My Thoughts on Humpty Dumpty, Science, and Humility
I have spent all my adult life working in three different worlds. For more than 40 years, I have been a scientist studying human behavior, particularly emotion, nonverbal behavior, and culture.
As a scientist, I have tried to discover things that are reliably true about people.
- What does the evidence tell us?
- What patterns can we identify?
- What theories best account for those patterns?
At the same time, I have spent even longer practicing, teaching, and coaching judo. Coaching asks a very different question than academia. I am not coaching athletes on average. I am coaching this one athlete.
- What makes this person tick?
- What helps this individual perform their best?
- How should I communicate with this particular person?
And during the last 20 years, I have increasingly worked in interviewing and other applied settings.
There again, the problem becomes intensely individual:
- How can I understand this one person better?
- How can I obtain better information from this single person sitting in front of me?
Living in these three worlds has led me to a realization that might initially sound strange:
The more I have learned about people, the humbler I have become about understanding a single person.
Inferential Humility Goes Deeper
In a recent paper and lecture, I introduced the idea of inferential humility, which suggests that behavior is evidence, not proof, that some behaviors contain meaningful signals, others contain noise; that evidence varies in diagnostic value; that alternative explanations remain possible; and that uncertainty cannot be eliminated.
The implication is straightforward: We should not claim more than the evidence justifies.
But I have continued thinking about inferential humility, and I now believe it goes deeper than that.
- It requires humility about our theories. What do our theories actually explain, and what might they leave unexplained?
- It requires humility about our expertise. Deep expertise in one domain does not necessarily translate into expertise about the whole person.
- It requires humility about our perspectives. The way I organize and interpret the world is my way of doing so; it may not be the way the person in front of me does.
- It requires humility about our process. How am I collecting information? How am I reasoning about it? What would cause me to revise my current understanding?
- And ultimately, it requires humility about the extraordinary complexity of the individual.
All of which raises a question I have been pondering: If we know so much about people, why is it still so difficult to understand the person in front of us?
The Problem Is Not Ignorance
Behavioral science knows an extraordinary amount about people.
We have enormous research literatures on emotion, cognition, personality, development, biology, neuroscience, nonverbal behavior, culture, communication, relationships, decision making, influence, interviewing, deception, negotiation, stress, risk, and countless other subjects.
During my career, I have watched scientific journals proliferate across increasingly specialized areas of inquiry. Scientists are producing more and more findings about increasingly specific aspects of human functioning.
Studies produce facts, and we have tons of facts. But facts are not necessarily the same as knowledge. And knowledge is not necessarily the same as wisdom.
I sometimes wonder whether our growth in knowledge—and especially wisdom about people—has kept pace with the enormous growth in facts, because knowing more things about people is not necessarily the same thing as knowing how to understand a person.
The Humpty Dumpty Problem
Science has produced what I think of as a Humpty Dumpty problem.
Science advances in large part by taking complicated phenomena apart. We isolate variables. We narrow questions. We develop increasingly sophisticated methods for studying increasingly specific things in artificial and controlled environments.
In other words, science naturally produces silos. Those silos are tremendously productive. We can develop deep knowledge about emotion, facial expression, physiology, cognition, personality, culture, or any number of other topics.
Universities institutionalize those silos. We have departments of psychology, biology, economics, sociology, anthropology, communication, neuroscience, medicine, business, and many others.
Even within a psychology department, everyone may be studying people, yet one researcher studies cognition, another emotion, another personality, another development, another culture, another social behavior, and another neuroscience.
Then come the specialties, subspecialties, and sub-subspecialties. Over the years, I have sometimes thought of myself and my colleagues as walking silos of information. And here is the problem: the structures that make specialization so productive do not automatically make integration easy (Figure below).
Scientific boundaries are extremely useful tools for studying people. But they are not necessarily boundaries inside people. We have become extraordinarily sophisticated about the pieces of people while remaining much less sophisticated about how those pieces work together in one person.
Eventually, somebody has to put Humpty Dumpty back together.
Specialization Is Not the Problem
I want to be clear about this. Specialization works. It is one of the great engines of scientific progress. And thank goodness for specialists.
If I have a complicated medical problem, I want someone who possesses deep expertise in precisely that problem. We do not need less specialization.
But specialization creates a paradox when we try to move from scientific knowledge to practice. The deeper we go into one piece, the easier it can become to see the person only through that one, single piece.
And the silos in which we work can also shape how we think. Academic debates frequently become organized around mutually exclusive alternatives: Is a phenomenon universal or culturally specific? Is biology or culture more important? Is the face or context more important? Is my theory right and yours wrong?
Constructive scientific debate is essential. Competing theories and disagreement can generate better research and better science. But sometimes I wonder whether we are asking the wrong kind of question.
Instead of simply asking, Who’s right?, perhaps we should also ask: Under what conditions might each be correct? How could two perspectives both contain something valid? How could each be partly right? How could both be incomplete—or even partly incorrect? And most importantly: Can we integrate apparently competing perspectives to produce something better and truer to what people are actually like?
I’ve Been Asking This Question Longer Than I Realized
Looking back over my own career, I realized that I have been wrestling with versions of this problem for a long time. When studies produced apparently conflicting conclusions about whether faces or contexts dominated judgments of emotion, I became interested in what kinds of contexts were being studied and under what conditions each mattered.
In debates about whether emotions and facial expressions were universal or culturally specific, I became interested in how biological and cultural influences could both contribute—and how their relative contributions might differ depending on what aspect of emotion we were trying to explain.
More recently, in studying nonverbal behavior and deception, I have become increasingly interested in moving beyond individual channels toward the larger communication package unfolding across channels and over time.
At times, I think specialization has allowed us—me included—to lose sight of the forest for the trees. That is often necessary when our scientific goal is to understand a particular tree. But practitioners eventually have to deal with the forest.
Practice Inherited the Pieces
The fragmentation does not end when scientific knowledge leaves the university. It follows that knowledge into practice. Professional training is often organized around specialized competencies: facial expressions, body language, emotional intelligence, culture, deception, interviewing, negotiation, personality, influence, communication, and many others.
Ours included. Humintell has participated in this structure too. I am not suggesting that everyone else has been doing something wrong while we somehow got it right. We have all inherited, to some extent, the pieces produced by science. And those pieces can be tremendously valuable. But eventually something happens.
And Then the Door Closes
Years ago, during a discussion about investigative interviewing, someone made a comment that has stayed with me: You can study all the theories. You can learn about facial behavior, voice, nonverbal behavior, language, personality, culture, interviewing techniques, and everything else. Then you and another person walk into the interview room. And then the door closes.
Now it is just the two of you. The science came into the room with you. But the science cannot substitute for the person sitting in front of you. You are no longer trying to understand people in general. You are trying to understand this person.
I have experienced the same thing in judo. I never coached the “average” athlete, the kind of generalized person our scientific findings necessarily describe. Principles of biomechanics mattered. Principles of physiology mattered. Principles of psychology mattered. Principles of training mattered.
But eventually I had to figure out what made this one athlete tick.
Understanding a person is therefore provisional. Scientific findings are an essential starting point, but they do not by themselves determine our understanding of an individual because real people don’t come in silos.
Real People Don’t Come in Silos
A person is simultaneously biological, cognitive, emotional, relational, historical, cultural, and contextual. They have a personality, goals, motives, relationships, a worldview, an organizational environment, immediate circumstances, and a personal history. They communicate verbally and nonverbally at the same time.
And culture is not simply another silo added to the list. People exist within multiple nested contexts. Culture operates through those contexts, shaping expectations, meanings, relationships, interpretations, and behavior. Context does not merely surround behavior. Context helps give behavior its meaning. The person is integrated even though the systems we use to study and teach about that person are specialized.
That leaves science and practice with two related but different intellectual problems.
- Specialization asks: How can we understand this piece more deeply?
- Integration asks: How does this piece operate with all the other relevant pieces when the whole system is functioning within an individual?
Both are necessary.
The Integration Problem Is a Reasoning Problem
Fragmentation can create predictable vulnerabilities. It can encourage single-cue thinking, technique dependence, channel isolation, context neglect, overgeneralization, and confirmation bias. And expertise in one piece does not necessarily translate into expertise in the whole person. Deep expertise should improve judgment. But it should not eliminate humility. High confidence is not necessarily high accuracy.
When we sit across from one individual, the questions change.
- What should I pay attention to?
- Which information is signal and which is noise?
- Which evidence deserves greater weight?
- What if different behavioral channels conflict?
- What does culture change?
- What does context change?
- What alternative explanations remain plausible?
- How confident should I be?
- And perhaps most importantly: What should make me revise my current understanding?
This last question is especially important because understanding a person should evolve as new information becomes available.
This leads me to a distinction that I think is critical: This is not just a knowledge problem. It is a reasoning problem.
We can possess excellent knowledge and still reason poorly with it. Knowing all the pieces is not the same as knowing how to think across the pieces—what matters, what relates, what conflicts, what deserves greater weight, what remains uncertain, and when our understanding should change. Better reasoning may help us put Humpty Dumpty back together again (Figure 2).
Four Forms of Inferential Humility
That brings me back to inferential humility. Four forms seem particularly important:
- Humility about evidence: What does the evidence actually support?
- Humility about knowledge: What might my theory or expertise be missing?
- Humility about the person: Does what is generally true actually apply to this individual?
- Humility about the process: What would make me revise my current understanding?
Humility is not ignorance. It doesn’t mean we know nothing. It means knowing the limits of what we know and regulating our conclusions accordingly. We can make judgments while remaining appropriately uncertain about them.
Inferential humility therefore requires recognizing not only the limits of our conclusions but also the limits of our own perspectives. No single theory, discipline, behavioral channel, or training method owns the whole person.
Evidence Still Has to Anchor Us
None of this means that anything can mean anything. Quite the opposite. Integration without evidence becomes intuition. Intuition can sometimes be useful. But if our goal is science-based practice, our judgments must remain anchored in the strongest available evidence. At the same time: Evidence without integration remains fragmented.
We need both. Humility means respecting what the evidence establishes—and what it does not.
Maybe We’ve Been Asking Only Half the Question
Practitioners naturally ask: What else should I learn? That’s a great question. We should continue learning. Science should continue producing better evidence, and professionals should continue developing expertise. But perhaps there is another question that deserves equal attention:
Perhaps we do not simply need another piece. Perhaps we need a disciplined way of reasoning across the pieces—one that respects scientific evidence, culture, context, individual variation, uncertainty, alternative explanations, and the possibility that our current understanding may need to change.
Humility Disciplines Expertise
Eventually, for many practitioners, the door closes, and there is just you and another human being. After more than 40 years of studying, coaching, interviewing, and trying to understand people, I have come to believe that one of the most important things I can bring into that room—along with everything I have learned—is humility.
Humility doesn’t diminish expertise. It disciplines expertise.
The more I have learned about people, the humbler I have become about understanding a person.
Science advances by taking people apart. Understanding people requires putting the pieces back together.
The challenge is not to abandon those pieces. It’s to understand what each piece tells us, what it does not tell us, and how the relevant pieces fit together when we are trying to understand the one individual in front of us.
So I am left with a question:
How do we bring scientifically grounded knowledge together—within culture and context, and without claiming more than the evidence allows—to better understand the individual in front of us?
And ultimately: What would an integrated approach capable of doing that actually require?