Donnellan called attention to what he called the referential use, as opposed to the attributive use, of a definite description.
Donnellan’s objection to the Theory of Descriptions is just that the theory overlooks the referential use; Russell writes as if all descriptions were used attributively.But, against Strawson, Donnellan complains that equally he did not see the attributive use, that Strawson writes as if all descriptions were used referentially, in a context, to draw somebody’s attention to a particular person, place or thing.Thus both Strawson and Russell were mistaken in thinking that definite descriptions always work in one way, because there is an ambiguity acknowledged by neither.Donnellan does not take a position as to what kind of ambiguity it is; in particular he does not try to decide whether the sentence (12) itself has two different meanings explaining the description’s evidently distinct “uses.”
The passage you quote is from Keith Donnellan's paper, "Reference and Definite Descriptions". Donnellan references a disagreement between Russell and Strawson about the correct way to understand definite descriptions, that is constructions that might have the form "the F". Russell in the classic paper "On Denoting" treats them as being equivalent to "the one and only thing that is F". So, for example, "The present king of France is bald" means, "there is one and only one thing that is the present king of France, and that thing is bald". Strawson in his paper "On Referring" criticises this analysis as failing to do justice to the fact that when people utter sentences that include definite descriptions they frequently, and perhaps typically, are referring to particular individuals and are not making general statements about things that might or might not exist.
Donnellan claims that Russell and Strawson are both right and both wrong. He holds that there are some uses of definite descriptions in English that are attributive (Russell's kind) and some that are referential (Strawson's kind). There are even some that might be either. He offers as an example, "Smith's murderer is insane". This could be used attributively to mean, the nature of this murder is such that the person who did it, whoever that is, is insane. Or it could be used referentially to mean, that person Jones who is in the dock and has confessed to killing Smith, is insane.
Kripke disagreed with Donnellan's claim that there are two different kinds of definite description. He leaned instead on the distinction that is familiar from the work of Paul Grice, namely the distinction between the conventional meaning of a sentence and the meaning or intention of a speaker when uttering the sentence. The latter may differ from the former because of pragmatic considerations. So we might understand the conventional semantic meaning of a sentence to be that of Russell's theory, while a speaker may have a referential intention when uttering the sentence.
Keith Donnellan. "Reference and Definite Descriptions". Philosophical Review 77, 1966, pp.281-304.
Paul Grice. "Logic and Conversation". Part 1 of Studies in the Way of Words. 1989, Harvard UP.
Saul Kripke. "Speaker's Reference and Semantic Reference" Midwest Studies in Philosophy vol. II, University of Minnesota Press, Minneapolis, 1977, pp. 259-76.
Bert Russell. "On Denoting". Mind 13, 1905, pp. 479-493.
Peter Strawson. "On Referring". Mind 59, 1950, pp. 320-44.