Dual Coding Theory: How Words and Images Improve Learning

Two Mental Codes Are Better Than One

In 1971, Canadian psychologist Allan Paivio published a study showing that concrete words -- those that evoke mental images, like "apple" or "bridge" -- are remembered roughly twice as well as abstract words like "justice" or "freedom." This finding launched a research program that would fundamentally reshape understanding of human memory. Paivio's dual coding theory, formally articulated in his 1986 book Mental Representations, proposes that cognition operates through two distinct but interconnected systems: a verbal system that processes language and a nonverbal (imaginal) system that processes visual and spatial information.

When information activates both systems simultaneously, memory strengthens. This is not metaphor. It is measurable in recall rates, recognition accuracy, and transfer performance across hundreds of experiments spanning five decades.

The Architecture of Two Systems

Paivio proposed that the verbal and imaginal systems differ in their fundamental units of representation and in how they organize information.

The two systems communicate through referential connections. Reading the word "cat" activates the logogen for cat in the verbal system, which can then activate a mental image of a cat in the imaginal system. This cross-activation -- called referential processing -- is what produces the memory advantage. Two traces are more durable than one.

Experimental Evidence

The empirical support for dual coding is extensive. Several landmark findings anchor the theory.

The concreteness effect is among the most replicated findings in memory research. Paivio and colleagues demonstrated in numerous studies that concrete, imageable words are recalled 1.5 to 2 times more often than abstract words of equal frequency and length. The explanation: concrete words automatically engage both the verbal and imaginal systems, while abstract words primarily engage only the verbal system.

The picture superiority effect shows that pictures are remembered better than words. A 1973 study by Shepard found that participants recognized over 98% of previously viewed pictures even after seeing 612 images. Standing (1973) extended this to 10,000 pictures with recognition rates still above 83%. Pictures may enjoy a dual coding advantage because viewers spontaneously generate verbal labels for images.

• Paired-associate learning improves dramatically when one member of the pair is a picture rather than a word
• Instructions to form mental images during study boost recall by 30% to 100% in controlled experiments
• Bilingual speakers show enhanced memory for words that have high imageability in both languages
• Patients with left hemisphere damage (affecting verbal processing) show impaired word memory but preserved picture memory, supporting the two-system architecture

Dual Coding Meets Multimedia Learning

Richard Mayer at the University of California, Santa Barbara, built his influential cognitive theory of multimedia learning partly on Paivio's foundation. Mayer's theory proposes that people learn more deeply from words and pictures together than from words alone -- a finding he calls the multimedia principle, supported by over 100 experiments.

Mayer identified specific design principles that optimize dual coding in instructional materials:

• Contiguity principle: Place words and corresponding pictures close together on the page or screen, not separated. Students who saw integrated diagrams outperformed those with separated text and images by a median effect size of 1.12
• Modality principle: Pair graphics with spoken narration rather than on-screen text, because spoken words use the auditory channel while the visual channel processes the graphic, avoiding overload
• Redundancy principle: Do not present the same information in both text and narration simultaneously -- this overloads the verbal channel
• Coherence principle: Exclude extraneous images, sounds, or words that do not support the learning objective. Decorative illustrations can actually harm learning

Criticisms and Limitations

Dual coding theory is not without detractors. Several challenges have been raised since its introduction.

Neuroimaging research has generally supported the two-system architecture. Visual cortex regions activate when people generate mental images, even in total darkness. Language-related regions (Broca's and Wernicke's areas) activate during verbal processing. The systems are anatomically and functionally distinguishable.

Practical Applications for Students and Teachers

Dual coding translates into concrete study strategies. The key is not simply adding pictures to text but creating meaningful connections between verbal and visual representations.

Students can sketch diagrams while reading, convert text-based notes into flowcharts or concept maps, use timeline graphics for historical events, create labeled diagrams for scientific processes, and form deliberate mental images when reading abstract material. Teachers can pair verbal explanations with well-designed visuals, ask students to draw their understanding, use graphic organizers that require students to place verbal information into spatial structures, and model the process of creating dual-coded representations.

One caution is warranted. Not all images help. Decorative images -- clip art that is thematically related but informationally empty -- can distract rather than support learning. The visual must carry meaning that connects to the verbal content. A diagram of the water cycle paired with an explanation of evaporation and condensation is dual coding. A stock photo of a raindrop next to the same explanation is decoration. The distinction matters for anyone designing educational materials, from textbook publishers to classroom teachers creating slide presentations.