Related Concept Videos
Language and Cognition
Language Development
The critical period for language acquisition suggests that the ability to acquire language is at its peak early in life. As people age, this proficiency decreases. Language development begins very...
Higher Mental Functions of the Brain: Language
Language formation and comprehension take place in the dominant hemisphere. The dominant hemisphere is responsible for understanding the meaning of spoken, written, or sign language, as well as the ability to communicate. For most people, the left hemisphere is the dominant one. The right hemisphere, then, gives tone and emotional context to the...
Components of Language
Language
Corballis and Suddendorf (2007) and Tomasello and Rakoczy (2003) highlight the role of language in...
Natural and Artificial Concepts
You might also read
Related Articles
Articles linked to this work by shared authors, journal, and citation graph.
Large language models and emergence: a complex systems perspective.
Imagining and building wise machines: the centrality of AI metacognition.
Related Experiment Video
Updated: Aug 6, 2025

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
Published on: December 6, 2024
The debate over understanding in AI's large language models.
Melanie Mitchell1, David C Krakauer1
1Santa Fe Institute, Santa Fe, NM 87501.
Large language models (LLMs) are debated for their humanlike understanding. This research explores arguments for and against LLM comprehension, proposing an extended science of intelligence to analyze distinct cognitive modes.
More Related Videos
Area of Science:
- Artificial Intelligence (AI)
- Cognitive Science
- Philosophy of Mind
Background:
- A significant debate exists within the AI research community regarding the extent to which large pretrained language models (LLMs) exhibit genuine understanding.
- This debate centers on whether LLMs comprehend language in a humanlike manner, including the physical and social contexts that language represents.
Purpose of the Study:
- To survey the current arguments for and against the claim that large pretrained language models understand language and its encoded contexts.
- To identify key questions for the broader sciences of intelligence prompted by this debate.
- To propose the development of an extended science of intelligence capable of analyzing diverse modes of understanding.
Main Methods:
- Literature review and synthesis of arguments presented in the artificial intelligence research community.
- Analysis of the implications of LLM capabilities for the broader scientific study of intelligence.
- Conceptual framework development for an extended science of intelligence.
Main Results:
- The study outlines the core tenets of the debate surrounding LLM understanding, presenting both supportive and critical viewpoints.
- It highlights emergent questions concerning the nature of intelligence, understanding, and cognition in both biological and artificial systems.
- The research identifies the need for a more comprehensive framework to study different forms of understanding.
Conclusions:
- An extended science of intelligence can offer valuable insights into the distinct modes of understanding, their respective strengths, and limitations.
- This extended science will be crucial for addressing the complex challenge of integrating diverse cognitive capabilities, both human and artificial.
- The findings suggest a path forward for a more nuanced understanding of artificial intelligence and its relationship to general intelligence.

