The present work investigates whether and how morphological decomposition processes bias the reading of Hebrew heterophonic homographs, i.e., unique orthographic patterns that are associated with two separate phonological, semantic entities depicted by means of two morphological structures (linear and nonlinear). In order to reveal the nature of morphological processes involved in the reading of Hebrew homographs, we tested 146 university students with three computerized experiments, each experiment focusing on a different level of processing. Participants were divided into three experimental groups given that the three experiments used the same stimulus lists. Evidence obtained from the analysis of the participants' processing time and processing accuracy points to a propensity to process heterophonic homographs by default as morpho-syntactically simple rather than complex words. Findings are discussed with reference to assumptions made by Dual-Route models regarding the importance of morphological knowledge in fast and accurate access of written words' representations which mediate the retrieval of their meanings with direct reference to the context in which they occur.