The Song in Your Head: Why We Assume Other People Understand More Than They Do

Imagine tapping "Happy Birthday" on a table with one finger.

You know exactly what you are tapping. The melody runs through your head while your finger marks the rhythm. You can almost hear the words. It seems hard to imagine that the person sitting across from you cannot recognize it.

Now switch seats.

There is no melody. No pitch. No lyrics. No musical instruments. Someone is knocking on a table.

That difference between what exists inside the sender's mind and what reaches the receiver is one of the most persistent problems in human communication.

In 1990, Stanford University doctoral researcher Elizabeth Newton built a remarkably simple experiment around it.

The Tapping Experiment

Newton assigned participants to roles generally described as tappers and listeners. Tappers selected familiar songs and communicated them by tapping the rhythm with their fingers. Listeners tried to identify the song.

Before learning whether the listener had succeeded, the tappers estimated how likely listeners were to recognize their tune.

They expected success roughly half the time.

Listeners were nowhere close.

There is a small historical wrinkle worth noting because the experiment has been repeated so often in books and articles that two versions of its numbers now circulate. The abstract of Newton's Stanford dissertation reports that listeners identified 2 of 150 tunes, while the version later popularized by Chip and Dan Heath in Made to Stick reports 3 of 120, or 2.5 percent. Newton's dissertation itself is titled The Rocky Road from Actions to Intentions. The important result is consistent across the accounts: people producing the message expected recognition around 50 percent of the time, while listeners recognized only a tiny fraction.

Newton described the underlying problem clearly. Our outward behavior contains far less information than the thoughts accompanying that behavior inside our own minds. We therefore tend to overestimate how easily another person can reconstruct what we mean.

The tapper hears a song.

The listener hears taps.

Much of human misunderstanding begins somewhere in that gap.

Key Terms

Curse of knowledge: Once we possess information, we have difficulty reproducing the mental state of someone who does not possess it.

Common ground: Information, assumptions, experiences, or meanings that participants in a conversation believe they share.

Perspective-taking: The cognitive effort involved in representing what another person can see, know, believe, or understand.

Illusion of transparency: The tendency to overestimate how visible our internal thoughts and emotions are to other people.

What Seems Obvious May Exist Only Inside Your Head

Psychologists Colin Camerer, George Loewenstein, and Martin Weber formally described the curse of knowledge in economic decision-making shortly before Newton's dissertation. Their experiments showed that people who possessed information had difficulty ignoring it when trying to predict the judgments of people who did not possess that information. Even when setting their own knowledge aside would have been useful, they struggled to do so (Camerer et al., 1989).

Raymond Nickerson later reviewed a much larger body of research and reached a related conclusion. When trying to estimate what another person knows, we naturally begin with what we know. That is often efficient. The problem appears when we fail to adjust enough for the differences between our knowledge and theirs (Nickerson, 1999).

Consider a physician explaining a diagnosis, an engineer describing a technical problem, or an insurance professional explaining deductibles and networks. After years of working inside a subject, distinctions that once required effort become automatic.

The expert may say, “This is straightforward.”

Straightforward to whom?

A person who has encountered the concept hundreds of times has compressed an enormous amount of background knowledge into a few words. The newcomer has none of that compression available.

The words may be identical. The information received is not.

This happens outside expertise as well. Families develop shorthand. Couples accumulate years of shared history. Friends remember old arguments and old loyalties. Religious communities inherit stories and symbols. Political groups develop vocabulary whose meaning appears self-evident inside the group.

Then someone from outside enters the conversation.

What one side hears as a complete song may reach the other side as tapping.

Why Would a Mind Work This Way?

There are at least two plausible explanations, and they can operate together.

Hypothesis 1: Perspective-taking is cognitively expensive

Humans are capable of representing other people's minds, but doing so perfectly would be extraordinarily demanding.

To understand what another person hears, I would have to temporarily separate what I know from what you know, estimate your experiences, determine which parts of my background information you share, predict how you interpret particular words, and continually update that model as the conversation progresses.

Research by Boaz Keysar and colleagues suggests that adults do not always perform this computation from the beginning. In experiments tracking people's eye movements during conversation, participants sometimes initially interpreted instructions from their own perspective even when they had enough information to know that the speaker's perspective differed. They could correct themselves, but the egocentric interpretation often appeared first (Keysar et al., 2000).

Prediction: Communication errors should increase when people possess substantially different information, when conversations move quickly, or when people have little opportunity to check whether their assumptions are shared.

The evidence generally fits that prediction.

Hypothesis 2: Assuming shared knowledge is often efficient

There is another possibility.

Perhaps beginning with our own knowledge is usually a sensible shortcut.

Human social life depends heavily on shared information. Families, work groups, coalitions, communities, and cultures accumulate common knowledge over time. If every conversation required us to rebuild another person's entire mental model from zero, communication would become painfully inefficient.

Imagine asking your spouse to hand you the car keys but first having to establish what a car is, what keys are, which car you mean, where you normally leave them, and why you need them.

Shared context allows enormous compression.

From this perspective, assuming some common ground may be less a defect than a normally efficient system that occasionally produces large errors.

Prediction: The problem should become worse when people incorrectly assume that two minds share more history, knowledge, values, or context than they actually do.

That may explain why some communication failures are especially intense between people who believe they already understand one another.

At Work: “I Thought Everyone Knew”

Organizations create perfect conditions for this problem.

Executives have been discussing a strategy for six months. By the time they announce it to employees, they have reviewed presentations, financial models, customer research, competitive threats, board discussions, and dozens of internal conversations.

Then comes the company meeting.

“We’re changing our operating model so we can get closer to the customer.”

The executives may hear six months of reasoning packed inside that sentence.

Employees hear the sentence.

When people later resist the change, leadership may conclude that employees oppose the strategy. Sometimes they do. Sometimes they simply never received enough information to reconstruct why the decision was made.

The same pattern appears between departments.

Engineering thinks sales should understand why a feature cannot simply be added. Sales thinks engineering should understand why the customer needs it immediately. Compliance assumes marketing understands the regulatory risk. Marketing assumes compliance understands the commercial cost of delay.

Each group has information the other does not.

Communication often breaks before disagreement even begins.

A useful operating discipline follows from this: separate disagreement from missing context.

Before concluding that someone rejects your argument, ask whether they actually received the information that made the argument convincing to you.

In Families and Friendships: “You Should Know What I Mean”

Close relationships create enormous common ground. They also create confidence that common ground exists.

That confidence can become dangerous.

“You know why that bothered me.”

“You should have known.”

“I thought it was obvious.”

“You know what I meant.”

Perhaps they did not.

Research on the illusion of transparency shows that people systematically overestimate how well others can detect their internal emotional states. We feel anxiety, frustration, attraction, embarrassment, resentment, or disappointment so strongly that it can seem externally visible. Observers often perceive much less than we imagine (Gilovich et al., 1998).

The person experiencing the emotion has access to everything: memories, physical sensations, expectations, fears, previous conversations.

The other person sees facial expression, tone, behavior, and words.

Again, one person hears the song. The other hears taps.

This does not excuse deliberate inattentiveness or bad behavior. It changes the first diagnostic question. Before deciding that someone ignored information, we should ask whether that information was ever truly available to them.

In School: The Distance Between Knowing and Teaching

Teaching requires something psychologically unusual: a knowledgeable person has to reconstruct ignorance.

A mathematics teacher no longer experiences an equation like a beginner. A historian already knows what happened next. A professor reading an academic paper recognizes terminology, intellectual lineages, and methodological assumptions that a student may encounter for the first time.

Expertise can therefore improve knowledge while making the beginner's perspective harder to recover.

Good teaching repeatedly rebuilds the missing context.

What does the student already know?

Which assumption did I skip?

Which word am I using as if its meaning were obvious?

Could the student explain this back to me without repeating my language?

The test of explanation is not whether the teacher said it clearly. It is whether another mind successfully reconstructed it.

Politics: Shared Facts Cannot Be Assumed

Political communication adds another layer because people may differ in knowledge, identity, trust, values, and information sources simultaneously.

Two people can use the same word — freedom, fairness, security, rights, government, justice — while attaching different histories and assumptions to it.

The resulting argument can look like disagreement over facts when part of the conflict concerns definitions, premises, or trusted sources.

This does not mean political disagreements are merely misunderstandings. People can understand one another perfectly and still prefer different policies because their interests or values differ.

But communication deteriorates quickly when each side assumes its framing is self-evident.

Then disagreement itself becomes evidence of stupidity, dishonesty, or bad intent.

The tapping experiment suggests a less dramatic possibility worth testing first: perhaps the other person is not hearing the same song.

Religion and the Problem of Inherited Meaning

Religious language may provide an even clearer example of compressed cultural knowledge.

Words such as faith, salvation, sin, karma, jihad, dharma, covenant, enlightenment, or grace can carry centuries of theology, interpretation, ritual, and lived experience.

Inside a tradition, a single word may activate an entire network of meaning.

Outside that tradition, the same word may activate something completely different — perhaps learned through news coverage, popular culture, political rhetoric, or experiences with a different religious community.

Both people may sincerely believe the meaning is obvious.

Neither is necessarily hearing what the other intends.

Cultural transmission makes this possible. Humans inherit far more than genes. We acquire languages, rituals, moral systems, categories, stories, technologies, and institutions from other people. Shared culture makes extraordinarily complex cooperation possible, but cultural inheritance also produces different pools of assumed knowledge.

Communication across those boundaries requires more explanation because less can safely be assumed.

Evidence, Interpretation, and Speculation

Evidence: People systematically overestimate how successfully others can infer information available primarily inside their own minds. Newton's tapping experiments, research on the curse of knowledge, studies of conversational perspective-taking, and work on the illusion of transparency all point toward this broader pattern.

Interpretation: Human communication appears to rely heavily on estimates of common ground. Starting from our own knowledge is efficient when those estimates are reasonably accurate and produces misunderstanding when they are not.

Speculation: From an evolutionary perspective, imperfect perspective-taking may reflect a tradeoff rather than a simple cognitive failure. Our ancestors lived in social worlds where repeated interactions, kinship, coalitions, shared subsistence, and local cultural traditions created considerable common knowledge. A communication system that assumes some shared context could have been economical while still leaving humans vulnerable when interacting across very different informational or cultural environments. Establishing that hypothesis would require comparative, developmental, and cross-cultural evidence beyond the tapping experiment itself.

Finding Where Communication Actually Broke

When communication fails, people usually examine the visible end of the process.

Why didn’t they understand?

Why didn’t they listen?

Why are they resisting?

Why can’t they see what is obvious?

A better investigation begins earlier.

What information was available to me that was unavailable to them?

Which part of my explanation depended on something I never actually explained?

Which words did I assume had the same meaning for both of us?

Which experiences made my conclusion seem obvious?

What would someone need to know before my statement would make sense?

This changes communication from a question of repetition to a question of reconstruction.

Saying the same thing louder does not add the missing melody.

What Would Change My Mind?

• Large, well-designed replications showing that people accurately estimate how much information listeners possess would weaken the claim that this is a systematic bias.
• Cross-cultural evidence showing large societies where communicators consistently avoid projecting their own knowledge onto others would suggest that the effect is strongly shaped by culture rather than being a broad feature of human cognition.
• Evidence that expertise reliably improves the ability to reconstruct novice perspectives would challenge the common assumption that knowledge itself contributes to the communication gap.
• Research showing that most serious interpersonal or intergroup conflicts persist unchanged after participants establish genuinely shared definitions, facts, and context would place stronger limits on how much explanatory weight this framework deserves.

Key Takeaways

• Knowing something changes our ability to imagine what it feels like not to know it.
• Communicators routinely possess context that never reaches the listener.
• Shared assumptions make communication efficient, but false assumptions about shared knowledge create predictable errors.
• Miscommunication should be investigated before disagreement is diagnosed.
• Expertise, intimacy, ideology, and shared culture can all increase confidence that other people understand more than they actually do.
• Clear communication requires checking what the other person reconstructed, rather than judging clarity by what we intended to convey.

The Space Between Two Minds

Newton's experiment was small. Its lesson travels unusually well.

We experience our thoughts with extraordinary richness. Other people do not.

They receive words, gestures, facial expressions, fragments of history, and whatever context we manage to transmit. From those pieces, they attempt to reconstruct what exists inside another human mind.

Usually, they reconstruct enough.

Sometimes they reconstruct something entirely different.

The mistake begins when we confuse the richness of what we intended with the amount of information we actually communicated.

At work, at school, across a dinner table, inside a marriage, between political groups, or across religious traditions, the first question after communication breaks down may be surprisingly simple:

Am I sure they heard the song, or did I only tap it?

References & Further Reading

Camerer, C., Loewenstein, G., & Weber, M. (1989). The curse of knowledge in economic settings: An experimental analysis. Journal of Political Economy, 97(5), 1232–1254.

Gilovich, T., Savitsky, K., & Medvec, V. H. (1998). The illusion of transparency: Biased assessments of others’ ability to read one’s emotional states. Journal of Personality and Social Psychology, 75(2), 332–346.

Heath, C., & Heath, D. (2007). Made to stick: Why some ideas survive and others die. Random House.

Keysar, B., Barr, D. J., Balin, J. A., & Brauner, J. S. (2000). Taking perspective in conversation: The role of mutual knowledge in comprehension. Psychological Science, 11(1), 32–38.

Newton, E. L. (1990). The rocky road from actions to intentions [Doctoral dissertation, Stanford University].

Nickerson, R. S. (1999). How we know—and sometimes misjudge—what others know: Imputing one’s own knowledge to others. Psychological Bulletin, 125(6), 737–759.