This thesis investigates the morphology-phonology interface in Malay. The work is largely a corpus-based reanalysis of prefixation and reduplication. Based on two large different written corpora of Standard Malay (henceforth SM), the analysis permits us to make reliable and robust generalizations about how the language actually works. The data reveal that the language has a distinct co-existing phonological system. I will show that these co-existent grammars can be handled with Optimality Theory (henceforth OT), specifically in co-phonologies. The reanalysis of prefixation places Malay in a wider context and examines, cross-linguistically, issues related to voicing and nasality. It is shown that nasal substitution, which is regularly used to eliminate nasal and voiceless obstruent clusters, fails to occur in some prefixed words. In the analysis, I propose that nonnative words are not subject to the same phonological requirements as those imposed on native words. The constraint rankings must therefore be different from those found in native words which result in the blocking of nasal substitution at prefix-root junctures. The application of nasal substitution at prefix-prefix junctures is mainly determined by morphological factors rather than phonetic factors, due to a morphology-phonology interface constraint, i. e. EDGE-INTEGRITY. The investigation of reduplication deals with total, partial and affixal reduplication. A recent theory of prosodic morphology - namely the Morpheme-Based Template or MBT (Downing 2006), motivated within OT (Prince & Smolensky 1993) - is applied to organize the morphological and prosodic factors that condition the size of prosodic morphemes (ibid. : 1). In the analysis, I propose that total and affixal reduplication are best treated as compounding, rather than affixation, due to the disyllabic minimality condition. Considering the Perak dialect, light and heavy reduplication have been captured by associating each morphological construction with a different co-phonology. This study also examines dialectal variation, comparing SM with three nonstandard dialects with respect to prefixation. The analysis discovers some significant facts about the language. Since both voiceless/voiced obstruents undergo nasal substitution, the *NC constraint has been replaced by a CRISP-EDGE[sigma] constraint. Nasal deletion and nasalisation are also the strategies used to eliminate nasal and voiceless obstruent clusters. The different strategies applied can be satisfactorily explained in OT with its variable constraint rankings.