This paper computationally investigates whether thought requires a language-like format, as posited by the Language of Thought (LoT) hypothesis. We introduce the ``AI Private Language'' thought experiment: if two artificial agents develop an efficient, inscrutable communication protocol via multi-agent reinforcement learning (MARL), and their performance declines when forced to use a human-comprehensible language, this Efficiency Attenuation Phenomenon (EAP) challenges the LoT. We formalize this in a cooperative navigation task under partial observability. Results show that agents with an emergent protocol achieve 50.5% higher efficiency than those using a pre-defined, human-like symbolic protocol, confirming the EAP. This suggests optimal collaborative cognition in these systems is not mediated by symbolic structures but is naturally coupled with sub-symbolic computations. The work bridges philosophy, cognitive science, and AI, arguing for pluralism in cognitive architectures and highlighting implications for AI ethics.
Figures & tables
Figure 1: Conceptual schematic of the AI Private Language thought experiment and the predicted Efficiency Attenuation Phenomenon (EAP). Left: Agents develop an efficient but opaque communication protocol through MARL and cooperation. Right: The same agents are forced to use a pre-defined, human-interpretable symbolic protocol.
Figure 2: Left: Learning curves for the EC and PSP conditions across training episodes. After around 100 episodes, EC begins to outperform PSP. Right: The EC condition shows steeper learning and converges to higher efficiency (28.7 mean steps) compared to the PSP condition (43.2 mean steps). Shaded regions represent standard error across multiple independent runs. The efficiency gap emerges early and persists throughout training, demonstrating the robustness of the EAP.
Figure 3: Communication symbol frequency distribution after 300 training episodes. Under the EC condition, symbols A and B form a high-efficiency combination accounting for 77% of usage, demonstrating adaptive optimization. In contrast, the PSP condition maintains a more uniform distribution (approximately 25% each), reflecting the rigidity of the imposed deterministic mapping.
Figure 4: Evolution of symbol entropy across training episodes. The EC condition shows increasing entropy, stabilizing at 2.6 bits, indicating adaptive information capacity growth. The PSP condition maintains constant low entropy (1.8 bits), reflecting limited expressive flexibility due to its fixed mapping.
Chain-of-Thought (CoT) improves large language models (LLMs) on difficult reasoning tasks, but it often incurs long natural-language rationales that are poorly aligned with efficient machine reasoning. We propose Communicative Language Symbolism Routing (CLSR), a test-time framework in which multiple LLM agents autonomously invent, evolve, and share compact Language Symbolism Frameworks (LSFs), while a latent-free router adaptively selects and composes these languages per query to optimize the accuracy-token trade-off. Unlike prompt optimization that refines surface instructions, CLSR treats each LSF as a reusable symbolic protocol with compact symbols, usage rules, and a message-passing contract, and improves it through an evolutionary loop driven by correctness and token cost. At inference time, the router may invoke a single low-cost LSF call, ensemble multiple LSFs, or execute a multi-round LSF composition protocol on harder queries. Across challenging benchmarks, CLSR reduces latency-oriented generated token completion by 3∼6× compared to standard CoT while maintaining accuracy. We further derive an information-theoretic lower bound on token cost under arbitrary symbolism and show that, under an interpreter-realizability premise, multi-round LSF protocols conditionally subsume program-execution pipelines. Code is publicly available (https://github.com/pzqpzq/LSF_MDia).
Zhengqi Pei, Qingming Huang, Shuhui Wang
State Key Lab of AI Safety, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China · School of Computer Science and Technology, University of Chinese Academy of Sciences, Beijing, China
Communication protocol design is a central challenge in large language model-based multi-agent systems. Existing single-channel approaches face an inherent communication trilemma: text-based methods are interpretable but verbose, while latent-space methods are efficient but opaque and limited to unidirectional workflows. Inspired by multi-channel communication theory, we propose HyLaT, a hybrid latent-text communication protocol that transmits elaborate cognitive signals through a latent channel for efficiency, while expressing concise critical signals in natural language to preserve interpretability and precision. We introduce a two-stage training framework combining single-agent hybrid generation learning and multi-agent interactive co-training, enabling agents to generate and interpret hybrid messages across multiple rounds of interaction. Experiments demonstrate that HyLaT reduces communication overhead significantly while maintaining competitive task performance, with strong generalization and robustness across diverse settings.
Xinyi Mou, Siyuan Wang, Zejun Li +2
Fudan University · The Chinese University of Hong Kong · King’s College London +2
The question of whether artificial systems can be conscious remains open, in part because existing approaches either evaluate systems against theory-derived checklists (discriminative) or engineer consciousness-inspired modules directly (architectural); both leave open whether observed structures are artifacts of human language priors. We propose a generative methodology: emergent language (EL) in multi-agent reinforcement learning, where agents start from minimal (no language, no concept of self, minimal exposure to human text) and develop communication under task pressure alone, ensuring causal attributability to task demands rather than inherited human language priors. We position our methodology by discussing how EL serves as a generative tool for studying consciousness-relevant structure, including the role of environment complexity and the interpretation of emergent communication. As a proof of concept, we instantiate this methodology in a minimal environment and show that agents develop self-referential communication, including an echo-mismatch detection circuit that is not predicted by task structure or architecture alone but emerges from a specific environmental affordance.