Syntactic chunking reveals a core syntactic representation of multi-digit numbers, which is generative and automatic

Area of Science:

• Cognitive Science
• Psycholinguistics
• Neuroscience

Background:

• Understanding the cognitive mechanisms of numerical representation is key to explaining human numerical literacy.
• The precise way literate adults process and store number-word sequences remains an open question.

Purpose of the Study:

• To investigate whether literate adults represent the complete syntactic structure of numbers.
• To determine how this syntactic representation influences short-term memory for number sequences.

Main Methods:

• Five experiments involving participants repeating number-word sequences.
• Systematic variation of word order to test grammatical versus non-grammatical sequences.
• Analysis of repetition accuracy and chunking behavior in relation to sequence length and structure.

Main Results:

• Repetition accuracy was significantly higher for grammatical number sequences compared to non-grammatical ones.
• Memorization improved with longer grammatical segments up to approximately four words, after which performance declined.
• Chunking of number words occurred automatically, even when it impaired performance or external cues were absent.

Conclusions:

• Literate adults actively represent and utilize the full syntactic structure of numbers.
• This representation facilitates automatic, hierarchical chunking in short-term memory, with an optimal segment size.
• The findings provide a detailed account of number syntactic structure representation, essential for numerical literacy.