# Learning to refer informatively by amortizing pragmatic reasoning

arXiv: https://arxiv.org/abs/2006.00418 · HTML: https://mlsanity.com/paper/2006.00418/

## Key facts

- Cited by 17 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 31 May 2020.

## Built on by (most cited first, 17 of 17)

- [From Word Models to World Models: Translating from Natural Language to the Probabilistic Language of Thought](https://mlsanity.com/paper/2306.12672/) (2023)
- [Talk Less, Interact Better: Evaluating In-context Conversational Adaptation in Multimodal LLMs](https://mlsanity.com/paper/2408.01417/) (2024)
- [Reference-Centric Models for Grounded Collaborative Dialogue](https://mlsanity.com/paper/2109.05042/) (2021)
- [Evaluating statistical language models as pragmatic reasoners](https://mlsanity.com/paper/2305.01020/) (2023)
- [Learning to Mediate Disparities Towards Pragmatic Communication](https://mlsanity.com/paper/2203.13685/) (2022)
- [Language Models are Bounded Pragmatic Speakers: Understanding RLHF from a Bayesian Cognitive Modeling Perspective](https://mlsanity.com/paper/2305.17760/) (2023)
- [Calibrate your listeners! Robust communication-based training for pragmatic speakers](https://mlsanity.com/paper/2110.05422/) (2021)
- [CoGen: Learning from Feedback with Coupled Comprehension and Generation](https://mlsanity.com/paper/2408.15992/) (2024)
- [Are LLMs good pragmatic speakers?](https://mlsanity.com/paper/2411.01562/) (2024)
- [Generating Pragmatic Examples to Train Neural Program Synthesizers](https://mlsanity.com/paper/2311.05740/) (2023)
- [Communicating with Speakers and Listeners of Different Pragmatic Levels](https://mlsanity.com/paper/2410.05851/) (2024)
- [NeLLCom-Lex: A Neural-agent Framework to Study the Interplay between Lexical Systems and Language Use](https://mlsanity.com/paper/2509.22479/) (2025)
- [Integrating Neural and Symbolic Components in a Model of Pragmatic Question-Answering](https://mlsanity.com/paper/2506.01474/) (2025)
- [Surprisal Minimisation over Goal-directed Alternatives Predicts Production Choice in Dialogue](https://mlsanity.com/paper/2605.00506/) (2026)
- [Color Overmodification Emerges from Data-Driven Learning and Pragmatic Reasoning](https://mlsanity.com/paper/2205.09172/) (2022)
- [When Contextual Inference Fails: Cancelability in Interactive Instruction Following](https://mlsanity.com/paper/2603.19997/) (2026)
- [Modeling Human-Like Color Naming Behavior in Context](https://mlsanity.com/paper/2604.25674/) (2026)

## Abstract (arXiv)

A hallmark of human language is the ability to effectively and efficiently convey contextually relevant information. One theory for how humans reason about language is presented in the Rational Speech Acts (RSA) framework, which captures pragmatic phenomena via a process of recursive social reasoning (Goodman & Frank, 2016). However, RSA represents ideal reasoning in an unconstrained setting. We explore the idea that speakers might learn to amortize the cost of RSA computation over time by directly optimizing for successful communication with an internal listener model. In simulations with grounded neural speakers and listeners across two communication game datasets representing synthetic and human-generated data, we find that our amortized model is able to quickly generate language that is effective and concise across a range of contexts, without the need for explicit pragmatic reasoning.