There is increasing recognition of the multimodal representational nature of science discovery practices and the roles of multiple and multimodal representations in students' meaning making in science (Lemke, 1998; Tytler, Prain, Hubber, & Waldrip, 2013; Tang, "International Journal of Science Education," 38(13), 2069-2095, 2016).…

Ambridge calls for exemplar-based accounts of language acquisition. Do modern neural networks such as transformers or word2vec -- which have been extremely successful in modern natural language processing (NLP) applications -- count? Although these models often have ample parametric complexity to store exemplars from their training data, they also…

The Cattell-Horn-Carroll (CHC) or three-stratum model of intelligence envisions human intelligence as a hierarchy. General intelligence (g) is situated at the top, under which are a group of broad intelligences such as verbal, visuospatial processing, and quantitative knowledge that pertain to more specific areas of reasoning. Some broad…

This commentary critiques Ambridge's radical exemplar model of language acquisition using research from the Longitudinal Study of Early Language, which has tracked the language development of 30+ children with Autism Spectrum Disorders (ASD) since 2002. This research has demonstrated that the children's capacity for abstraction at the grammatical…

Ambridge reviews and augments an impressive body of research demonstrating both the advantages and the necessity of an exemplar-based model of knowledge of one's language. He cites three computational models that have been applied successfully to issues of phonology and morphology. Focusing on Ambridge's discussion of sentence-level constructions,…

A theoretical model describing young students' (Grades 1-3) functional-thinking modes was formulated and validated empirically (n = 345), hypothesizing that young students' functional-thinking modes consist of recursive patterning, covariational thinking, correspondence-particular, and correspondence-general factors. Data analysis suggested that…

Elementary standards include multiplication of single-digit numbers and students advance to solve complex problems and demonstrate procedural fluency in algorithms. The ability to illustrate procedural fluency in algorithms is dependent on the development of understanding and reasoning in multiplication. Development of multiplicative reasoning…

A key property of relational representations is their "generativity": From partial descriptions of relations between entities, additional inferences can be drawn about other entities. A major theoretical challenge is to demonstrate how the capacity to make generative inferences could arise as a result of learning relations from…

As a programmer when solving a problem, a number of conscious and unconscious cognitive operations are being performed. Problem-solving is a gradual and cyclic activity; as the mind is adjusting the problem to its schemas formed by its previous experiences, the programmer gets closer and closer to understanding and defining the problem. The…

Learning progressions (LPs) are novel models for the development of assessments in science education, that often use a scale to categorize students' levels of reasoning. Pictorial representations are important in chemistry teaching and learning, and also in LPs, but the differences between pictorial and verbal items in chemistry LPs is unclear. In…

The visual-manual modality of sign languages renders them a unique test case for language acquisition and processing theories. In this commentary the authors describe evidence from signed languages, and ask whether it is consistent with Ambridge's proposal. The evidence includes recent research on collocations in American Sign Language that reveal…

Abstraction is a core principle of Distributional Semantic Models (DSMs) that learn semantic representations for words by applying dimensional reduction to statistical redundancies in language. Although the posited learning mechanisms vary widely, virtually all DSMs are prototype models in that they create a single abstract representation of a…

In 'Against Stored Abstractions,' Ambridge uses neural and computational evidence to make his case against abstract representations. He argues that storing only exemplars is more parsimonious -- why bother with abstraction when exemplar models with on-the-fly calculation can do everything abstracting models can and more -- and implies that his…

Model-based instruction offers numerous benefits to students, including increased content knowledge and critical thinking. This study explored the differences in the knowledge outcomes and reasoning processes employed by undergraduate students in an introductory biology lab as they constructed, revised, and simulated a computational model of a…

When asked to determine whether a syllogistic argument is deductively valid, people are influenced by their prior beliefs about the believability of the conclusion. Recently, two competing explanations for this belief bias effect have been proposed, each based on signal detection theory (SDT). Under a response bias explanation, people set more…