Language and Cognition - Interfaces

Core Idea

Language interfaces with multiple cognitive systems: phonological processing (sound), semantics (meaning), pragmatics (communicative intent), working memory, attention, and motivation. The nature of these interfaces constrains linguistic structure. For example, grammatical complexity is constrained by working memory capacity; garden-path effects reveal interface constraints between parsing and real-time comprehension. Understanding interfaces explains why some structures are universal, rare, or absent in languages.

How It's Best Learned

Study classic interface phenomena (garden-path sentences, right-branching preference, resumptive pronouns in position violations). Read psycholinguistic evidence on parsing strategies. Examine linguistic universals explained by interface constraints rather than core grammar. Learn experimental methods (eye-tracking, ERP, fMRI) that reveal processing difficulty. Consider evolutionary and developmental constraints on language.

Common Misconceptions

• Thinking interfaces are secondary to core grammar; interfaces often explain why grammatical principles take their specific form.
• Assuming all linguistic variation is grammatically encoded; much variation reflects interface properties and processing strategies.

Explainer

Linguistics traditionally separated syntax (the formal structure of language) from cognition (how the mind processes language). But interfaces between formal structure and cognitive processing are where much of linguistics' explanatory power lies. Language-cognition interfaces study how cognitive constraints shape linguistic structure, and how linguistic structure is adapted to serve cognitive capacities.

Key interface constraints:

Working memory: The capacity to hold and manipulate information is limited (roughly 4-7 items). This constrains:

• Embedding depth: Center-embedding (NP [VP [VP ...]]) requires keeping incomplete dependencies open. After one or two levels, capacity is exceeded. Languages typically limit natural center-embedding to 1-2 levels.
• Structure complexity: Simpler structures that don't require deep nesting are preferred. This may explain why some transformations exist (raising removes nesting; extraposition moves material rightward to reduce center-embedding).
• Garden-path effects: Initial parses that exceed memory are abandoned; reanalysis occurs. Frequent structures are preferred even if ambiguous, because they're less effortful.

Incremental processing: The mind processes language word-by-word in real-time, not all at once. This means:

• Right-branching is preferred: In right-branching structures, the head appears first, and modifiers follow incrementally. In left-branching, modifiers come first, unintegrated. Right-branching imposes lower memory load.
• Frequency effects: High-frequency structures are recognized faster, suggesting rapid pattern-matching based on exposure.
• Gaps create processing cost: Extracting elements (forming long-distance dependencies) creates gaps that must be held in memory until resolution. This is cognitively costly.

Attention and salience: Cognitive systems prioritize salient information:

• Preferred argument structure: Agents (typically subjects) are more salient than patients; active voice highlighting agents is preferred over passive.
• Word order effects: Agents appearing first is cross-linguistically preferred, suggesting SVO and SOV orders are more natural than VSO or OSV.

Pragmatic constraints: Communicative intent shapes structure:

• Information structure: New information appears in different positions depending on pragmatic context; languages vary in how they mark topic and focus.
• Politeness and register: Formality constraints shape syntax (politeness markers, indirectness).

Examples of interface explanations:

The garden-path effect in "The horse raced past the barn fell": The parser prefers the simpler analysis (raced as main verb), even though it leads to a dead-end. This isn't a grammatical principle but a processing strategy reflecting frequency and memory limits.

Resumptive pronouns: In "The boy that I saw him," a pronoun appears in the gap position. This violates formal island constraints but reduces processing cost. Many languages allow resumptives in difficult structures, suggesting they're licensed by interface constraints.

Right-branching universality: Cross-linguistically, structures where constituents branch rightward are more frequent than left-branching. This likely reflects the processing advantage: head-initial structures allow incremental, efficient processing.

Understanding interfaces explains not just why certain structures are universal, but why languages vary in predictable ways. A language might satisfy a working-memory constraint through extraposition (moving material rightward), raising (removing nesting), or resumptive pronouns (reducing gaps). Different strategies, same constraint.

This perspective bridges traditionally separated fields. Formal linguists studying what's grammatical, psycholinguists studying what's processable, and cognitive scientists studying capacity limits are all studying the same phenomena from different angles. Integrating these perspectives provides explanatory power neither field alone can achieve.