Syntactic Frameworks and Discourse Functions of Lexical Antonymy in Persian Language

Antonymy is a unique semantic relation between two lexical forms that are opposite while at the same time they share some basic similarities. Antonyms have attracted attention of linguists interested in lexical semantics. A number of linguists such as Lyons (1977), Cruse (1986) and Murphey (2003) have examined lexical and semantic characteristics of antonyms. Some other scholars have paid attention to discourse functions of antonymy among whom Jones (2002) have provided the most clear-cut theoretical framework. Jones used a test set of 56 word pairs that were well-known, conventional antonyms. They were not balanced across word class, morphological complexity, word length or frequency ranking, but were selected to be representative of the antonym relation. He extracted all instances of these antonyms co-occurring in sentences from a British newspaper corpus of 280 million words. Jones limited the analysis to a sample of 3000 sentences. Approximately every 30^th sentence was extracted from the corpus. He then adjusted it so that no more than 60% of the sentences involved adjectival antonymous pairs, in order to ensure that there were sufficient noun, verb, and adverb pairs within the sample. Next, the discourse function of each antonymous pair was identified. One of the categories introduced in Jones (2002) was ancillary function in which an antonym pair is used to create or highlight a secondary contrast within sentence/discourse. The second major antonym function is coordinated antonymy, in which the distinction between the two opposite is neutralized. Comparative antonymy involves measuring one antonym against the other. The distinguished function calls attention to the inherent distinction between the members of the antonym pair. Transitional antonymy expresses a movement or change from one location, activity or state to another. The negated antonymy function emphasizes one member of the antonym pair by using it with the negation of the other member. Jones's last category is idiomatic category, in which any instance of antonym co-occurrence would be recognized as a familiar idiom.  
   The previous studies conducted on some languages including Persian revealed that in spite of some differences across languages, almost all of them follow quite the same patterns in implementing different discourse functions. The present study enjoys theoretical as well as methodological concepts from syntax, semantics and discourse analysis to recognize Persian antonyms and identify their discourse function and then the relevant syntactic framework. Here, we attempt to answer the following four questions;
1. What is the share of each discourse function in Persian antonyms?
2. What is the relationship between grammatical framework and discourse function of antonyms?
3. What are the most frequent antonyms in Persian (non)virtual texts?
4. What is the semantic nature of antonyms as far as the kind of genre is concerned?
To shed light on the questions, a corpus of 4000 Persian sentences which contained antonyms was selected. They were extracted from written, verbal and virtual genres. The criterion was that each antonym whether noun, verb, adjective, adverb or pronoun had to occur in a single sentence. Next, the part of speech, discourse function and grammatical framework of each antonym were identified.
     The findings indicate that in Persian the coordinated discourse function is the most frequent function and other functions, i.e. negated, transitional, idioms, ancillary, and comparative follow that. It is also revealed that there is a significant relationship between some syntactic frameworks and discourse functions. For example, the syntactic framework of "X and Y" is significantly frequent in coordinated discourse function. Likewise, it is shown that there exists a relationship between the semantic nature of some antonymous pairs and their frequency in some genres. For instance, the antonymous pair 'real-virtual' is used just in virtual corpus. These kinds of findings help corpus linguistics researchers predict the type of corpus with an eye to some antonymous pairs.
     To be specific, it should be mentioned that out of 4000 antonyms, 1432 antonyms are nouns, 926 items are verbs, 800 ones are adjectives, 501 cases are adverbs, 89 antonyms are pronouns and finally, 252 antonyms form idiomatic expressions. As far as the discourse function is concerned, 1843 antonyms are coordinated, 533 items are labelled 'extra', 313 cases transitional, 253 antonyms are idiomatic, 226 items are ancillary and 87 ones are comparatives. Hence, the coordinated discourse function ranks the highest one and the negated, transitional, idiomatic, ancillary and comparatives are followed respectively.  It is also shown that there is some relationship between specific discourse functions and grammatical frameworks. For example, the syntactic framework X and Y is frequent in coordinated discourse function. Likewise, it is revealed that there is a connection between some antonyms and the genre they appear. As one example, we may refer to the antonyms 'vaaqe'i (real) against 'majaazi' (virtual) which appear just in Persian virtual genre. Studies like this would help corpus linguists predict the genre of texts based on the existing antonyms.