بررسی طبیعی‌بودن برخی پسوندهای فاعلی در تاریخ الوزرا

An investigation into certain
Persian Suffixes extracted from Tǎrix-ol-vozarǎ: natural morphology

1. Introduction
The present paper aims at comparing and contrasting certain number of some suffixes to discover which one is most natural (unmarked) and least marked within the Natural Morphology framework.The present paper is a corpus-based study of naturalness and markedness of Persian subject suffixes including '-gar, -kǎr, -mand, -ande, -sǎr, -vǎr, -ǎ, -ǎn, -ǎr, and -bǎn', extracted from Tǎrix-ol-vozarǎ, within Natural Morphology framework. Natural morphology is a cognitive-functional approach which was introduced in 1970s. In this approach, naturalness and markedness are two important terms which are different from frequency and simpleness. Naturalness and being unmarked are synonymous. Naturalness is not a definite issue; in other words, we can show naturalness of a language process in a continuum. There are 4 semiotic criteria for understanding naturalness of any word formation process including: 1. morphotactic transparency, 2. indexicality, 3. transparency bi-uniqueness 4. Shape of signan. According to morphotactic transparency, there is a relation between form and meaning: meanwhile there is a meaning and concept for each form. In the forms which have the highest transparency, we can separate stem and affix; while in some cases, affixation causes some phonological changes in stem and it result in reducing transparency. Indexicality means a signan should have just one signatum. Bi-uniqueness means that a signan should have just one signatum and vice versa. This criteria doesn’t accept synonymous, homonymy and polysemy. For example, whenever two words are synonyms, then the have violated bi-uniqueness. Also, productivity is important for this criteria. There are two sub-citeria for this parameter: Qualitative criteria and quantitative criteria. According to the first one, closeness of the meaning of a derived word to its stem, without any semantic change is focused. According to the later, new tokens of a process should be focused. The fourth criteria of naturalness, is about the number of syllables of each word or affix. One-syllable words and affixes are more natural than two-syllable ones. Data of the present paper are studied according these four criteria. As Persian subject suffixes form a fairly large category, a corpus-based study on them remains to be done.

2. Materials and Methods
In the present paper, subject suffixes are semantically studied; there are two types of meaning focused; dynamic and static. The first one is when a subject suffix about doing an action; for example, 'ahangar' is a person who makes tools using iron. This action has a process including start, continue and end. Subject suffix in this word has a dynamic meaning. An example for static meaning is in word 'tavǎnǎ'. In this word an action is not done; but it shows having a quality; then the meaning of suffix in such words is static. This categorization helps us to put some words in the category of subject which were not observed as subject, before. These terms (static and dynamic) are used in systemic functional grammar, about processes.
The writers of this paper make the data machine readable; so the book Tǎrix-ol-vozarǎ was typed and after that ant- conc software was used to extract subject suffixes; extracting suffixes, they were copied into Excel to filter the parts which were not needed. It was needed to study all the words derived using target suffixes, even the words in which target suffixes weren’t subjective ones; In other words, there were some cases in which target suffixes didn’t imply subjective meaning; In fact some implied some other meanings including gerund, object and so on. But in some other cases, target suffixes didn’t have the role of an independent affix. For instance, in some cases, '-kǎr' was an independent word. In this case, writers had to filter the extra data.

2. Results and Discussion
After filtering unnecessary data, all the sentences containing target derived words, were studied and these results were observed: the number of derived words containing suffix '–bǎn' is the most- 352 cases and suffix '–kǎr' has the least amount of derived words. Based on the findings of the study, '-gar' is the most while '–sǎr' is the least natural suffix in Tǎrix-ol-vozarǎ.. In other words, it is shown that suffix '-gar' has the least and suffix '-sǎr' has the most violation of four criteria mentioned above. Order of other suffixes in the ranking of naturalness is like this: ', -kǎr, -mand, -ande, -bǎn, -ǎ, -vǎr,and -ǎn'. As it is shown, after '-gar', '-kǎr' has the least violation of four main criteria. It shows that it has the second order of naturalness among subject suffixes in Tǎrix-ol-vozarǎ. It was also proved that '–ǎr' is not a subject suffix in Tǎrix-ol-vozarǎ and it is used as a gerund or object suffix.

3. Conclusion
Findings show that the most natural subject suffix in Tǎrix-ol-vozarǎ is '-gar' ; because it conform more to the four main criteria mentioned in natural morphology; then it could be said that it is the most natural subject suffix in Tǎrix-ol-vozarǎ.