A Corpus-based Grammatical Investigation towards a Computational Identification of Metaphor

The paper titled "A Corpus-based Grammatical Investigation towards a Computational Identification of Metaphor" focuses on the computational identification of metaphors in the Holy Qur’ān, particularly in Sūrat Hūd. The primary research question is to establish a grammatical criterion for identifying metaphors that can be applied computationally.

The study aims to create a framework for computationally detecting metaphorical expressions within the text of the Holy Qur’ān. It seeks to determine specific grammatical markers that indicate metaphorical usage, which can then be implemented in computer software for metaphor identification. The scope is centered on the grammatical structure of metaphors, extending the cognitive theory of metaphor by Lakoff and Johnson, and integrating insights from Arabic exegeses and computational linguistics.

The methodology involves a corpus-based analysis, utilizing both qualitative and quantitative approaches. The researchers manually identify candidate metaphors in the corpus using authentic Qur’ānic exegeses and then apply a set of grammatical criteria to these candidates. This process determines the degree of metaphoricity, which is a measure of how strongly an expression is metaphorical based on its grammatical structure.

The key findings indicate that certain grammatical structures, such as the combination of verbs with specific types of nouns or adjectives, are reliable indicators of metaphorical usage. These structures are graded on a continuum of metaphoricity, from literal to highly metaphorical expressions. The study also highlights the potential for developing software that can accurately detect metaphors in the Holy Qur’ān by applying these grammatical markers.

The study concludes that the identification of metaphors through grammatical markers is feasible and that this approach can significantly contribute to computational linguistics. The findings underscore the importance of integrating traditional linguistic analysis with modern computational methods, particularly in the context of sacred texts like the Holy Qur’ān. The implications are significant for both linguistic theory and the development of advanced tools for text analysis.