Reference production and grammatical encoding are jointly investigated.
•Participants described pictures showing an animate agent and an inanimate patient.
•Animacy drives pronominalization and word order.
•Demonstratives are used for referring to non-topics.
•The reference form for the patient depends on the reference form for the agent.
AbstractThis paper presents three picture description experiments investigating grammatical encoding and reference production in German. Participants described pictures showing transitive events with an animate agent and an inanimate patient. A preceding context established one of the referents as topic. The results show that animacy outranks topichood with regard to pronoun choice and choice of word order. Animate entities were pronominalized and produced sentence-initially more often than inanimate ones — independent of their topic status. The use of demonstratives, on the other hand, was mainly driven by topichood, with more demonstratives for non-topics. In addition, the choice of word order depended on the choice of referential expressions. Our findings extend existing evidence against a unified accessibility scale that simultaneously accounts for different types of referential expressions and for word order. We show how the consensus model of language production can be refined to account for our findings without invoking the problematic notion of accessibility.
KeywordsLanguage production
Accessibility
Referential expressions
Word order
Animacy
Topichood
Data availabilityStimuli, datasets and analysis code are available through a public OSF repository:
© 2025 The Authors. Published by Elsevier Inc.
Comments (0)