The observation of an action modulates motor cortical outputs in specific ways, in part through mediation of the mirror neuron system. Sometimes we infer a meaning to an observed action based on integration of the actual percept with memories. Here, we conducted a series of experiments in healthy adults to investigate whether such inferred meanings can also modulate motor cortical outputs in specific ways. We show that brief observation of a neutral stimulus mimicking a hand does not significantly modulate motor cortical excitability (Study 1) although, after prolonged exposure, it can lead to a relatively nonspecific modulation (Study 2). However, when such a neutral stimulus is preceded by exposure to a hand stimulus, the latter appears to serve as a prime, perhaps enabling meaning to the neutral stimulus, which then modulates motor cortical excitability in accordance with mirror neuron-driving properties (Studies 2 and 3). Overall results suggest that a symbolic value ascribed to an otherwise neutral stimulus can modulate motor cortical outputs, revealing the influence of top-down inputs on the mirror neuron system. These findings indicate a novel aspect of the human mirror neuron system: an otherwise neutral stimulus can acquire specific mirror neuron-driving properties in the absence of a direct association between motor practice and perception. This significant malleability in the way that the mirror neuron system can code otherwise meaningless (i.e. arbitrarily associated) stimuli may contribute to coding communicative signals such as language. This may represent a mirror neuron system feature that is unique to humans.