English Subject-auxiliary inversion

In normal declarative English sentences, let's call them 'canonical' sentences, the verb comes after the subject.

• Bob is walking the Great Wall of China.
• Bob likes elephants.

But in other types of construction and in certain other situations the auxiliary verb comes before the subject. One notable instance is in questions:

• Is Bob walking the Great Wall of China?

Here we se the auxiliary verb BE occurring before the Subject. If there is no auxiliary verb in the normal declarative version of the sentence then we need to insert the auxiliary DO:

• Does Bob like elephants?

My question here is twofold. Firstly:

1. What are the different constructions or environments in which either we require subject-auxiliary inversion in modern English, or can optionally use it?*

Secondly, and more importantly:

2. Is there any generalisaton we can make about these constructions. Does subject-auxiliary inversion have some kind of meaning in all of these, or a subset of these?

*_{Not including poetry, for obvious reasons.}

Solution 1:

The only time outside of poetry that I can think of where it is not optional would be in the so-called tag question.

Bob likes elephants, doesn't he?

Solution 2:

Offhand, I can only think of two, uh, constructions ... or environments:

Not only did they dislike verismo opera intensely, they weren't even remotely interested in music as an art form.

And this, from Robert Louis Stevenson's Heather Ale:

The king rode, and was angry.
Black was his brow and pale:
To rule in a land of heather
And lack the Heather Ale.

Solution 3:

Bob likes elephants. 

Sure, this is in canonical order with an S/V/O pattern.  It's also indicative and declarative. 

Does Bob like elephants? 

This isn't declarative.  It's interrogative. 

Should Bob like elephants, we can plan a visit to the zoo. 

This isn't indicative, although it might not be fashionable to call it subjunctive anymore. 

Not only does Bob like elephants, he practically worships them. 

This still feels indicative and declarative, but the entire clause is negated from the outside.  It's not as straightforward an indicative declaration as it is when it stands on its own.  If the clause including its external negation seems indicative, then the clause per se, excluding that qualification, may well be something else. 

One thing that traditional grammar gets wrong is the way it lumps its declarative/interrogative distinction in along with its indicative/subjunctive distinction.  There's a tangle of two or three orthogonal properties that are traditionally labeled mode: indicative, interrogative, subjunctive, exclamatory and imperative are values that all leap to mind. 

Something pulled that snarl together.  This inversion marks some non-default value in that tangled mess.  Do I think that it's a likely candidate?  Do I ever!