Proceedings of the Annual Meeting of the Cognitive Science Society

Abstract mental representation is fundamental for humancognition. Forming such representations in time, especiallyfrom dynamic and noisy perceptual input, is a challenge forany processing modality, but perhaps none so acutely as forlanguage processing. We show that LISA (Hummel &Holyaok, 1997) and DORA (Doumas, Hummel, & Sandhofer,2008), models built to process and to learn structured (i.e.,symbolic) representations of conceptual properties andrelations from unstructured inputs, show oscillatory activationduring processing that is highly similar to the cortical activityelicited by the linguistic stimuli from Ding et al. (2016). Weargue, as Ding et al. (2016), that this activation reflectsformation of hierarchical linguistic representation, andfurthermore, that the kind of computational mechanisms inLISA/DORA (e.g., temporal binding by systematicasynchrony of firing) may underlie formation of abstractlinguistic representations in the human brain. It may be thisrepurposing that allowed for the generation or emergence ofhierarchical linguistic structure, and therefore, humanlanguage, from extant cognitive and neural systems. Weconclude that models of thinking and reasoning and models oflanguage processing must be integrated—not only forincreased plausiblity, but in order to advance both fieldstowards a larger integrative model of human cognition.