This paper looks at the need for a better understanding of the impediments to critical thinking in relation to graduate student work. The paper argues that a distinction is needed between two vectors that influence student writing: (1) the word-level-sentence-level vector; and (2) the grammar-inferencing vector. It is suggested that much of the work being done to assist students is only done on the first vector. This paper suggests a combination of explicit use of deductive syllogistic inferences and computer-aided argument mapping is needed. A methodology is suggested for tackling assignments that require students to "make an argument". It is argued that what lecturers understand tacitly, now needs to be made a focus of deliberate educational practices. (Contains 3 figures.)