Linguistic-Memetic Intelligence

This paper introduces the theory of linguistic-memetic intelligence (LMI): the culturally evolved intelligence carried by language and instantiated through human, institutional and machine hosts. It proposes that language contains memetic algorithms, recurring transformations that act on representations, alter discourse states and activate further transformations. Their organized interaction supports reasoning, explanation, persuasion, coordination and prediction across generations of hosts. The paper defines the central concepts of LMI, memetic algorithms and hosts; examines how linguistic transformations are acquired, selected and executed; considers the relationship between LMI and large language models; and outlines an empirical programme for identifying and testing these transformations in textual corpora. It also discusses the possibility of constructing artificial systems in which explicit, learnable transformations operate recursively over representations expressed in language.