In the assignment for Monday (pp. 60-73), Kripke presents the description theory of names in the form he intends to criticize and then begins to offer a detailed criticism of it; this criticism is continued in Wednesday's assignment (pp. 73-90).
The theory is presented at the end of Lecture I (pp. 64f, 68) and summarized again at the beginning of Lecture II (p. 71); Kripke explains the final point (C) further in his Addenda (pp. 160-162). Here is a slight restatement of the theory with an indication (in lighter type) of the role of its various components. At the right, I've noted locations where Kripke criticizes various aspects of the theory.
| The description theory | Criticisms | |||
| I. As a definition, let us say for speaker A and designating expression X: | ||||
| (1) the cluster that A associates with X is the collection φ1, φ2, ... of properties φ such that A believes 'φX'--i.e., such that A believes 'X has the property φ'. | ||||
| II. Then the following hold: | ||||
| A. (2) there are properties in the cluster that A associates with X which are believed by A to pick out an individual uniquely; | (2) pp. 80-82 | |||
| B. The reference of the name X in the language as A speaks it is determined by the cluster that A associates with X in the sense that | ||||
| (3) any object with a (weighted) majority of the properties in the cluster is the referent of the name, and | (3) pp. 82-85 | |||
| (4) if no unique object has a (weighted) majority, then the name does not refer | (4) pp. 86-87 | |||
| C. Having most of the properties φ1, φ2, ... in the cluster that A associates with the name X is part of the content of the name for A in the sense that we can say of the sentence 'If X exists, then X has most of the properties φ1, φ2, ...' that it | ||||
| (5) is known a priori by A | (5) pp. 66-67, 87 | |||
| (6) is a necessary truth in the language as A speaks it | (6) pp. 61-62, 74-78 | |||
| III. Moreover, this account of reference is non-circular in the sense that | ||||
| (C) any involvement of the concept of reference in the properties of the cluster is ultimately eliminable | ||||