Artificial Symbol Learning With Training - Experiment 2 materials

Materials for Experiment 2 in the paper "Learning Artificial Number Symbols with Ordinal and Magnitude Information".
Article abstractThe question of how numerical symbols gain semantic meaning is a key focus of mathematical cognition research. Some have suggested that symbols gain meaning from magnitude information, by being mapped onto the approximate number system, whereas others have suggested symbols gain meaning from their ordinal relations to other symbols. Here we used an artificial symbol learning paradigm to investigate the effects of magnitude and ordinal information on number symbol learning. Across two experiments, we found that after either magnitude or ordinal training, adults successfully learned novel symbols and were able to infer their ordinal and magnitude meanings. Furthermore, adults were able to make relatively accurate judgements about, and map between, the novel symbols and non-symbolic quantities (dot arrays). Although both ordinal and magnitude training was sufficient to attach meaning to the symbols, we found beneficial effects on the ability to learn and make numerical judgements about novel symbols when combining small amounts of magnitude information for a symbol subset with ordinal information about the whole set. These results suggest that a combination of magnitude and ordinal information is a plausible account of the symbol learning process.
© The Authors

论文《结合序数与幅值信息学习人工数字符号》（Learning Artificial Number Symbols with Ordinal and Magnitude Information）中的实验2所用材料。

论文摘要
数字符号如何获得语义内涵，是数学认知研究的核心议题。部分研究者提出，符号可通过映射至近似数量系统（approximate number system）从幅值（magnitude）信息中获取语义；另有研究者则认为，符号的语义源于其与其他符号之间的序数（ordinal）关系。本研究采用人工符号学习范式，探究幅值与序数信息对数字符号学习的影响。在两项实验中，我们发现：无论接受幅值训练还是序数训练，成年受试者均能成功习得全新符号，并推导其序数与幅值语义。此外，成年受试者可对全新符号与非符号数量（点阵列（dot arrays））做出较为精准的判断，并实现二者间的映射。尽管序数训练与幅值训练均足以赋予符号语义，但我们发现，若将符号子集的少量幅值信息与全体符号的序数信息相结合，受试者学习全新符号并完成数值判断的能力将获得增益。上述结果表明，幅值与序数信息的结合可合理解释符号学习的过程。

© 作者