Persian Proposition Bank

Persian Proposition Bank Azadeh Mirzaei*, Amirsaeid Moloodi** *Department of Linguistics, Allameh Tabataba'i University, Tehran, Iran **Department of Foreign Languages and Linguistics, Shiraz University, Shiraz, Iran E-mail: azadeh.mirzaei@atu.ac.ir, amirsaeid.moloodi@shirazu.ac.ir Abstract This paper describes the procedure of semantic role labeling and the development of the first manually annotated Persian Proposition Bank (PerPB) which added a layer of predicate-argument information to the syntactic structures of Persian Dependency Treebank (known as PerDT). Through the process of annotating, the annotators could see the syntactic information of all the sentences and so they annotated 29982 sentences with more than 9200 unique verbs. In the annotation procedure, the direct syntactic dependents of the verbs were the first candidates for being annotated. So we did not annotate the other indirect dependents unless their phrasal heads were propositional and had their own arguments or adjuncts. Hence besides the semantic role labeling of verbs, the argument structure of 1300 unique propositional nouns and 300 unique propositional adjectives were annotated in the sentences, too. The accuracy of annotation process was measured by double annotation of the data at two separate stages and finally the data was prepared in the CoNLL dependency format. Keywords: semantic role labeling, Persian Proposition Bank, predicate-argument information, propositional noun, propositional adjective, CoNLL dependency format. 1. Introduction As we know, the amount of information obtained from the syntactic analyses of sentences is limited and one needs to go to a higher level which is semantic analyses to achieve the propositional content of the sentence. The following examples elaborate on the necessity of semantic role labeling. 1. Ali broke the window. OBJ 2. The window broke. SBJ In the above sentences, the active verb "to break" has two different syntactic realizations. In the first sentence "the window" is the verb’s direct object and its subject in the second one. Although "the window" has two different syntactic roles in these sentences, it plays the same underlying semantic role (PATIENT) in both. Conversely, in some cases the same syntactic realizations may result in different meanings. 3.     Ali to question gets back ‘Ali gets back to the question.’ ARG0:  (Ali) ARG1:  (to the question) REL:  (gets back) 4.     issue to planning gets back ‘The issue gets back to planning.’ ARG1:  (issue) ARG2:  (to planning) REL:  (gets back) As it is obvious from the examples (3) and (4), the syntactic realizations of the both sentences are the same (||subject, prepositional object (to)||) but the assigned numbered arguments are different and so there are semantically two different verbs. As a result the semantic analysis of the sentences is essential to achieve propositional content. The first Persian Proposition Bank like the other Proposition Banks (Kingsbury & Palmer, 2002; Palmer et al., 2005; Burchardt et al., 2006; Xue & Palmer, 2009; Palmer et al., 2008; Ohara et al., 2004; Taulé et al., 2008) aims to construct the argument structure of the predicates and in order to analyze the sentences semantically, the functional …

Content truncated to 500 words