Reading comprehension: English Language Notes for TOPIK

Reading Comprehension – Advanced Study Notes

Audience: Graduate students in Linguistics, Cognitive Science, or Applied Linguistics
Context: Preparing for the TOPIK (Test of Proficiency in Korean) with a focus on English‑language reading comprehension.

Goal of these notes – to provide a MIT‑level treatment that starts from the fundamentals of language and cognition, builds a rigorous theoretical framework, and culminates in practical, research‑based strategies for mastering reading comprehension at an advanced university level.

1. What Is Reading Comprehension?

Term	Definition	Key Features
Reading	The act of decoding written symbols into linguistic units (words, phrases) and linking them to mental representations.	Visual → phonological → lexical → syntactic → semantic.
Comprehension	The construction of a coherent mental model of the text that integrates prior knowledge, text structure, and inference.	Requires information integration across multiple levels.
Reading Comprehension	The synthesis of decoding, linguistic parsing, and model construction, yielding the ability to answer questions, summarize, critique, or apply the text.	A multi‑layered, interactive process.

2. First Principles

Information Theory
- Shannon’s entropy: H = - Σ p(x) log₂ p(x) measures the uncertainty of a symbol.
- Redundancy: Language is highly redundant (≈75 % in English), providing multiple cues for comprehension.
Cognitive Psychology
- Working Memory (WM): Capacity ~ 7 ± 2 chunks (Baddeley, 1992).
- Long‑Term Memory (LTM): Stores lexical, grammatical, and world knowledge.
- Attention & Monitoring: The metacognitive component that flags comprehension gaps.
Computational Linguistics
- Predictive Coding: The brain actively predicts upcoming words; mismatch triggers error signals (Friston, 2005).
- Bayesian Inference: Prior knowledge P(K) and evidence P(E|K) combine to yield posterior probability P(K|E) of a meaning.

3. Theoretical Models of Reading

Model	Core Idea	Strengths	Weaknesses
Bottom‑Up (Lexical‑Driven)	Decoding → Word recognition → Syntax → Semantics.	Simple, focuses on decoding.	Ignores context and prediction.
Top‑Down (Context‑Driven)	Prior knowledge → Prediction → Lexical matching.	Explains speed reading, context effects.	Over‑reliance on prior knowledge may lead to misinterpretation.
Interactive‑Compensatory	Bottom‑up and top‑down processes mutually influence each other.	Balances decoding & context.	Complexity in modeling.
Cognitive Load Theory	WM limits reading speed and comprehension.	Explains effect of text difficulty.	Doesn’t address linguistic structure.
Predictive Coding Model	Hierarchical generative model with prediction errors.	Integrates neural evidence.	Requires computational modeling.

Takeaway: In advanced reading, the interactive‑compensatory framework is most useful because it accounts for both decoding errors and contextual predictions, especially for a non‑native learner tackling complex English texts.

4. Linguistic Levels Involved

Level	Typical Processing Task	Key Tools
Orthographic	Letter‑word mapping	Word‑frequency lists, sub‑lexical decoding
Morphological	Inflection, derivation	Morphological parsing, compounding
Lexical	Word meaning & sense disambiguation	Word‑net, semantic fields
Syntactic	Sentence structure	Constituency trees, dependency graphs
Semantic	Contextual meaning, inference	Pragmatic rules, world knowledge
Discourse	Coherence & cohesion	Cohesion markers, discourse relations

5. Cognitive Processes: A Step‑by‑Step Breakdown

Visual Perception & Attention
- Eye‑movement patterns: Fixations