because it scratches an itch people can’t name: the feeling that we live in a world where everyone talks, and nobody explains.

We’re saturated with claims. We’re starved for mechanisms.

And that gap makes people easier to manipulate.

A high-compression meme doesn’t need to be accurate. It just needs to be legible in two seconds and identity-safe to share. It offers the comfort of a clean story. Feynman offers something harder: a disciplined curiosity that refuses to pretend.

He’s doing what every leader, educator, journalist, and organizer needs to do right now:

• Slow the frame down.

• Clarify the question.

• Separate certainty from understanding.

• Admit the limits of explanation without collapsing into cynicism.

This connects to a cognitive pattern that shows up in my research: people often believe they understand complex systems until they’re asked to explain the mechanism step-by-step. Then confidence drops. Sometimes extremism drops with it. That matters for civic life. It suggests a path out of ideological shouting: move from slogans to mechanisms.

No “gotchas,” just a shared experiment.

Quick Scan: copy/paste lines for real life

Use these in Slack, meetings, community forums, editor chats, and comment threads.

• “I want to make sure I’m tracking before I react.”

• “Define X the way you’re using it here.”

• “What’s the concrete example you have in mind?”

• “Which cases count as X?”

• “What’s the causal path—step by step?”

• “What would count as evidence on this?”

• “Are you implying X, or am I adding that?”

• “What outcome are you aiming for?”

• “What constraint are you optimizing for?”

• “What would change your mind?”

Carry it into your week

Pick one setting where you usually rush—work, civic discourse, or online.

Then run this field test three times before Sunday:

1. Identify the “magnet word” (the term doing the work).

2. Ask one mechanism or scope question.

3. Stop. Let the other person answer.

Track three signals:

• Did the temperature drop?

• Did the next turn clarify or escalate?

• Did you notice your own assumptions earlier than usual?

If you want, post your example as a Note. Real scenes sharpen the series.

Next: As promised in the last piece

We’ll tackle: when questions are used as weapons. How do we recognize a bad-faith “just asking questions” tactic and respond without taking the bait? How can we keep dialogue productive when questions are meant to provoke? That’s up next in the series.

A department meeting. The project lead breezes through updates, dropping an unfamiliar acronym and a vague reference to “the initiative.” Around the table, a few people exchange glances. You’re lost—did you miss a memo? You swallow hard. The director is in the room, and the last thing you want is to sound clueless. The moment to ask slips by, and the meeting moves on. You’re stuck, hoping you can bluff through the task. You glance around: nobody else speaks up either. Outwardly, everything seems fine, except that no one really knows what’s been agreed to.

Subscribe now

Moments like this carry heat. It’s the heat of status and social risk. It’s the flush of “I should know this” mixed with “If I ask, I’ll look incompetent.” It’s the tension of a public setting, colleagues and bosses in the room, and the clock pushing everyone forward. In this heat, asking “What do you mean?” feels like grabbing a live wire. So we nod along. We stay quiet. A simple point of confusion hardens into a silent fault line. The meeting adjourns with shaky alignment.

Sometimes silence feels safer than admitting we’re lost.

Why Asking for Clarity Feels Dangerous

Why is a clarifying question so fraught? It starts with face – our social self-image, the need to appear competent and respectable. A “What do you mean?” question can ding face in two directions at once. If I ask you to explain, it hints you weren’t clear (your face suffers). If I admit I’m confused, I risk looking ignorant (my face suffers). (In everyday terms, losing face means feeling embarrassed or humiliated.)

Context cranks up the risk. The impact of asking depends on power, relationships, and timing. We speak differently to a peer than to a CEO. Asking “What do you mean?” in private is easy; asking it of a person with more authority in front of others is another story. The bigger the power gap or audience, the more a clarification can feel like a challenge or a spotlight on someone’s mistake. Silence often wins out.

Our conversation habits reinforce this. Often, the original speaker will clarify themselves before anyone asks, an instinctive way to save face for everyone (remember Repair?). Directly asking someone to clarify breaks that norm. That’s why we tend to cushion it (“Sorry, just want to be sure I understand…”). We’re trained to avoid making others (or ourselves) look bad.

On top of that, many of us carry internal “voice” rules from past experience: Don’t question the boss in public. Don’t sound unsure, or you’ll look weak. These kick in automatically. We weigh the risk of speaking up (shame, angering someone, looking stupid) versus staying quiet. Too often, quiet wins. The normal fix for confusion (simply asking) gets vetoed by fear. The result? Meetings end with hidden misunderstandings, and discussions become parallel monologues, all because asking felt too risky.

Why We Don’t Ask: A Menu of Barriers

If we pull back the curtain, we find many reasons people hold back from asking for clarity. Here are some common barriers, each with what it feels like, what it sounds like inside, and what it costs:

• Status Risk (Looking Stupid) – Feels like: a pit in your stomach. Sounds like: “Everyone will think I’m clueless.” Cost: The clarification never gets aired, and you move forward with a shaky grasp of the plan.

• Power Risk (Challenging Up) – Feels like: tiptoeing through a minefield with a superior. Sounds like: “If I ask, will I be seen as second-guessing them?” Cost: Deference beats understanding. Bad ideas go unchecked; good ideas get lost in miscommunication.

• Identity Risk (Sensitive Territory) – Feels like: heart pounding at a loaded phrase. Sounds like: “If I ask what they mean, they’ll think I’m attacking their beliefs.” Cost: People talk past each other, each using the same words differently. Conflict flares over false assumptions.

• Audience Risk (Public Performance) – Feels like: red cheeks, aware that others are watching. Sounds like: “I can’t ask that in front of everyone—I’ll look dumb.” Cost: Discourse stays surface-level. Everyone nods along, and confusion goes underground until it explodes.

• Time Risk (Keep Moving) – Feels like: a spike of anxiety as the meeting races on. Sounds like: “No time to clarify—just decide and move on.” Cost: A false consensus. The team “decides” without true alignment, and time saved now is time lost fixing problems later.

• Norm Risk (Culture of Silence) – Feels like: walking on eggshells in a “no questions” culture. Sounds like: “Here, asking for clarification equals dissent.” Cost: An echo chamber. Questions are frowned upon, so misunderstandings accumulate and innovation stalls.

(Notice: personal history can amplify all these fears—a nervous temperament or past punishments make any question feel riskier. And in some settings, a general cynicism (“questions are just traps”) makes even sincere clarification feel suspect.)

These barriers often pile on. Imagine asking a powerful figure to clarify a politically loaded point in a high-stakes forum. Most of us would rather bite our tongue. Misunderstandings persist, bad decisions go uncorrected, and trust erodes.

Unasked questions turn into hidden problems.

Counter-Moves: How to Lower the Threat and Ask Anyway

Even in a high-heat moment, you can ask for clarity in a safer way. Here are a few counter-moves that reduce the threat. Each comes with a “steal-this” script:

• Stance First (Declare Intent): Preface your question with why you’re asking, to show goodwill. For example: “I want to make sure we’re on the same page. Can I ask a quick clarification about X?”

• Offer a Guess (Candidate + Escape Hatch): Give your best guess of their meaning, and invite correction. For example: “When you mention X, do you mean Y? (If I’m off, let me know.)” This way, if you’re wrong, you’ve made it easy for them to adjust without feeling challenged.

• Pinpoint the Gap (Targeted Trouble): Ask about a specific part of what they said, not the whole thing. For example: “Sorry, which initiative are we referring to, exactly?” Focusing on the unclear term keeps it neutral and contained.

• Blame the Lingo, Not the Person (Depersonalize): Frame it as clarifying a term, not questioning them. For example: “The term ‘X’ can mean different things; can we clarify how we’re using it here?” Now you’re teaming up to clarify language, rather than putting them on the spot.

• Buy Time for Clarity (Pause the Rush): If things are moving fast, propose a tiny pause to define terms. For example: “Can we take 30 seconds to clarify Y before we decide? It might save us time later.” This signals that your aim is shared understanding, not delay.

60-Second Tool: Should I Ask?

If you’re hesitating, run this mental checklist:

1. Name the barrier. Are you afraid of looking dumb (status)? Of who’s listening (audience)? Of angering someone (power)? Pinpoint the fear.

2. Choose a safe approach. Match a strategy to the fear. If it’s about tone or respect, lead with your intent (“I want to understand…”). If it’s about power dynamics, make the question extra polite or offer a guess (“Are you saying X?”). If it’s a sensitive topic, focus on the term, not the person (“Can we define Y?”).

3. Ask and stop. Pose one clear question and then listen. Don’t ramble or stack multiple questions. A concise, neutral question is far less threatening, and it invites a clear answer.

That’s it: identify what’s stopping you, adjust your approach, and ask. A few seconds of courage can prevent hours of confusion.

Quick Scan: Ready-to-Use Clarification Lines

• “I want to understand before I react…”

• “When you say X, do you mean Y?”

• “Which cases count?”

• “Walk me through the mechanism.”

• “What would count as evidence here?”

• “Are you implying X, or am I reading that in?”

• “What outcome are we aiming for?”

Each of these phrases gets clarity without casting blame. They’re short, sincere, and focused on the idea, not the person. Feel free to adapt them to your voice. The goal is to have a few go-to moves so that when confusion hits, you’re prepared to respond instead of freezing up.

Carry It Into Your Week: Clarification Challenge

Try a simple experiment this week. Use one of the quick-scan scripts above three times: maybe once in a meeting, once in a personal conversation, and once online. Each time, observe what happens. Does the temperature of the conversation drop a bit? Do you gain clarity or uncover a key detail? Does the dialogue shift in a positive way?

At week’s end, see what changed (if anything). You might be surprised how often a single clarifying question can redirect a discussion. If you try it, consider sharing your observations in the comments or Substack Notes; we can all learn from the stories.

Next: When Questions Get Weaponized

We’ve focused on asking in good faith to improve understanding. Next, we’ll tackle the flip side: when questions are used as weapons. How do we recognize a bad-faith “just asking questions” tactic and respond without taking the bait? How can we keep dialogue productive when questions are meant to provoke? That’s up next in the series.

When conversations heat up, reality has a tendency to split. A simple misunderstanding can snowball into two people talking past each other, each convinced they’re “right” and the other is being obtuse or hostile. If you’ve ever watched a minor disagreement turn into a full-blown argument in minutes, you know how easily shared reality can fracture under pressure. Words that were meant one way get heard another way. Clarifying questions start to feel like personal attacks. Before you know it, it’s no longer a dialogue – it’s two dueling monologues in separate realities.

It doesn’t have to go that way. Shared reality doesn’t maintain itself. We maintain it — one repair at a time. When the room gets hot, a few simple protocols can keep everyone on the same page. Think of these as specialized moves for asking “What do you mean?” without fanning the flames. They’re designed to lower defensiveness, untangle confusion, and get you and the other person seeing the same picture again. Here are three protocols to help re-sync understanding when it matters most:

Protocol 1: State Your Intent Before Your Question

Summary: Lead with a reassuring preface to signal curiosity, not accusation. Under stress, people often perceive a clarifying question as a challenge or attack. By stating your intent upfront, you disarm that reflex. Essentially, you’re saying, “I’m asking because I want to understand, not because I’m trying to score points.”

Why it helps: This extra sentence lowers the threat level. It assures the person that you’re not setting a trap or judging them – you simply want clarity. In psychological terms, you’re giving them a safe space to explain themselves rather than putting them on the defensive. The question that follows is far more likely to land as curiosity rather than accusation.

Example:
A: “We can’t keep extending the deadline. This team needs to be more disciplined.”
B: (feeling a flare of anxiety and potential offense) – “I want to make sure I get your point before I react. When you say ‘be more disciplined,’ what do you mean in this context?”

In this scenario, Person B first assures A that their goal is understanding, not pushback. By prefacing the question with “I want to make sure I get your point before I react,” B shows respect for A’s perspective. The follow-up question (“What do you mean by ‘be more disciplined’ here?”) then comes across as a genuine ask for clarification. A, sensing B’s good-faith stance, is more likely to explain rather than get defensive. This protocol interrupts the typical fight-or-flight pattern and buys both parties time to recalibrate rather than escalate.

Protocol 2: Offer a Hypothesis (with an Escape Hatch)

Summary: Instead of bluntly asking for meaning, paraphrase what you think they mean and invite correction. This is a two-part move: (1) Candidate interpretation – you gently hypothesize their meaning, and (2) Escape hatch – you explicitly give them room to correct you if you’re wrong. The format is something like: “When you say X, do you mean Y? If not, I’m curious what you do mean.”

Why it helps: First, offering a guessed meaning shows you’re actively trying to see it from their side. It’s a sign of good faith – you’re meeting them halfway by attempting to articulate their point. Second, the “if I’m wrong, let me know” part (the escape hatch) ensures they don’t feel cornered. You’re making it clear that you’re not wedging them into a yes-or-no trap; you’re open to being corrected. This turns a potential confrontation into a collaboration: now you’re two people refining a meaning together, rather than adversaries.

Example:
A: “I’m just so fed up with our department’s support on this project.”
B: “Let me check if I’m tracking: when you say ‘support,’ do you mean you feel the other team isn’t responding enough? If that’s not it, what do you mean exactly by ‘fed up with support’?”
A: “Yeah, exactly – I feel like whenever we ask for help, it takes days to get a reply. I don’t mean they’re bad at their jobs; I mean we’re low priority for them and it’s frustrating.”

Here, B tested a reasonable interpretation of A’s complaint. By phrasing it as “Do you mean X? If not, correct me,” B showed humility and a sincere effort to understand. A, now assured that B is not twisting their words, clarified the real issue (slow response times, not personal incompetence). Notice how this protocol flips the script from potential argument to joint problem-solving. A feels heard, and B gains a more precise grasp of A’s concern. The escape hatch (“if that’s not it...”) kept things low-pressure, giving A freedom to say “No, that’s not what I meant” without feeling judged.

Protocol 3: Zoom In on the Specific Gap

Summary: Aim your question at the specific piece that’s unclear, instead of saying a blanket “What do you mean?” Identify what type of misunderstanding might be happening and ask about that. Are you unclear on who they mean, which case they mean, or how something works? Pinpoint it. In practice, this means using a targeted prompt: “Who specifically?” “Which part are we talking about?” “How would that work?” – whichever gets to the heart of your confusion.

Why it helps: A generic “What do you mean?” can be taken as “You’re not making sense.” It’s open-ended and might put someone on the spot to justify themselves broadly. In contrast, a pointed question shows that you’re zeroing in on the idea, not the person. It also narrows the scope, making it easier for them to clarify one thing at a time. We often have trouble in conversation because one term or reference was fuzzy; by isolating that, you prevent the entire exchange from derailing. In essence, you’re debugging the conversation: finding exactly where the understanding broke and fixing just that piece.

Example:
A: “They really dropped the ball on this one. It’s honestly unethical.”
B: “I’m with you that something’s off. Quick clarification: when you say ‘they,’ who do you have in mind? And what part strikes you as unethical?”
A: “By ‘they’ I mean upper management, not the frontline team. The unethical part is that the decision was made without consulting those it impacts – it feels deceitful.”
B: “Got it – so the concern is about top leadership making secret calls that hurt the rest of us.”

In this exchange, B didn’t ask a vague question; instead, B pinpointed two potential gaps: the referent (who is ‘they’?) and the value judgment (which aspect is ‘unethical’?). This targeted approach helped A refine their statement: A specified the group (“upper management”) and the issue (lack of consultation). B could then reflect back the clarified meaning. By zooming in on ambiguous words and asking directly about them, you restore mutual understanding on precise points. It also shows respect – you’re engaging with what they actually said, not what you might have assumed they meant.

Pocket Summary

• Lead with intent: Before a tough question, reassure the other person why you’re asking. A line like “I want to understand before I respond” signals goodwill and lowers defensiveness.

• Paraphrase and ask: Try a candidate interpretation of their point, and invite correction. “Do you mean __? If not, let me know what you do mean.” This shows you’re trying to get on their wavelength, not catch them out.

• Get specific: Pinpoint the unclear element – whether it’s who a pronoun refers to, what a term encompasses, or how a process is supposed to work. Ask that specific question. It’s easier to answer and less likely to be misread as an attack.

• Goal: Keep the conversation in a shared reality by actively repairing meaning gaps. The point isn’t to win or be right in the moment, but to make sure both of you are actually talking about the same thing. Only then can real resolution or progress happen.

Field Test: Carry It into Your Week

These protocols are skills – and like any skill, they get stronger with practice. This week, look for an opportunity to test them out in real life. It doesn’t have to be a five-alarm conflict; even a minor misunderstanding or a confusing remark from a friend is a chance to try a shared reality move. For example:

• At work: The next time a discussion in a meeting starts feeling tense or convoluted, jump in with a Protocol 1 move. e.g. “Hey, before we go further, I just want to make sure I understand your perspective. You’re saying ___, is that right?” Notice how setting your intent changes the tone of the response you get.

• At home: If a family or roommate conversation hits a bump, practice Protocol 2. Paraphrase what you think they mean, and add “Correct me if I’m wrong.” You might say, “So, you’re upset because it sounded like I dismissed your idea – is that it? If not, help me understand.” Pay attention to their reaction; do they seem relieved to clarify rather than argue?

• Online or text: The next time you read a message that rubs you the wrong way, instead of firing back or stewing, try a Protocol 3 approach in your reply. Ask about a specific word or ask for an example: “When you said ____, did you have a particular situation in mind?” This can prevent an all-caps comment war by steering the exchange toward specifics and away from assumptions.

Treat these experiments as low-stakes training. You’re building your “clarification muscle memory” so that when high-stakes pressure does hit, you’ll be more comfortable using these tools naturally. The field test is simple: step into a moment of potential misunderstanding and consciously apply one protocol. Afterwards, reflect: Did the conversation feel different? Did tension ease? Each time you do this, you’re reinforcing a powerful habit. You’re proving to yourself (and others) that even under pressure, you can choose accuracy over adrenaline, alignment over ego.

Carrying These Protocols Into Your Week

You’re putting out fires all day, and you’re preventing them. You’re helping create a culture (at work, at home, in your friend group) where asking “What do you mean?” is welcome and normal, not a cause for offense. Little by little, you’re keeping shared reality intact when it matters most, one clarification at a time. And that skill compounds – after all, big conflicts are often just small misunderstandings that went un-repaired. This week, see what shifts when you take the initiative to repair in the moment. You might be surprised how much smoother things go when everyone’s operating in the same reality.

I’ve written about this pickup story before. I used it in a piece about “ghost maps,” the private mental maps we keep using long after the territory has changed. I’m bringing it back here because it’s the cleanest example I know of repair in the wild: the moment a relationship could have slid into blame, and a single clarification rebuilt the world we were sharing.

This series is about the question “What do you mean?” as cognitive infrastructure.

Subscribe now

Part 1 laid down the architecture: pragmatics, speech acts, implicature, the machinery of meaning when things are calm. Repair is the same machinery under pressure. It’s what keeps a conversation from splitting into two different realities.

Repair is the moment you choose accuracy over adrenaline.

The elevator that wasn’t there

I was picking up my youngest daughter after an event.

I parked where I always parked—near the elevator. The usual spot. The spot my body could find without thinking.

She wasn’t there.

Minutes passed. I called. No answer. I called again. When she finally picked up, my tone came out stern, frustrated, and certain she was being difficult. I was already halfway into a story: She’s ignoring me. She’s giving me attitude. She’s making this harder than it needs to be.

Then she said, calmly and genuinely confused:

“There is no elevator here.”

That sentence did three things in an instant:

1. It contradicted my certainty.

2. It exposed my hidden assumption: I had treated my map as the map.

3. It forced a reset: I stopped judging her behavior and started checking reality.

Nobody had told me the pickup spot had changed. I was in the wrong place. She wasn’t disrespectful. She was exactly where she was supposed to be.

I didn’t need a better argument. I needed a better alignment.

Misunderstanding doesn’t always sound like confusion. Sometimes it sounds like confidence.

That pivot, moving from posture to recalibration, is what this essay is about.

What repair is

Repair is a system built into ordinary conversation. It’s the set of moves humans use to fix problems in speaking, hearing, and understanding while the conversation is still happening.

Repair is not rare. It’s constant. You can hear it in the smallest moments:

• “Wait—who?”

• “When you say they, who are you talking about?”

• “Do you mean midnight as in 12:00 AM, or end-of-day midnight?”

• “Hold on. I’m not tracking. What do you mean?”

Repair is how we keep a shared world stable enough to think together.

Conversation isn’t a pipeline. It’s a coordination problem. Repair is the coordination tool.

How conversation analysts figured this out

The most important thing conversation analysts did was stop relying on after-the-fact explanations.

They recorded real talk: phone calls, dinner table discussions, workplace moments—and they watched what people did, turn by turn. Their basic logic is simple:

• If you want to know what someone understood, don’t guess.

• Look at what they do next.

The next turn shows how the previous turn landed. If the next turn treats the prior turn as an insult, a promise, a joke, a threat—that treatment becomes visible in the interaction itself.

That’s why repair matters so much. It’s what happens when the “next turn” reveals that two people are no longer standing on the same ground.

The preference for self-correction

Here’s a key finding that changes the way you hear conflict:

Conversation is built to let people fix their own meaning.

If a problem shows up, the structure of talk typically gives the original speaker first rights to correct, clarify, adjust, or restate. That matters because direct correction often pulls rank. It can trigger defensiveness. It can feel like a status move. It can feel like a verdict.

Repair does something smarter. It slows the interaction down and gives the other person room to steer their own meaning back into clarity.

Correction grabs the steering wheel. Repair taps the brake and hands the wheel back.

That’s one reason “What do you mean?” can be so powerful. When it works, it invites self-correction without humiliation.

When it fails, it sounds like prosecution.

So the question isn’t whether we repair. The question is whether we repair in a way that the other person can hear.

The repair ladder

Not all repair moves are equal. People typically climb a ladder from vague to specific, depending on what kind of trouble they’re dealing with.

Here’s the ladder in plain language, with examples you’ll recognize.

1) Open trouble: “Huh?”/“What?”

This signals a problem, but it doesn’t specify what kind.

A: “We should move it to the other one.”
B: “Huh?”
A: “The other meeting—the Thursday slot.”

2) Targeted trouble: “Who?”/“Where?”/“When?”

This points to a missing reference.

A: “They already decided.”
B: “Who’s ‘they’?”
A: “The committee.”

3) Repeat or partial repeat

This often puts pressure on the trouble spot without turning it into a fight.

A: “It’s obvious.”
B: “Obvious?”
A: “To me it is—because I’ve been tracking it for months.”

4) Candidate understanding: “You mean X?”

This is one of the most useful moves you can learn. It’s fast when you’re close. It’s respectful when you’re wrong and give an escape hatch.

A: “This policy is about safety.”
B: “When you say safety, do you mean public safety… or political safety?”
A: “Public safety. I’m talking about harm reduction.”

5) Meaning trouble: “What do you mean?”

This signals: I heard the words. I can’t map them into a stable interpretation.

A: “People just need to take responsibility.”
B: “What do you mean by responsibility here?”
A: “I mean ownership of the consequences, not punishment.”

“What do you mean?” is the repair move that targets the meaning layer.

It’s the move you reach for when the conversation has started producing heat because the interpretation has drifted.

Why repair breaks under pressure

Repair is cognitive, and it’s social.

A repair question can land as:

• curiosity, or

• accusation.

The difference often has nothing to do with the words and everything to do with the implied stance: Are you inviting me to clarify, or are you setting me up to lose face?

Under stress, people treat questions as evaluations. In high-conflict environments, questions get read as traps. Online, questions get weaponized so often that a sincere request for meaning can feel like a disguised attack.

Add one more ingredient: identity.

When someone’s claim is fused to their moral standing, a repair request can feel like a challenge to who they are, not what they mean. That’s how a conversation about a word turns into a fight about worth.

When identity takes the wheel, repair feels like disrespect.

And this is where my elevator story matters. I wasn’t “confused” in that parking garage. I was certain. My certainty was doing identity work: I’m the responsible one. I’m the parent. I’m the one being disrespected.

Repair took one sentence to interrupt that storyline: “There is no elevator here.”

It didn’t shame me. It reoriented me.

Repair as civic infrastructure

If you zoom out, repair becomes a public problem.

A lot of public discourse now runs on compressed language—short posts, clipped clips, headlines without context, slogans standing in for arguments. Words like “freedom,” “DEI,” “patriotism,” “violence,” “law and order,” “racism,” “woke,” and “accountability” carry dense moral payloads. People use the same vocabulary while operating different internal models.

When repair disappears, we get predictable outcomes:

• We argue against meanings that were never intended.

• We assign intent where there’s ambiguity.

• We moralize what is often a coordination failure.

• We escalate because we feel unheard, not because we’re “right.”

Repair keeps disagreement from turning into theater.

The collapse of repair is one way a society loses the ability to think together.

That’s why this series keeps returning to the same hinge: the ability to ask for meaning without turning it into a fight.

Three repair moves you can use immediately

You don’t need to become a conversation analyst to practice repair. You need a small set of phrases you can trust when the room gets hot.

1) The “stance first” repair

Start by declaring your intent before you ask.

“I want to understand your model before I react. What do you mean by X here?”

This lowers threat and buys you clarity.

2) Candidate understanding with an escape hatch

This is the most practical move for real life.

“Let me check I’m tracking. When you say X, do you mean Y?”
“If that’s not it, what do you mean?”

The escape hatch matters. It signals: I’m not cornering you. I’m testing a hypothesis.

3) The “pinpoint the trouble type” move

Use the taxonomy you’ve already seen in this series. Choose one target and ask that question.

• Reference: “Who are we talking about, specifically?”

• Scope: “Which cases count, and which cases don’t?”

• Mechanism: “How does X lead to Y—step by step?”

• Implication: “Are you implying X, or am I reading that into it?”

Strong repair starts with diagnosis.

Why I’m using the elevator story again

Because it keeps me honest.

I can write about repair like it’s a public virtue. The elevator story shows repair as a private discipline. It shows the cost of skipping it. It shows how quickly a human mind fills in missing information with narrative, then treats that narrative as evidence.

It also shows the humility built into repair: the willingness to update your map while you’re still holding the wheel.

Repair asks for one thing: a pause long enough for reality to re-enter the conversation.

Repair is a pause with a purpose.

Carry it into your week

Pick one repair move and use it three times before Sunday.

Here’s the simplest one:

“Let me check I’m tracking. When you say X, do you mean Y?”

Use it once at home. Once at work. Once online.

Then notice what changes:

• Does the temperature drop?

• Does the other person clarify?

• Do you feel less pressure to “win”?

• Do you hear your own assumptions earlier?

If you want to share, drop your example in a Note. The series gets sharper when we build this from real interactions, not abstract ideals.

Next

Next comes the thing most readers actually want: scripts.

You already have the taxonomy, the diagnostic. The next tool post will be protocols: exact phrasing for work meetings, family friction, comment threads, and teaching. The point stays the same: a repair move that preserves dignity while restoring meaning.

Repair is the bridge. It takes meaning out of theory and puts it where we live: in the middle of the sentence, when your pulse has already started rising.

Shared reality doesn’t maintain itself. We maintain it—one repair at a time.

Skip the two foundational links below if you have already read them.

Subscribe now

The premise

A lot of American discourse right now is “argument” built on a missing step: we never agree on what a key word is doing. “Freedom.” “Safety.” “Woke.” “Accountability.” “Midnight.” We treat the word like a shared object, then we punish each other for not holding the same one.

This post is the first carry-it-into-your-week anchor.

It’s a taxonomy of the most common “meaning failures,” so you can ask for the right kind of clarity: definition, scope, mechanism, evidence, implication, values, and more.

In conversation analysis terms, this is repair (more to come on the repair topic later): how humans flag trouble and fix understanding mid-stream. Repair exists because problems of speaking/hearing/understanding are normal.

In cognitive science terms, it’s also grounding: updating common ground so the conversation can keep moving without having to guess.

A 60-second diagnostic

Before you ask anything, pick the failure point. Most confusion falls into one of these bins:

1. What are we talking about, exactly? → Reference

2. Which cases count? → Scope/boundaries

3. How does it work? → Mechanism

4. Is it true? → Evidence/standard

5. What’s being suggested without being said? → Implicature/presupposition

6. Why does it matter? → Values/stakes

7. What do you want done? → Action/next step

If you pick the bin, your question gets sharper and the temperature drops.

Tone key (so it lands as curiosity)

A clarification question can sound like a cross-examination if the other person is fused to the claim. So lead with one sentence that protects face and signals collaboration:

• “I want to understand your model before I react.”

• “Help me map what you mean, so I don’t straw-man you.”

• “I may be missing context, can you tighten this for me?”

Then ask your targeted question.

Politeness research treats these as ways of reducing the perceived threat of a request.

The taxonomy: 12 targets for “What do you mean?”

Use this like a menu. You don’t need all twelve. You need the right one.

1) Definition

Use when: a keyword is doing heavy lifting.
Ask: “When you say X, what do you mean by it in this context?”

Example: “When you say ‘accountability,’ do you mean consequences (punishment), transparency, or ownership?”

2) Reference (who/what exactly?)

Use when: the claim feels vague because it points at a fog.
Ask: “What specifically are you referring to: who, what event, what policy, what example?”

Example: “When you say ‘they,’ who do you mean: leadership, voters, media, agencies?”

3) Scope (boundaries + edge cases)

Use when: two people agree on the center and fight on the edges.
Ask: “Which cases count, and which cases don’t?”

Example: “When you say ‘free speech,’ are you talking about legal protection, platform rules, or social norms?”

4) Time (when + how long?)

Use when: disagreement is really about timeline.
Ask: “What time window are we talking about: today, this year, decades? What’s ‘before’ and ‘after’ here?”

Example: “When you say ‘recently,’ do you mean last week, last month, post-2020?”

5) Mechanism (how does X lead to Y?)

Use when: people are trading conclusions without a causal story.
Ask: “Walk me through the mechanism, how does X produce Y, step by step?”

Example: “How exactly does that policy change lead to that outcome? What’s the chain?”

(Mechanism questions are powerful because they force the claim to show its work.)

6) Evidence (what would count?)

Use when: the argument is stuck in “I feel” vs “prove it.”
Ask: “What evidence are you using, and what evidence would change your mind?”

Example: “What would you accept as a credible source here: data, firsthand reporting, audits, court records?”

7) Source/provenance (where did this come from?)

Use when: the claim arrives preloaded with certainty.
Ask: “Where did you first encounter this, and what makes that source reliable to you?”

Example: “Is this from a primary document, an article, a clip, a secondhand summary?”

8) Implicature (what’s being suggested?)

Use when: the conclusion is implied, not owned.
Ask: “Are you implying X, or am I reading that into it?”

This is classic Grice: people infer meaning to keep conversation coherent, and that inference is where conflict often ignites.

9) Presuppositions (what must be true for this to make sense?)

Use when: the framing smuggles in assumptions.
Ask: “What are we assuming as a given here? What would have to be true for your claim to hold?”

Example: “That question assumes intent; are we sure intent is established?”

10) Values/criteria (what are we optimizing for?)

Use when: facts are being used to serve a value conflict.
Ask: “What value is driving your position: fairness, safety, liberty, loyalty, harm reduction?”

Example: “Is your priority preventing harm, preserving autonomy, or enforcing consistency?”

11) Stakes/intent (what are you trying to accomplish?)

Use when: the conversation feels like it has hidden goals.
Ask: “What outcome are you aiming for by saying this: persuasion, warning, venting, coordination?”

Example: “Do you want me to agree, to act, or to understand where you’re coming from?”

12) Action/next step (what do we do now?)

Use when: meaning is clear enough, and the real gap is coordination.
Ask: “What would a good next step look like?”

Example: “What decision are we trying to make by the end of this conversation?”

Quick Scan (copy/paste)

If you only save one thing from this post, save this:

• “Define X in this context.”

• “What exactly are you referring to?”

• “Which cases count?”

• “What time window?”

• “What’s the mechanism?”

• “What evidence would change your mind?”

• “Where did this come from?”

• “Are you implying X?”

• “What assumptions are baked in?”

• “What value is driving this?”

• “What are you trying to accomplish?”

• “What’s the next step?”

Carry it into your week (starting today)

Pick one target from the taxonomy and use it three times before Sunday (if possible).

Here’s the rule: no rebuttal until you’ve run one targeted clarification question.

If you want a simple start, use Scope or Mechanism. They’re high-leverage and low-drama.

Careful promise: when the series is complete, I’ll compile the taxonomy + scripts into a single printable field guide for readers who want the whole kit in one place.

Next Up

Next: Repair—what conversation analysts learned from the way humans fix misunderstandings mid-stream. Repair is the bridge between “meaning in theory” and “meaning under pressure.”

Why “What do you mean?” is cognitive infrastructure hiding in plain sight

This isn’t the first time I’ve had this conversation, but in August 2007, inside a dimly lit tent during a military training exercise, a group of Marines and Sailors drifted from logistics into something heavier: obedience, moral certainty, and the old line, I’m just following orders.

Subscribe now

My Lai came up. The Holocaust came up. One Marine said, with total conviction, that he would never harm civilians. Another admitted he wasn’t so sure, boot camp and service conditions you in ways you don’t fully see.

It was turning into a values contest. Identity. Honor. Who’s the “good” Marine.

Then someone asked a simple question: “What do you mean by ‘conditioning’?”

And the room changed.

The debate stopped being a moral performance and became an inquiry into a mechanism, how training shapes behavior, how obedience gets built, how a human being can be guided into actions they would swear, in calmer conditions, they’d never take. One clarification question changed the level of analysis. We moved the conversation from posturing to reasoning.

That moment is this series in miniature.

Because here’s the thing most people miss: meaning isn’t an object that travels from one mind to another. Meaning is negotiated in real time. Without a repair move, people can use the same moral vocabulary while running completely different internal models.

So when I say this series is about “What do you mean?” I’m not talking about manners; rather, I’m talking about a tool that keeps humans from talking past each other, at work, at home, online, and in public life.

Part 1 is the foundation. To see why this question works, we need a firm grasp of three ideas: pragmatics, speech acts, and implicature.

Quick orientation on what you’ll walk away with from this series:

• A taxonomy of clarification targets (definition, scope, evidence, mechanism, implication, values, presuppositions, and more).

• A set of micro-protocols (short scripts) for meetings, classrooms, online chats, and high-stakes conversations.

• A repeatable habit for media literacy: slowing down meaning before the narrative takes over.

Each post ends with a field test, because this work is meant to be used.

1) Pragmatics: meaning in use

Most people were taught that language is basically a labeling system: words point at things, definitions settle disputes, and the dictionary is the referee.

Pragmatics comes in and says: watch what happens when people actually talk. Watch what the words do in context: who’s speaking, to whom, with what history, under what pressure, toward what goal. Pragmatics is the study of meaning as it’s used and inferred in real situations, not just encoded in sentences.

A phrase like “That’s interesting” can be sincere curiosity, a warning shot, or a polite exit ramp. A question like “Are you done?” can be a request for information, or a threat in business casual.

Pragmatics is why “What do you mean?” is a family of questions aimed at different failure points: definition, scope, reference, implication, evidence, mechanism, and more.

Later in this series, I’ll name those targets and give you a menu you can use in minutes.

2) Speech acts: words are actions

Speech act theory gives you a clean lens: every utterance has at least three layers.

• Locution: the words spoken

• Illocution: the act performed by saying them (requesting, warning, accusing, inviting, promising)

• Perlocution: the effect it has (calming, provoking, motivating, humiliating)

We’re attuned in everyday conversation not primarily to sentences as strings of words, but to the acts those sentences perform.

This is why so many arguments feel “irrational.” People are often fighting over illocution, not content.

Think about how many conflicts start here:

• “I’m just saying…” (The listener hears an accusation.)

• “Relax, it was a joke.” (The listener hears contempt.)

• “I’m asking a question.” (The listener hears cross-examination.)

Your nervous system is reacting to what you think the other person is doing to you with their words.

A sharp version of “What do you mean?” targets that layer directly:

“Are you asking me to fix this, or are you venting?”

The goal is conflict prevention and not pedantry.

In status-heavy settings, clarification can sound like a challenge: Are you unclear? Are you wrong? Are you being evasive? The craft is asking in a way that protects dignity while still demanding precision.

3) Implicature: the unsaid that runs the show

Now we get to the part that drives modern discourse insane: implicature, meaning that is implied rather than stated.

Grice’s core insight was simple: in conversation, we assume cooperation. We assume people are trying to be understood. So when a statement seems incomplete, we infer what “must” be meant. That inference is often where fights begin.

Example:

• Person A: “The budget is tight.”

• Person B hears: “So you want to cut the project.”

• Person A hears the pushback as: “You’re refusing reality.”

Implicature is where someone can signal without owning the claim. It’s also where media manipulation thrives: scope creep (“everyone knows…”), presuppositions hidden inside questions (“why are they lying?”), insinuations that travel faster than evidence.

A media-literate “What do you mean?” interrogates the implied payload:

“When you say everyone, who exactly?”

“What would have to be true for that claim to hold?”

“Are you implying X, or am I reading that into it?”

This is the difference between getting captured by a narrative and forcing a claim to show its work.

4) Why this matters in cognitive science

If pragmatics explains why misunderstanding happens, cognitive science explains why we’re so confident while misunderstanding each other.

Two biases matter here.

The illusion of transparency

We routinely overestimate how visible our intent is to others, how much our internal context “comes through.” Gilovich, Savitsky, and Medvec call this the illusion of transparency: we assume people can read our internal states far better than they can.

That’s why speakers feel misunderstood, and listeners feel blamed for “not getting it.”

The illusion of explanatory depth

We also overestimate how well we understand mechanisms. Rozenblit and Keil showed that people often feel they understand causally complex systems in much greater depth than they actually do—until they’re asked to explain how the thing works step by step. Confidence collapses on contact with mechanism.

This is where “What do you mean?” turns into an intellectual tool.

When you ask for the mechanism (How, specifically, does X lead to Y?), you’re clarifying and stress-testing certainty.

Fernbach and colleagues found that when people are asked to explain how political policies work, perceived understanding drops, and extreme attitudes often soften. Mechanism questions induce a kind of intellectual humility.

That matters for media literacy because modern platforms reward the opposite: hot takes, compressed claims, identity signaling, speed, and “I won.”

“What do you mean?” is a speed bump for the nervous system and a stress test for the claim.

5) Clarification is interaction design

A good clarification question is accurate, and it’s hearable. It reduces threat, signals goodwill, and then asks for precision.

Here are three micro-protocols I use (and teach) because they work across domains:

• Workplace precision (A/B menu): “I want to execute this cleanly. When you say X, do you mean A or B?”

• Heated debate (mechanism check): “Walk me through the mechanism. How, specifically, does X lead to Y?”

• Crucial conversation (fact vs. story): “I noticed [observable fact]. I’m telling myself a story that it means [interpretation]. Is that what you meant?”

Notice the pattern: you’re not starting with a verdict. You’re starting with a request for meaning that keeps both people in the conversation.

This is where this series becomes a toolkit.

Over the next few posts, you’ll get two things you can keep on your desk:

1. A taxonomy of clarification targets (definition, scope, mechanism, implication, evidence, values, presuppositions, and more).

2. A set of protocols, short scripts for meetings, classrooms, online heat, and high-stakes conversations.

Careful promise: when the series is complete, I’ll compile the taxonomy + protocols into a single, printable field guide for readers who want the whole kit in one place.

Each part will end with a field test. This is meant to be usable and, with personal effort, understood.

6) Preview: repair—the human system for fixing understanding mid-stream

So far we’ve been in architecture: how meaning gets built (and how it breaks) through action, implication, and context.

Next, we go one level deeper into repair.

Conversation analysts noticed something most of us ignore: everyday talk has a built-in “quality control” system. Humans constantly test, correct, and renegotiate understanding turn by turn. When something goes off, we don’t abandon conversation; we need to repair it.

Schegloff, Jefferson, and Sacks famously described repair as addressing problems of “speaking, hearing, and understanding,” and they documented a strong preference for self-correction—people typically try to fix their own talk once a problem is flagged.

“What do you mean?” is one of the cleanest forms of other-initiated repair: a signal that the problem is meaning and not how loud it was said.

Next week: repair—what conversation analysts learned from how humans fix understanding mid-stream. We’ll look at why repair is structured the way it is, why conversation prefers self-correction, and how those patterns give us a practical template for clarification that works in real life.

That’ll be my bridge for you to “carry it into your week” posts that will anchor the series: the taxonomy and the protocols.

Field test (5 minutes)

This week, replace one rebuttal with one clarification. Ask exactly what you’re uncertain about (definition, scope, evidence, mechanism, or implication) and watch what happens to the temperature of the conversation.

If you try it, tell me what changed.

References (starter list)

• Pragmatics (Stanford Encyclopedia of Philosophy).

• Speech Acts (Stanford Encyclopedia of Philosophy).

• Implicature (Stanford Encyclopedia of Philosophy).

• Schegloff, Jefferson & Sacks (1977), “The preference for self-correction in the organization of repair in conversation” (PDF).

• Gilovich, Savitsky & Medvec (1998), “The illusion of transparency…” (PubMed).

• Rozenblit & Keil (2002), “The misunderstood limits of folk science…” (PMC).

• Fernbach et al. (2013), “Political extremism is supported by an illusion of understanding” (PubMed).

In everyday talk, we often think we understand each other, but research shows miscommunication is rampant. We live in a high-velocity, high-context world that prizes brevity over clarity, and we suffer from the “illusion of transparency,” assuming our own intentions and definitions are clear to others. The result is a crisis of coherence, where people use the same words for radically different ideas.

Subscribe now

In this environment, the simple question “What do you mean?” becomes a powerful (if overlooked) repair tool. Rather than revealing ignorance, it’s a sophisticated intervention that halts misunderstanding and restores precision. As my research will show, in our current polarized certainty, “‘What do you mean?’ is our most powerful technology for rebuilding the public sphere”.

From Socrates’ search for clear definitions to modern conversation analysts, scholars show that asking for clarification is the primary mechanism for semantic repair, a check against biases like the Illusion of Explanatory Depth. In other words, it’s a piece of cognitive infrastructure: a low-cost question that aligns our private maps of meaning and keeps our conversations intelligible.

A Pragmatic Toolkit Across Disciplines

This series treats “What do you mean?” as a phrase and as a toolset grounded in multiple fields. Philosophy and linguistics teach us that meaning is slippery. Socrates used “What is…?” to test hidden assumptions; Wittgenstein reminded us that “the meaning of a word is its use” in context. Austin and Searle’s speech-act theory tells us every utterance has a force (warning, request, etc.), and “What do you mean?” often decodes that force.

Grice’s maxims show we constantly infer “implicatures” (the unsaid), and this question forces speakers to make their intended meaning explicit. Conversation analysts like Sacks and Schegloff identify “What do you mean?” as a classic other-initiated repair: it literally stops the line of talk until understanding is restored.

And Habermas’ theory of communicative action shows that every claim carries unspoken validity tests (truth, rightness, sincerity). Asking “What do you mean?” puts those on the table.

In cognitive science, we see why we need it: we habitually overestimate our understanding (“hallucinating depth”) and fear appearing ignorant. Education and conflict-resolution experts, in turn, note that practicing clear questions builds intellectual humility and reduces escalations.

• Philosophy & Theory: Socrates’s elenchus, Wittgenstein’s language games, Austin and Searle's speech acts, Habermas's discourse – each perspective shows why “What do you mean?” sharpens our concepts and aligns our conversational frameworks.

• Pragmatics & Conversational Repair: Grice’s Cooperative Principle tells us we assume cooperation; when that fails, we seek implicatures. “What do you mean?” makes the implicit explicit. Conversation-analytic research finds it to be a formal repair initiator – more than just “huh?”, it challenges the meaning itself.

• Cognitive Science: Biases like the Illusion of Explanatory Depth mean we think we understand complex ideas until forced to explain them. “What do you mean?” is a simple debiasing move: it punctures our false sense of clarity.

• Civic Discourse: Precision in meaning is civic infrastructure. In democratic deliberation, every argument carries truth and moral claims, and asking “What do you mean?” is how we test them. It’s a practice of listening before judging – a habit of intellectual humility that our polarized times desperately need.

Together, these lenses show that clarification is not soft or frivolous but rigorous. It’s about treating language as fragile machinery: if we don’t ask, we let it break down.

What This Series Offers

Over the coming posts, we will build a practical toolkit. You’ll get:

• A taxonomy of clarification targets: We’ll break “What do you mean?” into precise categories (e.g., definition of a term, scope of a claim, underlying values or evidence, the mechanism or how behind a statement, etc.). Each target will come with real-world examples, so you learn to aim your question accurately.

• Micro-protocols for everyday use: Exact phrasings to try in different settings – in meetings, classrooms, family discussions, or heated debates. (For example, protocols to clarify scope in a work project or separate intent from impact in a tense exchange.) These are ready-made question “scripts” you can adapt on the spot and improve over time.

• A practice in cognitive & civic literacy: Beyond one-off tips, this series encourages a habit: treating clarification as a default move. That means updating your own mental models without defensiveness. It’s a skill for both personal growth and public discourse, a way to ask “What do you mean?” as a normal, respectful step in any conversation, rather than a confrontation. Over time, this builds a culture where precision in language is expected and valued.

Series Outline

Each post will tackle a different aspect of this practice. Tentatively, we will cover:

• Meaning in use: Why clarification works (drawing on pragmatics and speech-act theory).

• Repair in conversation: How breakdowns happen (the mechanics of repair in talk).

• Barriers to asking: Status, face, and cognitive traps that make us stay silent.

• Taxonomy of clarification: The series’ core toolkit, laying out a structured taxonomy of question-types to ask when seeking clarity. It explains categories like definitions (e.g., “What exactly does X mean?”), scope/context (e.g., “Does this apply universally or only in specific cases?”), implications, mechanisms, values, etc., and why each matters (for example, definition questions break down jargon into simpler terms, and context questions narrow the scope to avoid assumptions).

• Protocols for practice: Practical wordings and scenarios for meetings, teaching, and conflicts.

• Scaling up: Clarification as civic practice (thinking beyond one-on-one talk).

• Field test: A prompt to try one clarification instead of a rebuttal this week.

By the end, you’ll have a clearer sense of when and how to ask “What do you mean?” plus a roadmap to revisit. Think of this as a reader’s contract: we’ll explore why it matters, share frameworks to apply it, and invite you to experiment.

Above all, the goal is simple: in any conversation, big or small, reach for understanding before assumption. The payoff is precision in our shared language – the bedrock of constructive dialogue.