Large language models generate classical Arabic maqamat with varied quality
Beyond Poetry: Can Large Language Models Generate Classical Arabic Maqamat?
Computation and LanguageArtificial Intelligence
Summary
Large language models (LLMs) are good at creating many types of text, but it’s unclear if they can write classical Arabic stories called maqamat, which have special rhyme, style, and structure rules. The authors tested five different AI models to see how well they could create maqamat using different ways of prompting. They found that the way the models were prompted affected how well they matched the style, with some prompting methods helping models add more rhymes and style features. Different models also varied in how closely their texts matched traditional maqamat conventions. This study helps understand how AI can handle complex cultural and literary styles.
What this means in practice
- •For language technology developers: Improve Arabic language generation tools by integrating prompting strategies that enhance classical maqama style in AI-generated texts.
- •For cultural heritage digitization teams: Use AI models tuned for maqama generation to help create digital literary artifacts preserving classical Arabic prose traditions.
Authors
AbdulRahman A. Morsy, Aya Zirikly
Abstract
Large language models (LLMs) have shown strong performance in creative text generation, yet their ability to produce culturally grounded and stylistically constrained literary forms remains underexplored. Prior work has focused largely on modern language varieties and poetry, while classical prose traditions such as maqama remain largely unstudied. The maqama is a classical literary genre characterized by rhymed prose (saj), dense rhetorical ornamentation, and episodic narrative structure, making it a challenging testbed for evaluating whether LLMs can move beyond surface fluency toward deeper literary competence. In this paper, we present the first controlled evaluation study of maqama generation with LLMs, comparing five models under zero-shot, few-shot, and rule-based prompting, and evaluating outputs through both human annotation and an LLM-as-a-judge framework across dimensions such as rhetorical richness, saj density, structural coherence, and stylistic authenticity. Our results show that prompting strategy plays a strong role in stylistic quality: few-shot prompting most consistently improves saj density, while its effects on rhetoric and coherence vary by model, with the strongest models (GPT-4o and GPT-5.4-mini) benefiting most from rule-based prompting on these dimensions, though zero-shot prompting yields the highest aggregate scores across all five models. We further observe systematic differences between models in stylistic alignment with Arabic maqama conventions, and corroborate our findings with a second independent LLM judge, paired statistical significance testing, and non-LLM proxy measures of saj.