Explicit position cues describe conceptual structure
If one buys into the idea that concept learning is few-shot scaffolding, one might also argue that the mental models superimposed on new and previous knowledge are also essential to the process. Without additional structural guidance (e.g. by using plain fluid search), the learner has to derive the relevant conceptual structure by themselves. This additional step could potentially be augmented as well, through explicit position cues for placing concepts in semantic space. For instance, parallels might be called forth based on reliably internalized knowledge in an effort to better accomodate new concepts (e.g. when learning what a neuron is, a system could summon an analogy along the lines of “the forest is to the tree as the brain is to the neuron”). Essentially this would lead to a constructive pattern matching tool for mental models. As many other tools for thought, it might have unintended consequences.