Springe direkt zu Inhalt

Dobler, F., Henningsen-Schomers, M. R., Pulvermüller, F. (2024): Verbal Symbols Support Concrete but Enable Abstract Concept Formation: Evidence From Brain‐Constrained Deep Neural Networks

News vom 19.05.2024

Dobler, F., Henningsen-Schomers, M. R., Pulvermüller, F. (2024). Verbal Symbols Support Concrete but Enable Abstract Concept Formation: Evidence From Brain‐Constrained Deep Neural NetworksLanguage Learning, 2024, 1-38. doi: 10.1111/lang.12646

Highlights
  • Abstract Concepts Rely on Language: Abstract concepts (e.g., “justice,” “democracy”) form stable neural representations only when associated with verbal symbols, indicating that language is essential for their formation.
  • Concrete Concepts Form Without Language: Concrete concepts (e.g., “sun,” “run”) develop robust and sustained neural representations based solely on sensory and motor experiences, even without verbal input.
  • Verbal Symbols Enhance All Concepts: Associating words with both concrete and abstract concepts strengthens their neural representations, but it’s crucial for abstract concepts to achieve stable activation.
  • Supports Linguistic Relativity: The findings back the strong version of the linguistic relativity hypothesis, suggesting language is necessary for forming and stabilizing abstract concepts.
  • Mechanistic Insight into Concept Formation: Brain-like neural networks use language to bind disparate features, enabling the formation of abstract concepts through verbal symbols.
Abstract

Concrete symbols (e.g., sun, run) can be learned in the context of objects and actions, thereby grounding their meaning in the world. However, it is controversial whether a comparable avenue to semantic learning exists for abstract symbols (e.g., democracy). When we simulated the putative brain mechanisms of conceptual/semantic grounding using brain-constrained deep neural networks, the learning of instances of concrete concepts outside of language contexts led to robust neural circuits generating substantial and prolonged activations. In contrast, the learning of instances of abstract concepts yielded much reduced and only short-lived activity. Crucially, when conceptual instances were learned in the context of wordforms, circuit activations became robust and long-lasting for both concrete and abstract meanings. These results indicate that, although the neural correlates of concrete conceptual representations can be built from grounding experiences alone, abstract concept formation at the neurobiological level is enabled by and requires the correlated presence of linguistic forms.

3 / 46