If someone reads the sentence about Blake aloud, you can't tell whether there is a comma after "collection" and before "Songs of Innocence and Experience" or not. And yet you can understand—pretty well—what the sentence means. For people who view spoken language as the only real language, and grammar as the system of conventions that permits spoken language to be understood as intended, there is no grammatical issue related to the presence or absence of the comma after "collection" and before "Songs of Innocence and Experience" in your last example.
What you have instead is a punctuation issue—a matter specific to written language. The Chicago Manual of Style is one of a number of books dedicated to written language and specifically to formulating guidelines that try to standardize punctuation, formatting, and other aspects of written language. Because written language permits various distinctive visual treatments of words, phrases, sentences, paragraphs, subsections, sections, chapters, parts, volumes, etc., it can (if handled consistently) convey ideas with a degree of precision that spoken language cannot easily equal.
In the case of the comma in the sentence about Blake, including the comma if Blake wrote only one collection of poems and omitting the comma if he wrote more than one collection of poems would provide a simple, immediate signal to readers Blake's poetry collection output—if everyone agreed that that's what including or omitting the comma after "collection" and before "Songs of Innocence and Experience" meant. In advising writers and editors to enforce that convention consistently, The Chicago Manual of Style attempts to promote a world in which written English possesses this type of sophisticated signaling and precision of meaning.
Of course, that isn't the world we live in. Lots of people who write have never even heard of The Chicago Manual of Style (or The AP Stylebook or The Oxford Style Guide or A Uniform System of Citation or The MLA Style Manual or any of the dozens of other style guides that exist), and many others are aware of it but (for one reason or another) reject it. And without agreement on what a comma located in such-and-such a place means, we are in much the same boat that bridge players would be in if there were no agreement about what a bid of one club or one no-trump meant.
Compounding the problem, writers tend to apply punctuation and other style rules inconsistently, either because they are unsure of a particular rule or because they forget or make a mechanical mistake. The net result is that comma use in restrictive and nonrestrictive clauses is wildly inconsistent and therefore of limited practical value, at least until you can figure out how consistently the writer seems to enforce style-guide preferences regarding such punctuation.
Many people look at style guide advice as "prescriptive"—meaning that the style guide's authors are trying to boss writers around and tell them what is right or wrong in some absolute sense. My own view is that style guide advice is not so much prescriptive as aspirational: it imagines a world in which consistent punctuation and formatting of like constructions would enable writers to convey their intended meaning more accurately than is possible in a world where disagreement over desirable forms is very common and where people who share the style guides' aspirations are few.
If I see a sentence that begins with
William Blake's collection, Songs of Innocence and Experience, ...
and I have no basis for gauging whether the writer supports or ignores The Chicago Manual of Style, my working assumption is that the writer is telling me that Blake wrote one collection and that its title was Songs of Innocence and Experience. Conversely, if I see a sentence that begins
William Blake's collection Songs of Innocence and Experience ...
I assume that the writer is telling me that Blake wrote more than one collection but that Songs of Innocence and Experience is the one of present interest. I am not an especially credulous reader, however. If I see other instances where the writer shows signs of not being particularly careful about his or her comma use, I quickly adopt a skeptical view of all the person's punctuation use. That is, I cease to assume that I know what the writer means when, for example, I see commas (or no commas) setting off what might be either a restrictive clause or a nonrestrictive clause.
To return to the posted question, asking about the grammatical rightness or wrongness of this or that punctuation choice runs athwart of many grammarians' view of what grammar is. Asking about the potential usefulness of consistently applying punctuation in accordance with a particular style rule will undoubtedly draw the reaction that any style rule is as good as any other, since they all are made up out of thin air and widely ignored.
And yet I think that aiming for consistency in punctuation and formatting is worthwhile—and indeed quite useful for readers who are attuned to such consistency. You don't have to believe that using commas to signal a nonrestrictive clause is a matter of virtue to see how it can sometimes help resolve ambiguity if the writer makes a sustained effort to punctuate restrictive clauses consistently and distinctively. We can't expect punctuation to do more than that in our world.
