Thoughts from high-level processes often arise involuntarily and in response to external stimuli. The reflexive imagery task (RIT) has been used to identify the kinds of conscious thoughts and imagery that can arise involuntarily in response to stimuli. In most RITs, subjects are presented with stimuli and instructed to not perform a mental operation in response to the stimuli. For example, subjects are presented with 2 dots after having been instructed to not count any objects. Often, subjects cannot suppress the operation and experience the involuntary imagery (e.g., “two”). RITs have not examined whether such effects arise for insights or syntactic processing. In our “insight” RIT, subjects were presented with 20 word pairs (e.g., wax–flame) and instructed to not think of any associates (e.g., “candle”). Subjects thought (involuntarily) of associates for 68% of the pairs. In our “syntax” RIT, subjects learned to associate nonsense visual symbols with words (e.g., “green” and “sun”). After training, in the read condition, subjects read the strings formed from pairs of these symbols (e.g., “green sun”). In the suppress condition, subjects were instructed to suppress reading. Subjects reported involuntary reading of these strings on 48% of the 20 trials. The mean accuracy rate of the judgments regarding nonsensicality (e.g., green sun) in the suppress condition was high (86%), corroborating that subjects were reading the sentences. Together, the data from our new tasks shed light on the capabilities of the mechanisms that, in everyday life, engender the often “high-level” contents that occupy our conscious minds.