This is a sequel to the post "Shepherd on the First Cause," which should be read first.
The key to understanding Shepherd's comments lies in grasping her theory of causation. (It is useful to keep in mind that Hume is her constant foil.) Here is a rough attempt to characterize this theory.
Shepherd sees causation as a "mixture of qualities." Suppose you have a cause (C) and an object (O) or co-cause. Each of these has a number of properties or qualities. When C acts on O (or combines with O), the qualities 'mix'. The mixed result is the effect. So, for instance, I punch my fist into clay, the resulting impression is a 'mixture' of some of the properties of my fist and some of the properties of the clay.
Take another example, which will perhaps give a clearer idea of the significance of the view. There is a book on a sturdy table. That the book does not fall through the table is necessitated by the combined properties of the book and the table. It is not impossible, of course, for books to fall through tables; but it is only possible if some of the properties of either the book, or the table, or both, are changed. That is, a change can be induced in the situation only by introducing new properties into the mix. These new properties are causes of new situations. This provides Shepherd with a very strong response to Hume's view that "Whatever begins to exist has a cause" is not a necessary proposition. On Shepherd's view of causation, it is necessary, because every change of properties requires the introduction of new properties. On the mixture view of causation, Shepherd thinks, anything new is necessarily an effect of a new introduction of properties. This, of course, is exactly right; you can't change the properties of a situation without changing its properties; if you have a new set of properties, it can only be because some properties in the set are new.
This is a very elementary summary of her view; I hope to cover more precise details in later posts.