Because of its ubiquitous and predominantly overt nature, grammarians have never been able to disregard the phenomenon of case. While the traditional grammar coined dozens of names for different forms of case, due to treating case simultaneously along syntactic, semantic and morphological lines, resulting in a confused tangle, the generative enterprise, especially the minimalist program, introduced a certain amount of clarity by separating these notions. However, Pashto grammarians remained oblivious of these phenomenal achievements in the area of case. All that we have are a few traditional accounts of case in Pashto. To fill this gap, this thesis takes the minimalist idea of agreement in terms of features between a functional head and a nominal (Schütze, 1997; Chomsky, 2001) as its starting point. For Pashto language, it is proposed that a) ϕ-features agreement between the functional category T and a nominal results in assigning nominative Case to that nominal, b) ϕ-features agreement between υ or Voice and a nominal results in assigning accusative Case to that nominal, c) ϕ-features agreement between the functional head Appl and a nominal results in assigning dative Case to that nominal, and d) agreement in terms of [N] feature between the functional head D and a possessor results in assigning genitive Case to that nominal. Predominantly, Pashto nominals show nominative-accusative pattern in the present and future tenses, and accusative-nominative (ergative-absolutive) pattern in the past tense. The verb agrees with the subject in the present and future tenses, and with the object in the past tense. This thesis proposes that this has to do with the close relation that exists between the past tense and the passive voice in Pashto, a well-established fact already reported for other Indo-Iranian languages. To deal with this split-ergative nature of Pashto, we xii propose that υ (not maintaining Chomskian distinction between υ* and υ) in Pashto past tense is defective (Chomsky, 2001), lacking [uϕ] features, hence unable to assign accusative Case. Following Collins (2005), we further propose (for Pashto) that [uϕ] in the past tense and the passive are withheld by the functional head Voice to itself. As such, whereas in the present and future tenses, υ is responsible for accusative Case assignment, in Pashto past tense and passive voice constructions, Voice assigns the accusative Case. We studied structural case assignment in various Pashto constructions; such as, monotransitives, ditransitives (datives), ao (and) conjoined subject constructions, unaccusatives, unergatives, copular, passives, determiner phrases, clitics, conjoined constructions, relative clauses, and the constructions where either one or both of the arguments are clauses with the overall conclusion that agreement (in terms of features) between a functional head and a nominal results in assigning structural Case to that nominal. This thesis, being the first of its kind, proposes argument structures/derivations for all of the above mentioned Pashto constructions, thus laying a solid foundation for future research on Pashto clause structure, light verb constructions, the nature of determiner phrases, Pashto verb base form, and the difference in continuous and indefinite aspects.