Machines of Meaning
Davide Nunes, Luis Antunes
TL;DR
The paper reframes the quest for machine language understanding by rigorously defining meaning in a computational setting and arguing for grounding beyond anthropocentric intuition. It surveys symbol theories, grounding, and semantic philosophies to show how symbols acquire referents through world-contexts and agent goals, proposing machines of meaning as capable but contingent on proper grounding. It identifies two core obstacles—the grounding/prediction frame problem and the fixed lexicon limitation of current neural models—and outlines avenues such as incremental online learning, compression-based encodings, and likelihood-free inference to enable open-ended semantics. The work cautions against equating statistical language modelling with true understanding and emphasizes ethical, risk-aware evaluation of AI capabilities. It aims to broaden research directions by integrating multimodal signals, structure from linguistics, and scalable grounding to build more robust, open-ended symbolic systems.
Abstract
One goal of Artificial Intelligence is to learn meaningful representations for natural language expressions, but what this entails is not always clear. A variety of new linguistic behaviours present themselves embodied as computers, enhanced humans, and collectives with various kinds of integration and communication. But to measure and understand the behaviours generated by such systems, we must clarify the language we use to talk about them. Computational models are often confused with the phenomena they try to model and shallow metaphors are used as justifications for (or to hype) the success of computational techniques on many tasks related to natural language; thus implying their progress toward human-level machine intelligence without ever clarifying what that means. This paper discusses the challenges in the specification of "machines of meaning", machines capable of acquiring meaningful semantics from natural language in order to achieve their goals. We characterize "meaning" in a computational setting, while highlighting the need for detachment from anthropocentrism in the study of the behaviour of machines of meaning. The pressing need to analyse AI risks and ethics requires a proper measurement of its capabilities which cannot be productively studied and explained while using ambiguous language. We propose a view of "meaning" to facilitate the discourse around approaches such as neural language models and help broaden the research perspectives for technology that facilitates dialogues between humans and machines.