Recently, high intensity short-term therapy with a heavy emphasis on verbal language (called constraint induced language therapy, CILT) has gained momentum in aphasiology. However, the entire extent of its applicability and limitations has not been fully studied, especially with regard to specific aphasic deficits. This thesis sought to: 1) determine the efficacy of the originally published CILT protocol (o-CILT) with a deficit specific population (four individuals with agrammatic aphasia) and 2) examine the potential effect of a modified CILT protocol, which additionally focused on grammatical accuracy (g-CILT). Results revealed differences between the performance of individuals with agrammatism in this study and previously published CILT data. Findings also demonstrated that participants receiving g-CILT produced more significant gains on tests of aphasia severity and grammaticality, while individuals receiving o-CILT showed more highly significant changes on discourse measures of grammaticality. This paper suggests that, for individuals with agrammatism, CILT in its original form may not evince significant changes on tests of aphasia severity and grammatical production and a grammatical modification appears to increase the efficacy of CILT.