Language comprehension is the process of extracting meaning from spoken or written language — arguably the most complex and rapid cognitive achievement humans perform routinely. In normal conversation, we process about 150-200 words per minute, accessing word meanings, building syntactic structures, computing sentence meanings, drawing inferences, and updating our discourse model, all in real time with minimal conscious effort.
Levels of Processing
Comprehension involves multiple levels of processing operating in parallel. Lexical access retrieves word meanings from the mental lexicon. Syntactic parsing builds hierarchical phrase structures. Semantic interpretation computes the meaning of phrases and sentences. Pragmatic interpretation determines the speaker's intended meaning in context. Discourse processing integrates sentence meanings into a coherent representation of the text or conversation.
Incremental Processing
A fundamental finding is that comprehension is incremental: listeners and readers do not wait until the end of a sentence to begin interpretation. Eye-tracking studies show that readers fixate words that disambiguate sentences immediately, visual world experiments show that listeners direct their eyes to likely referents before they are named, and ERP studies show that semantic anomalies are detected within 400 ms of word onset. This incrementality means that comprehenders are actively predicting upcoming input based on what they have already processed.
Language comprehension is increasingly understood as a predictive process. Comprehenders generate expectations about upcoming words, structures, and meanings based on context. These predictions are reflected in reduced processing (faster reading times, smaller N400) for predictable words and increased processing for unpredictable ones. Predictive processing may explain the speed and efficiency of comprehension: by pre-activating likely upcoming information, the system reduces the computational demands of processing each word from scratch.