On the Shallow Processing (Dis)Advantage: Grammar and Economy

Area of Science:

• Psycholinguistics
• Computational Linguistics
• Cognitive Science

Background:

• The "good-enough" processing framework suggests shallow, extra-grammatical processing is favored over deep, grammatical processing.
• This framework posits that simpler, less resource-intensive processing strategies are generally preferred by readers and listeners.

Purpose of the Study:

• To challenge the prevailing "good-enough" processing theory.
• To present theoretical and experimental evidence that deep, grammatical processing is more efficient for certain linguistic phenomena, specifically anaphora.

Main Methods:

• Theoretical argumentation based on psycholinguistic principles.
• Review and synthesis of existing experimental evidence concerning anaphoric dependency resolution.
• Analysis of processing costs associated with shallow versus deep derivations in language comprehension.

Main Results:

• Evidence indicates an economy hierarchy in anaphora resolution, where deep, grammatical derivations are less costly than shallow, extra-grammatical ones.
• When both shallow and deep derivations are possible, the deep derivation is preferred, contradicting the "good-enough" model's core assumptions.

Conclusions:

• The "good-enough" processing framework, particularly the shallow-deep dichotomy, is an oversimplification.
• A rethinking of the "good-enough" processing model is necessary, considering the efficiency of deep grammatical processing in anaphora resolution.