The Morphological Annotation of Reduplication-Circumfix Intersection in Indonesian

Abstract

In this paper, I report on the implementation of a morpheme-level annotation scheme for Indonesian [1], particularly the annotation of reduplications. Utilizing the NooJ program [2] and a set of novel linguistic resources, the majority of reduplications formed according to a number of distinct patterns were successfully annotated. However, one reduplication-circumfix intersection pattern could not be annotated. This is because the current NooJ morphological grammar is not designed to read the hyphen symbol, an orthographical cue that connects a root to its copy in Indonesian reduplication. Failure to read this symbol disconnects the opening and closing elements of the circumfix that surrounds its reduplication base. To overcome this problem, I introduced additional circumfix rules into the current morphological grammar without using any hyphen symbols in the rule definitions. The circumfix elements in the rules are linked as a single dependent unit using syntactic grammar whose rules NooJ allows to contain the hyphen symbol. However, this method promotes undesirable ambiguities. To overcome this side effect, I have modified the existing syntactic grammar to eliminate these ambiguities.