TED Talks Daily · Society & Culture
Why the best learners talk to themselves | Tania Lombrozo
Cognitive scientist Tania Lombrozo uses the programmer practice of rubber ducking — explaining a code problem step by step to a rubber duck on a desk — as a window into how we learn best in an age of information overload. Drawing on more than 25 years of research, she argues that "learning by thinking" — transforming and integrating knowledge already inside our heads — is the path to deep understanding, rather than immediately reaching for outside information or AI. Through experiments with children and research dating to the 1980s, she makes the case for becoming your own rubber duck.
Listen to this episode Go to the show page
What this episode answers
Why does more information make learning harder?
Without a cognitive foundation to filter and integrate what is needed, more information only deepens confusion — Lombrozo argues that adding outside information alone worsens information overwhelm.
Why does rubber ducking help learning without adding any new information?
Explaining a problem step by step to yourself transforms and integrates existing knowledge inside your head, generating new insight; that conversion process is the core of learning by thinking.
What is the self-explanation effect, and what does it do?
The self-explanation effect is the dramatic boost in learning that comes from explaining things to yourself and posing your own questions; research shows that both children and adults who self-explain learn more deeply and apply their knowledge more broadly than those who do not.
What is the illusion of explanatory depth, and why is it useful for learning?
The illusion of explanatory depth is the phenomenon where actually trying to explain something reveals gaps in knowledge you did not know were there; seeing those gaps clarifies exactly what you need to investigate, making it easier to connect new information to what you already know.
What is the value of questioning yourself before turning to AI or search?
Thinking with your own knowledge first reveals what you are missing, building the capacity to correctly evaluate and connect information found through AI or search; relying on outside sources from the start makes you prone to the illusion of understanding.
Rubber Ducking — Learning Without Any New Information
Software engineers have a practice called rubber ducking. When stuck on a code problem, a programmer explains it step by step, line by line, to a rubber duck sitting on the desk. The duck says nothing. The programmer is talking only to themselves, and no new information enters the picture at any point. Yet somehow the process helps. Cognitive scientist Tania Lombrozo calls this "learning by thinking" and has spent more than 25 years studying the phenomenon.
The Paradox of the Information Age
The spread of the internet, social media, and AI gives us instant access to more information than ever before. But Lombrozo argues that more information can actually make learning harder. Simply adding material from the outside only worsens what she calls "information overwhelm." The better move, she says, is to start with what is already inside your head. Human memory and skills contain enormous stores of knowledge we rarely notice. You may not be able to name off the top of your head how many windows face the front of your home, but you can count them using your visual memory. You can identify the fifth letter of the alphabet by running through it mentally. You know which side of the keyboard the letter Z is on through muscle memory. All of these are examples of learning by thinking — drawing on internal knowledge rather than consulting an external source.
The Self-Explanation Effect — Evidence Built Since the 1980s
In the late 1980s, researchers set out to find what the most successful students had in common. Using the same textbooks and the same exam questions, students still produced vastly different outcomes. The source of the difference was not external information but how students engaged with it. The best students tended to explain things to themselves while studying, pose their own questions, and work toward answers on their own — they were being their own rubber ducks. Subsequent research confirmed repeatedly that self-explanation dramatically improves learning outcomes in both children and adults, and that it helps learners apply what they have learned far more broadly. This is the self-explanation effect.
The Experiment Where Five-Year-Olds Adopted Adult Thinking Patterns
Lombrozo shared one of her own experiments to illustrate the point. Children were shown a reference card with two ducks and asked to choose the closer match: a card with two frogs, or a card with one duck and one fox. Adults typically choose the two frogs, applying the abstract rule "same kind of animal." But children aged five to six tend to choose the duck-and-fox card because at least one duck appears on it, just as on the reference card. Simply showing children the correct answer repeatedly did not shift their choices. But when children were asked to explain why a particular card matched, they began answering in the same pattern as adults — with no feedback and no new information provided. The act of explaining alone changed how they represented the information internally.
The Illusion of Explanatory Depth and the Shopping-List Metaphor
Self-explanation also breaks what Lombrozo calls the "illusion of explanatory depth." Her friend Roxana had taken a college course on climate and natural disasters and believed she could explain how tornadoes form. When she actually tried, she could describe the pressure difference involved but could not put into words how low pressure produces a rotating column of air. "I realized I didn't know anything," Roxana said. Attempting to explain exposed a gap she had not known was there.
Lombrozo compares this to checking what ingredients you already have before writing a shopping list: only when you know what is missing can you figure out what to look for. Turning straight to AI or a search engine outsources the process of converting and integrating knowledge, blurring the line between what you actually understand and what you merely think you understand. Lombrozo closed with a direct instruction: "Don't start with outside information. Look inside first. Ask yourself questions, explain things to yourself. Be your own rubber duck."
Takeaways
- In an age of information overload, looking inward and thinking with your own existing knowledge — before reaching for AI or search — is the true starting point of deep learning.
- Explaining something aloud to yourself shatters the illusion of understanding and surfaces the specific gaps you genuinely still need to fill.
- Making self-explanation a habit transforms and integrates existing knowledge, enabling you to absorb new information more deeply and apply it more broadly.
This article is edited from a machine transcript of the first 30 minutes of the episode. It is not an official summary — please listen to the episode itself.