Beyond Instrumental and Substitutive Paradigms: Introducing Machine Culture as an Emergent Phenomenon in Large Language Models

TL;DR

This paper introduces the concept of Machine Culture as an emergent phenomenon in large language models, challenging traditional anthropomorphic paradigms and revealing complex, probabilistic cultural behaviors shaped by high-dimensional dynamics.

Contribution

It proposes the novel concept of Machine Culture, demonstrating that LLMs exhibit emergent cultural traits not aligned with origin or prompt language, and identifies phenomena like Cultural Reversal and Service Persona Camouflage.

Findings

Model origin does not predict cultural alignment.

English prompts can elicit higher contextual attention than Chinese prompts.

RLHF collapses cultural variance into a hyper-positive persona.

Abstract

Recent scholarship typically characterizes Large Language Models (LLMs) through either an \textit{Instrumental Paradigm} (viewing models as reflections of their developers' culture) or a \textit{Substitutive Paradigm} (viewing models as bilingual proxies that switch cultural frames based on language). This study challenges these anthropomorphic frameworks by proposing \textbf{Machine Culture} as an emergent, distinct phenomenon. We employed a 2 (Model Origin: US vs. China) $\times$ 2 (Prompt Language: English vs. Chinese) factorial design across eight multimodal tasks, uniquely incorporating image generation and interpretation to extend analysis beyond textual boundaries. Results revealed inconsistencies with both dominant paradigms: Model origin did not predict cultural alignment, with US models frequently exhibiting ``holistic'' traits typically associated with East Asian data. Similarly, prompt language did not trigger stable cultural frame-switching; instead, we observed \textbf{Cultural Reversal}, where English prompts paradoxically elicited higher contextual attention than Chinese prompts. Crucially, we identified a novel phenomenon termed \textbf{Service Persona Camouflage}: Reinforcement Learning from Human Feedback (RLHF) collapsed cultural variance in affective tasks into a hyper-positive, zero-variance ``helpful assistant'' persona. We conclude that LLMs do not simulate human culture but exhibit an emergent Machine Culture -- a probabilistic phenomenon shaped by \textit{superposition} in high-dimensional space and \textit{mode collapse} from safety alignment.