Abstract This article presents a case of (non-canonical) optional subject marking observed exclusively in spontaneous speech. Persian is a pro-drop language with obligatory subject agreement marked by verbal suffixes and a set of clitic pronouns serving different functions, e.g., pronominal objects encliticized to the verb. In standard Persian, subject agreement for the 3rd person singular (3SG) is consistently realized by zero in past tenses; however, as shown in this study, in spoken language, speakers occasionally use the 3SG pronominal clitic = eš instead of zero. This non-standard usage is generally neglected, if not ignored, in previous research. Drawing on evidence from historical studies and naturally produced data from spoken and transcribed Persian, this study highlights the empirical shortcomings of previous claims that the subject agreement marker - eš results from grammaticalization of the object clitic = eš as a means of repairing or levelling the verbal agreement paradigm. Instead, it is argued that while written Persian preserves more parsimonious constructions, the redundant use of - eš to index the subject, similar to object clitic doubling, is part of the indexing system of spoken Persian and can be understood in terms of fluid differential subject marking. Differences in indexing between the written (i.e., offline language production) and spoken (i.e., spontaneous language production) modalities are not unusual and may be caused by different factors. Studying these differences in transcribed spoken data, this study identifies distinct patterns of information packaging in the two modalities and offers an explanation in terms of cognitive principles of language processing and production to account for the difference in indexing between spoken and written Persian.
Pegah Faghiri (Thu,) studied this question.