The Problem: Explaining Reading Comprehension

You read a sentence: “The spy hid the documents in the violin case before the concert.” In less than a second, your brain has done something remarkable. You haven’t just decoded words—you’ve constructed a mental scenario involving espionage, a specific hiding place, timing constraints, and probably a crowded concert hall.

How does this happen? How do black marks on a white page become a vivid mental world?

In 1988, psychologist Walter Kintsch proposed the Construction-Integration (CI) model, which remains the most influential and empirically supported theory of reading comprehension in cognitive science. Understanding this model reveals what’s actually happening in your brain when you read—and why some readers comprehend so much better than others.

What Research Shows: The Two Phases of Comprehension

The construction-integration model proposes that comprehension unfolds through two distinct phases that cycle rapidly as you read.

The Construction Phase: Messy Activation

When you encounter a word or phrase, your brain doesn’t carefully select only the relevant meaning. Instead, it activates everything that might be relevant—a messy, unfocused explosion of associations.

This construction phase is “dumb” in a sense—it follows simple rules of spreading activation without considering context or coherence. Everything connected to the input gets a boost of activation, regardless of relevance.

This might seem inefficient, but it’s actually clever. By initially over-activating, the system ensures that the right meanings are always somewhere in the activation pattern. The filtering comes next.

The Integration Phase: Coherent Selection

After the messy construction phase comes integration. Here, the activated elements interact with each other through a process called spreading activation with constraint satisfaction.

Elements that fit together—that support each other and create coherent meaning—strengthen each other’s activation. Elements that conflict or don’t fit fade away. It’s like a room full of conversationalists where compatible ideas find each other and irrelevant ones drift to the periphery.

After integration, you’re left with a coherent representation of what the text means. This representation then feeds into the next construction phase as you continue reading, and the cycle repeats.

The Deeper Analysis: Three Levels of Representation

The Kintsch model proposes that readers build three distinct levels of mental representation during comprehension. Understanding these levels explains why some readers remember everything while others forget immediately.

Level 1: The Surface Structure

This is the exact wording—the literal string of words you read. Surface structure is fleeting. You typically retain it only for a few seconds, just long enough to process it. This is why you can’t usually recall the exact words of a paragraph you read minutes ago, even if you understood it perfectly.

Level 2: The Textbase

The textbase captures the meaning of what the text explicitly states, stripped of exact wording. It’s a network of propositions—idea units—derived from the text.

For “The spy hid the documents in the violin case,” the textbase would include propositions like: HIDE(spy, documents), IN(documents, violin case), and their connections. The textbase preserves meaning without preserving form.

Many readers stop here. They can recall what the text said, summarize its main points, answer factual questions. But they haven’t achieved deep comprehension.

Level 3: The Situation Model

The situation model is where real comprehension lives. It’s not a representation of the text—it’s a representation of the situation the text describes.

For our spy sentence, a rich situation model might include: the spy’s anxiety, the size and weight of the violin case, the concert hall setting, other people who might notice, the risk of discovery, what happens to spies who get caught. None of this is in the text—it comes from your background knowledge integrating with text information.

Why Background Knowledge Matters So Much

The CI model explains why background knowledge is so crucial for comprehension—and why you can’t simply “learn to read” and then read anything.

During the construction phase, your prior knowledge activates alongside text information. During integration, knowledge and text fuse together to build the situation model. Without relevant background knowledge, you can only build a thin textbase—the words make sense, but the deeper meaning doesn’t form.

This has profound implications. Two readers with identical “reading skills” will comprehend the same text very differently depending on their background knowledge. Reading comprehension isn’t purely a skill—it’s a skill that operates on knowledge.

Implications for Readers: Using the CI Model

Understanding how your brain builds meaning suggests several practical strategies.

Activate Knowledge Before Reading

Since comprehension depends on integrating text with prior knowledge, you can improve comprehension by deliberately activating relevant knowledge before you read. Think about what you already know about the topic. Generate questions you hope the text will answer. This primes the construction phase to pull in relevant connections.

Monitor for Situation Model Failure

When you realize you’re just following words without building a mental movie, that’s a sign your situation model isn’t forming. Stop and ask: Do I understand the situation being described, or just the words? If you can’t visualize or explain the scenario, you’re stuck at the textbase level.

Build Knowledge to Build Comprehension

If you consistently struggle to comprehend texts in certain domains, the problem might be knowledge, not skill. Reading more in that domain—even easier texts—builds the knowledge base that enables future comprehension. The science of reading increasingly emphasizes knowledge building as essential to comprehension development.

Use Inference as a Comprehension Check

After reading a section, try to answer questions the text didn’t explicitly address. If you can make reasonable inferences, you’ve built a situation model. If you can only recall what was stated, you’re stuck at the textbase. This self-testing reveals the depth of your comprehension.

What This Means for Your Reading

The construction integration model reveals that comprehension isn’t a single thing that happens once—it’s a cyclical process that builds meaning through repeated construction and integration phases, ultimately creating a mental model of the situation described.

You’re not a passive receiver. Comprehension requires your active contribution. Your knowledge, your inferences, your mental model-building—these aren’t extras; they’re essential to understanding anything beyond the surface.

Comprehension depth varies. You can “understand” text at very different levels. Recognizing when you’re stuck at the textbase (knowing what it said) versus achieving a situation model (understanding the reality described) is crucial metacognitive awareness.

Knowledge enables skill. Reading comprehension isn’t purely a technique you can master independently of content. The more you know, the more you can understand. This is why wide reading across topics improves comprehension generally—you’re building the knowledge base that feeds future situation models.

Kintsch’s model shows that your brain is a meaning-making machine of remarkable sophistication. Every time you read, you’re running construction and integration cycles, building layered representations, fusing text with knowledge. Understanding this process gives you leverage: you can work with your brain’s natural mechanisms rather than fighting against them.