Controlled Language (CL) rules specify constraints on lexicon, grammar and style with the objective of improving text translatability, comprehensibility, readability and usability. A significant body of research exists demonstrating the positive effects CL rules can have on machine translation quality (e.g. Mitamura and Nyberg 1995; Kamprath et al 1998; Bernth 1999; Nyberg et al 2003), acceptability (Roturier 2006), and post-editing effort (O’Brien 2006). Since CL rules aim to reduce complexity and ambiguity, claims have been made that they consequently improve the readability of text (e.g., Spaggiari, Beaujard and Cannesson 2003; Reuther 2003). Little work, however, has been done on the effects of CL on readability. This paper represents an attempt to investigate the relationship in an empirical manner using both qualitative and quantitative methods.