Communicate to Play: Pragmatic Reasoning for Efficient Cross-Cultural Communication in Codenames

TL;DR

The paper tackles pragmatic miscommunication across cultures by extending Rational Speech Acts to Cross-Cultural Communication (RSA+C3). It builds Codenames Duet agents through contrastive embedding learning and LLM prompting, then models socio-cultural context and memory to infer interlocutor culture during live interaction. The work demonstrates that RSA+C3 improves collaboration and win rates when players differ in sociocultural background, supported by interactive evaluation on the Cultural Codes dataset and cross-cultural prompts. This approach offers a scalable, interaction-driven pathway to align cross-cultural communication in AI agents, with publicly available code and broader implications for culturally aware language and reasoning systems.

Abstract

Cultural differences in common ground may result in pragmatic failure and misunderstandings during communication. We develop our method Rational Speech Acts for Cross-Cultural Communication (RSA+C3) to resolve cross-cultural differences in common ground. To measure the success of our method, we study RSA+C3 in the collaborative referential game of Codenames Duet and show that our method successfully improves collaboration between simulated players of different cultures. Our contributions are threefold: (1) creating Codenames players using contrastive learning of an embedding space and LLM prompting that are aligned with human patterns of play, (2) studying culturally induced differences in common ground reflected in our trained models, and (3) demonstrating that our method RSA+C3 can ease cross-cultural communication in gameplay by inferring sociocultural context from interaction. Our code is publicly available at github.com/icwhite/codenames.

Figures (10)

• Figure 1: RSA+C3: Rational Speech Acts framework with Cross-Cultural Communication. Here we model interactions in Codenames Duet between the British clue giver and the American guesser. (1) In regular gameplay, the clue giver selects a target and generates a clue without considering the guesser's background. (2) Using RSA+C3, the giver considers what word the guesser may select based on their demographic background and generates a different clue accordingly. The avoidwords will cause the game to end in an immediate loss and the neutralwords have no effect on the success or failure of the game.
• Figure 2: Player modeling using LLM-prompting and trained word embeddings. The efficacy of the Llama2 chat models at simulating human players, including both the giver and guesser, varied across model size and task. Trained word embeddings consistently outperformed untrained word embeddings and generally outperformed LLM-prompting with the exception of the giver clue selection task.
• Figure 3: Comparison of guess accuracy using embeddings trained on cultural splits against baseline GloVe and different cultural training splits. The large difference of 9% on the data of Master+Doctorate cultural split, between the GloVe trained on Master+Doctorate and GloVe trained on the remaining data (i.e. the difference between the orange and green bars) indicates that there are cultural patterns found in the Graduate+Bachelor data that do not occur in the remaining data. There are similar large differences in accuracy between GloVe trained on split and GloVe trained on the other split in the cultural splits on country and politics.
• Figure 4: Target guessing with cultural context. Reranking potential target words based on the probabilities output by the Llama2 model simulating the clue giver and word guesser led to varying levels of guesser-aligned target word selections. Inclusion of cultural context (e.g. political leaning, personality) sometimes improved alignment with the guesser based on model size and selected demographic.
• Figure 5: Clue generation with cultural context. Leaning notably led to an increase in accuracy for giver alignment for the 7B model while including all demographics for the 13B model led to more accurate giver-aligned generations.