Morphologically derived words can neither be identified by dictionary look-up nor be accounted for with a syntactic parser in NLP. Mandarin Chinese involves several productive morphological rules. This paper proposes a set of rules to identify reduplicatives in Mandarin Chinese. This set of rules will be used to complement dictionary look-up and DM generation rules in the word segmentation module. The co-occurrence restriction of adjuncts in reduplication is also discussed and expressed in ICG mechanism to improve parsing results. |