Truth, meaning, and intention in Large Language Models
Abstract
The arrival of commercially available chatbots that are able to produce human-like language has made philosophical thought experiments come to life. Such language models produce mostly coherent language, but how do their outputs compare to what we usually assume about language uttered by a human agent? In particular, can their outputs be assumed to be related to truth, and can they be considered meaningful in any sense? In the first part of the thesis, some basic philosophical concepts are explored to provide a background, such as the nature of truth, meaning, and intention. In the second part, a high-level description of Large Language Models is provided. In the third part, these subjects meet, and the outputs of Large Language Models are evaluated according to philosophical concepts. A major issue is the question of grounding: are the outputs of these models grounded in the real world? One argument from different areas of literature is developed and refuted. Following that, the demand for intention is criticized, and a weakening of this principle is suggested. A great obstacle for a meaningful interpretation is the tokenization process, which is discussed as well. Another contribution is a sketch of how the outputs of Large Language Models could be interpreted in truth-conditional semantics. Lastly, future approaches for the philosophically informed study of Large Language Models are summarized.
Keywords
Not available
Specialisations
Similar theses
- Learning to adapt: From structured domain adaptation to efficient in-context learning for large language models Zihao Xu · Rutgers University · 2026
- On the Safety Alignment of Large Language Models Mengxuan Hu · 2026
- Documenting and Improving Prompts for Large Language Models Amandine M. Caut · Dalarna University Library · 2026
- Large Language Models for Symbolic Music Understanding Meng Yang · 2026
- Efficient algorithms for scaling large language models Zhang, Zhenyu, Ph. D. · 2026
- Solving Graph Problems with Large Language Models Angelo Zangari · University of Illinois Chicago · 2026
