During the processing of a compound word, its constituents become available and are easier to recognize due to the morphological segmentation of the compound. We investigated themorphological segmentation of compounds across three experiments. Specifically, we investigated whether morphological segmentation of compound masked primes is always attempted, whether it can occur based on orthographic information that is available prior to accessing the lexical representation of the compound word, and whether the early stages of morphological segmentation are influenced by semantic information. In Experiment 1, we found that morphological segmentation of compound masked primes occurred automatically and that the early stages of this process were influenced by semantic information. In Experiment 2, we found that morphological segmentation was based on orthographic information that was available prior to access of the lexical representation of the whole compound word. Finally, in Experiment 3, we found that orthographic and semantic information influenced morphological segmentation separately. Our findings suggest that morphemes become available early in compound processing and are extracted automatically to help process the whole compound word.

.