Document Type : Research article
Authors
1 Assistant Professor, Department of English Language, Faculty of Literature and Humanities, Lorestan University, Khorramabad, Iran.
2 Assistant Professor, Department of Persian Language and Literature, Faculyt of Language and Literature, University of Kurdistan, Sanandaj, Iran.
Abstract
Highlights
Extended Abstract
Introduction
The ezafe construction is one of the most central syntactic devices in Persian. In its standard form, it is marked by the unstressed vowel -e/-ye, which links a head noun to its modifier, possessor, complement, or descriptive dependent. Although this construction is usually represented in writing through kasra, historical and dialectal evidence shows that Persian and related Iranian varieties have not always expressed this relation in a uniform way. In some Iranian dialects, and occasionally in Persian textual transmission, the relation normally marked by ezafe appears with vāv instead of kasra. This phenomenon, which may be called “ezafe vāv,” creates a significant philological and grammatical problem, because its surface form is identical to the ordinary coordinating conjunction vāv ‘and’. As a result, scribes, editors, and commentators may misinterpret an ezafe structure as a coordinate construction, or conversely emend a genuinely coordinate phrase into an ezafe construction. This study proposes a corpus-based hypothesis about a group of such constructions in the Shahnameh and in other early Persian texts. The main concern is not merely orthographic variation, but the grammatical and editorial consequences of a structure in which vāv functions as a marker of dependency rather than coordination. The problem is especially important in classical Persian textual criticism, where editors often decide between competing manuscript readings on the basis of semantic plausibility, metrical constraints, and the authority of particular manuscripts. If vāv can, under certain conditions, represent ezafe, then several readings traditionally treated as corrupt, secondary, or coordinative may need to be reconsidered. The study therefore aims to distinguish between genuine coordinate structures and ezafe constructions containing vāv, and to clarify the grammatical environments in which such a substitution is possible.
Research Methodology
The research adopts a corpus-linguistic approach to the Shahnameh, using the complete electronic text of Ferdowsi’s epic as the main corpus. The corpus consists of 49,609 couplets, approximately 569,534 tokens, and 17,782 word types. In order to facilitate processing, the corpus was divided into smaller subcorpora, each containing roughly 50,000 words. This division made it possible to search the text more efficiently and to examine recurring syntactic patterns in manageable segments.
The analysis focused on structures of the type “noun + vāv + noun,” because this is the main environment in which ambiguity between coordination and ezafe arises. Using AntConc, all instances of vāv were first extracted from the corpus. For each occurrence, the immediate left and right contexts were examined, especially one word before and one word after vāv. Since the corpus was not morphologically tagged, grammatical classification was carried out manually. Cases in which both the preceding and following words were nouns were separated from other patterns. This produced a large set of potential “noun + vāv + noun” constructions, from which approximately 14,079 relevant instances were identified.
From this larger set, seventy examples were selected randomly for closer analysis. These examples were examined not only in the edited text of the Shahnameh, but also in major manuscripts and printed editions where available. The purpose was to determine whether particular occurrences of vāv should be interpreted as ordinary coordination or as a non-standard marker of ezafe. Four major examples received detailed attention: pārsi o dari “Persian and Dari / Persian-Dari,” bām o sarāy “roof and house / roof of the house,” zabān o gomān “tongue and thought / tongue of thought,” and pilān o kūs “elephants and drums / drums of the elephants.” The first two examples also occur in other early Persian texts, while the latter two were examined primarily within the Shahnameh. The study also considered problematic cases such as māh o mehr, where an originally coordinate structure may have been incorrectly understood as ezafe.
Discussion
The analysis shows that vāv cannot be treated simply as a graphic or scribal error in all cases. In several recurrent structures, the semantic relation between the two nouns is not coordinative but possessive, dependent, or relational. For example, in bām o sarāy, the broader poetic context suggests not “the roof and the house” as two separate entities, but “the roof of the house.” Similar reasoning applies to zabān o gomān, where the expression is difficult to interpret as “tongue and thought” in a simple coordinate sense, but becomes meaningful if understood as “the tongue of thought,” that is, the expressive or verbal capacity of the mind. In the case of pilān o kūs, the manuscript variation between pilān o kūs and pilān-e kūs suggests a deeper instability in the transmission of a phrase referring to drums mounted on or associated with elephants. The occurrence of both forms across manuscripts indicates that scribes may have alternated between a perceived coordinate reading and an ezafe reading.
The study further argues that the phenomenon is not unrestricted. Not every “noun + vāv + noun” structure can be interpreted as ezafe. The most important grammatical finding is that ezafe vāv appears mainly in possessive or relational ezafe constructions, where the second noun can be understood as having, possessing, or being associated with the first. In such cases, the phrase may be paraphrased by a relation of possession or belonging: for instance, “the arm has strength” in zur o bāzu “strength of the arm,” or “thought has a tongue” in zabān o gomān “tongue of thought.” This semantic test helps distinguish ezafe vāv from ordinary coordination.
By contrast, explanatory or appositional ezafe structures do not normally permit this substitution when they take the form “noun + noun.” Thus, a phrase such as māh o mehr should not automatically be emended to māh-e mehr, because the semantic relation does not support the same possessive dependency. Such cases are better interpreted as genuine coordination unless there is strong contextual or manuscript evidence to the contrary. The study therefore challenges overly regularizing editorial practices that replace vāv with kasra merely to conform to standard Persian syntax.
The example pārsi o dari is particularly significant because it has often been understood as referring to two linguistic varieties, “Persian and Dari.” However, the analysis suggests that in some contexts it may represent pārsi-ye dari, that is, “Dari Persian,” especially where the text clearly refers to a single language rather than two separate varieties. This interpretation has implications for the historical understanding of linguistic terminology in early Persian sources.
Conclusion
This study demonstrates that some ambiguous vāv-marked structures in the Shahnameh and other early Persian texts are best understood as instances of ezafe vāv rather than ordinary coordination. The phenomenon appears to have entered manuscript transmission through dialectal or scribal habits and later became entangled with standard coordinate constructions. A corpus-based method makes it possible to move beyond isolated examples and identify recurrent syntactic and semantic patterns.
The main grammatical conclusion is that vāv may replace the ezafe marker primarily in possessive or relational ezafe constructions, especially in “noun + noun” structures where a dependency relation can be semantically recovered. In other structural types, particularly explanatory or appositional constructions, such substitution is much more restricted and may lead to incorrect editorial decisions. The findings therefore contribute both to Persian historical grammar and to the methodology of editing classical Persian texts. They show that manuscript readings containing vāv should not be dismissed automatically as corrupt or merely coordinative. Instead, each case must be evaluated through the combined evidence of corpus frequency, manuscript distribution, syntactic structure, semantic relation, and textual context.
Keywords
Main Subjects
)