Traditional notions of dialect are qualitative, and perceived dialects are difficult to reconcile with the evidence of actual production of dialect feature. Traditional dialect boundaries are similarly tied to qualitative judgements. More recently, various quantitative approaches have been attempted to try to capture dialect boundaries, each with its won problems. We have achieved some success with an objective quantitative analysis of separate dialect features from the Linguistic Atlas of the Middle and South Atlantic States (LAMSAS). Our findings suggest that a mathematical modeling of areal variation of dialect features, combined with a reassessment of traditional notions of dialect, could contribute to our knowledge of language, as well as document regional differences at the time of the LAMSAS survey.