The success of a software project is largely dependent upon the quality of the Software Requirements Specification (SRS) document, which serves as a medium to communicate user requirements to the technical personnel responsible for developing the software. This paper addresses the problem of providing automated assistance for assessing the quality of textual requirements from an innovative point of view, namely through the use of a decision- tree-based text classifier, equipped with Natural Language Processing (NLP) tools. The objective is to apply the text classification technique to build a system for the automatic detection of ambiguity in SRS text based on the quality indicators defined in the quality model proposed in this paper. We believe that, with proper training, such a text classification system will prove to be of immense benefit in assessing SRS quality. To the authors' best knowledge, ours is the first documented attempt to apply the text classification technique for assessing the quality of software documents.