Mutations in the Aristaless-related homeobox (ARX) gene are found in a spectrum of epilepsy and X-linked intellectual disability disorders in children. During development Arx is expressed in pallial ventricular zone (VZ) progenitor cells which give rise to the excitatory projection neurons of the cortex. Arx-/Y mice were shown to have decreased proliferation in the cortical VZ resulting in smaller brains; however, the basis for this reduced proliferation was not established. To determine the role of ARX on cell cycle dynamics in cortical progenitor cells, we generated cerebral cortex specific Arx mouse mutants (cKO). The loss of pallial Arx resulted in the reduction of cortical progenitor cells, particularly affected was the proliferation of intermediate progenitor cells (IPCs). The mechanism was determined to be due to an overexpression of CDKN1C, an inhibitor of cell cycle progression, in the cortical VZ and SVZ of Arx KOs throughout corticogenesis. We also identified CDKN1C through transcriptional profile analysis of Arx KO cortices and showed that ARX is a direct regulator of Cdkn1c transcription. Later in development and postnatally cKO corticies showed a reduction of upper layer but not deeper layer neurons consistent with the IPC defect. The phenotype of the adult cKO mice is different from all other Arx mutant mice. They have no discernable seizure activity and are less anxious, less social, and more active when compared to their wild type littermates. Anatomically, there are significant changes with reduced cortical thickness and a hypoplastic corpus callosum and anterior commissure both consistent with a perturbation in cortical connectivity. Together, these data suggest specific structural and behavioral anomalies, common in patients with ARX mutations, are specifically due to alterations in pallial progenitor function. Furthermore, and of considerable interest, our data demonstrate that some of the neurobehavioral features found in patients with ARX mutations are not due to on-going seizures, as is often postulated, as the confounding variable of epilepsy was eliminated in these behavior analyses.