Developing the Persian WordNet of Verbs; Issues of Compound Verbs and Building the Editor Masoud Rouhizadeh NLP Research Laboratory Shahid Beheshti University Tehran, Iran mrouhizadeh@gmail.com Mahsa A. Yarmohammadi NLP Research Laboratory Shahid Beheshti University Tehran, Iran yarmohamadi@gmail.com Mehrnoush Shamsfard NLP Research Laboratory Shahid Beheshti University Tehran, Iran m-shams@sbu.ac.ir Abstract In this paper we mostly focus on the behavior of Persian compound verbs and the way we propose to deal with them. Most of the Persian verbs are compound verbs and they are formed by two major patterns of combination and in- corporation. In many cases the compound verbs are semantically transparent. This beha- vior of the verbs has some important conse- quences in the Persian semantic lexicon hence; we design an editor to fully support it. The system architecture is three-tier model and in analysis, design and implementation of this editor we used prototyping methodology. The database consists of 11 tables which are re- lated to each other by definite relations and store Persian verbs, nouns, adjectives, adverbs, prepositions and synsets. The results are com- patible to other WordNets and the information is exportable to XML. 1 Introduction Persian is the official language of three countries and it is also spoken in more than six other coun- tries. There is no doubt in the necessity of con- structing basic language processing resources and tools for it, like many other less-studied lan- guages. On the other hand, one of the most ur- gent problems in language technology is the lex- ical semantics bottleneck, the unavailability of domain-independent lexica with rich semantic information on lexical items. Such lexica could greatly improve the quality of current applica- tions. There have been some attempts for reaching this goal (Famian & Aghajaney 2006; Keyvan et al., 2006; Mansoori & Bijankhan, 2008); howev- er, most of them are only considering design of the structure and, in practice, limited sets of words or lexemes are entered in the lexicon. This paper is a report of an ongoing project of developing the Persian WordNet of verbs, per- suading our previous work (Rouhizadeh et. al. 2007 and 2008). It is a part of a larger project of building a semantic lexicon for Persian called FarsNet (Shamsfard, 2008). Here we mostly focus on the behavior of Per- sian compound verbs and the way we propose to deal with them. Then we will review the editor of the WordNet of Persian verbs which is designed to handle the compound verbs phenomena in Persian. This paper is divided into two parts; first, we give some theoretical considerations about Per- sian compound verbs and then we briefly review the editor of the Persian WordNet of verbs. 2 Compound verbs in Persian WordNet Persian verbs can be divided into two major morphological categories: simple and compound verbs. Compound verb formation is highly pro- ductive in Persian. The number of simple verbs in Persian today, is less than 200 verbs while the number of compound verbs is more than 4000. compound verb formation is highly productive in Persian today. Persian compound verbs show interesting semantic behavior and a good seman- tic lexicon of Persian should deal with such par- ticular characteristics. In the following subsec- tions we briefly review different types of com- pound verb formation in Persian and their se- mantic properties, then, we review the conse- quences of these properties in Persian WordNet. 2.1 Persian compound verbs and their se- mantics According to Dabirmoghaddam (1997) there are two major types of compound-verb formation in Persian which are Combination and Incorpora-