Where to put the enclitic -ne?

Question

The enclitic particle -ne can be used to form a binary (yes/no) question. I would like to know how attaching it to different words changes the meaning of the question. I have a clear idea of how it should behave, but that may well be because this is exactly how the Finnish enclitic -ko/-kö works, so I want confirmation from someone without a Finn's bias.

This is my idea:

The question concerns the word -ne is attached to. For example, consider these questions related to Gaius cum Marco loquitur:

1. Loquiturne Gaius cum Marco? — Does Gaius speak with Marcus? (The question is whether speaking takes place.)
2. Gaiusne loquitur cum Marco? — Is Gaius the one who speaks with Marcus? (The question is whether Gaius is the one who speaks.)
3. Cum Marcone loquitur Gaius? — Is Marcus the one Gaius speaks with? (The question is whether Marcus is the one spoken to.)

(For some reason I prefer to put the word with -ne towards the beginning of the sentence.) It is most common to attach -ne to the predicate, and it does change the meaning to put it elsewhere. Here is another set of examples:

4. Legisne librum bonum? — Are you reading a good book?
5. Bonumne librum legis? — Is the book you are reading good? Or maybe: Are you reading a good (with emphasis) book?
6. Librumne bonum legis? — I don't seem to be able to parse this as naturally. Perhaps: Are you reading a good book (emphatically, as opposed to watching a good movie)?

Answer 1

Your idea is correct.

Lewis-Short is not terribly clear:

added in a direct question, as an interrogation mark, to the first or principal word of the clause

but, if you know German, Georges is much clearer: ‑n(e) is attached to the focus of the question, therefore mostly at the beginning of the sentence.

To answer your question completely, a quotation from Terentius, Phormio, v.851 is very interesting: Sed isne est, quem quaero an non?

• It shows ‑ne attached to the word asked about, instead of the verb: is it he?
• It is an example of what can precede the ne-marked word in direct questions: connectives and such stuff, not real constituents.
• It reminds us that alternatives can be introduced by an; sometimes ‑ne is repeated instead.

If you read on at the Georges link, you see that, when used as a conjunction in indirect questions, ‑ne can be attached to words in whatever position; however this is not the case for direct questions.

In particular, I have strong doubts about your cum Marcone: I couldn't find a single example of a prepositional phrase Prep+Noun+ne, and I would say it is impossible: the forms justified by the extant corpus are Nonne cum Marco? (confirmation expected) or Num cum Marco? (denial expected).

Mecumne loqueris? does work and is OK with me, even if I could find it only in modern exercise books, but here ‑cum works as a suffix, almost a case ending, and it is a construction closer to mihine than to the PP above. Possibly even magnone cum gaudio could work, but I couldn't find any examples. I don't know what prevents enclitics to be attached to nouns after prepositions: but I'd bet the same holds for ‑que and ‑ve as well.

A difference with Finnish (I think: please correct me if I'm wrong) is that ‑ko/‑kö is used exclusively and obligatorily in yes/no questions, while ‑ne is not obligatory (at least in later periods) and can sometimes (rarely) be attached to interrogative adverbs and pronouns: see the above links for examples.