Humans build novel tools, external knowledge structures (markers, maps etc.), and internal structures (analogies, mentalmodels etc.) to facilitate cognition. Humans also recombine these building strategies to suit any task. Other organismsgenerate such structures as well, but they use them to optimize single tasks. This suggests that the human species’ cognitiveadvantage stems from the capability to recombine built structures, and the resulting extended mind. Chandrasekharan& Stewart (2007) hypothesized that this capacity could emerge from reinforcement learning. We tested this proposal,by studying three foraging models, which examined whether novel recombinations of building (external and internalnavigation structures) emerged in reactive agents, from just reinforcement learning. Results showed that recombinationdoes not emerge with just reinforcement. This was because the building of external structures provided a very high rewardprofile, including free riding, thus acting as an attractor, blocking the recombination strategy. We discuss the implicationsof these results.

Science teaching is a complex socio-cognitive practice, where teachers simultaneously collaborate with and influence student minds. This process is distributed across textbooks, explanations, blackboard activities, student questions, student performance, etc.; and enactive, as teachers act out scientific mechanisms using descriptions, gestures, teaching props, models etc. This complex process of Teacher Cognition (TC) is not well understood, as existing studies are disparate, and based on disjointed approaches. The lack of a TC theory limits the design of systematic education policies - currently based on intuitions about teacher cognition - leading to policy guidelines that teachers find difficult to implement. Recent developments in embodied/enactive cognition theory – particularly the enactive simulation model of language and the enactive simulation theory of other minds – provide useful ways to develop models of TC, especially related to science teaching. Here we extend and develop a finer-grained version of a recent enactive model of TC, using classroom data.

Forming new science concepts, such as photosynthesis, is one of the ways students transition from everyday thinking to a more scientific worldview. Academic language (AL) is a central mediator of this transition. However, it is unclear how the features of AL (such as nominalisation and encapsulation), and classroom dynamics based on AL, promote new concept formation. We present an analytical framework to explore the way AL mediates the transition to scientific thinking, extending the simulation model of language understanding. We use this framework to develop a distributed cognition model of classroom dynamics, where the teacher first develops a standard mental simulation from textbook descriptions. She then uses this standard simulation to nudge students' individual mental simulations towards a 'gist simulation', which roughly captures the textbook model. We are currently examining how this model, along with an interactive media system to teach AL, could advance our understanding of science classroom dynamics.

Embodied cognition theories predict that changing motor control would change cognitive control, as cognition is considered to emerge from action in this theoretical approach. We tested this prediction, by examining the attention and cognitive control capabilities of a group of school students (12-13-year-olds) trained to write using both hands (experimental group, N=28), compared to a group of age- matched children (control group, N=33) who did not receive such training. The key tasks used were the attentional network test (ANT) task and the hearts and flowers (HF) task. Results from the ANT task showed that there was no significant difference in the three attentional networks between the groups. However, results from the HF task showed that the experimental group had better inhibitory control. This second result provides support to the embodied cognition prediction that cognitive control and motor control are related, and the former can be changed to some extent by changing the latter.

Behavioral and neuroanatomical studies show that reading comprehension is based on mental simulation, as modality-specific sensorimotor representations and processes related to the meaning of the text are activated during reading. Supporting this simulation view of reading comprehension, cognitive semantics analyses show that changes in lexico-grammatical features are accompanied by changes in mental simulation. We hypothesized that Academic Language (AL) texts with different linguistic structuring will alter students’ mental simulations, particularly their perspectives differently, which can be tracked through its effect on attention. We first changed the linguistic structuring of AL text – particularly nominalization – to change the encoded perspective. We then tracked the effect of this change on the global and local perspectives of student mental simulations, using the Navon task. Initial results show a trend where changes in the passage structure led to a local attention effect. We discuss the implications of this indicative result, and ongoing work to further examine this finding.