Scansion of lines in Homer involving εἰνὶ θρόνῳ

Question

Scanning Homeric verse is something I'm not very experienced at yet, and I have a question about these two lines involving the phrase εἰνὶ θρόνῳ:

σείσατο δ’ εἰνὶ θρόνῳ, ἐλέλιξε δὲ μακρὸν Ὄλυμπον, (Iliad 8.199)

ἕζετο δ᾽ εἰνὶ θρόνῳ· τὼ δ᾽ ἀΐξαντε πετέσθην. (Iliad 15.150)

The only way to scan the first one seems to be this:

σείσατο | δ’ εἰνὶ θρό|νῳ, ἐλέλ|ιξε δὲ |μακρὸν Ὄ|λυμπον,

I would think that the second ι in εἰνὶ would be long by position, but if I do that, I can't make the rest of the line scan.

The second one comes out similarly:

ἕζετο | δ᾽ εἰνὶ θρό|νῳ· τὼ | δ᾽ ἀΐξ |αντε πετ|έσθην.

(In ἀΐξαντε, the α is long phonetically.)

I'm new to this kind of thing, and one thing I'm not really very clear on is how strictly the rules apply, or to what extent you can just make a vowel long or short because it's necessary. I'm having a hard time understanding the logic of the subject. I have Pharr, who lists a long list of rules, but it's not clear whether the rules are absolute or can be bent, and some of the rules seem like they can contradict one another, so it's not obvious what priority to give them.

It does seem like it would be odd artistically to devote an entire ponderous spondee to a humble preposition like εἰνὶ.

A similar situation seems to come up here:

ἤν τίς τοι εἴπῃσι βροτῶν, ἢ ὄσσαν ἀκούσῃς (Odyssey 1.282)

The wiktionary entry for βροτός specifically remarks that the initial βρ of this word has to be treated anomalously, the evidence being the meter of this line. (The genitive βροτῶν is extremely common.) Looking around for words that had the same phonetic pattern, I found ξένος, which seems to evade the metrical difficulties because in Homer it's ξεῖνος, so we get ξείνῳ and ξείνων rather than ξένῳ and ξένων.

Answer 1

This is known as correption, and in particular Attic correption, which displays this more frequently than Homeric verse.

From Halporn, Ostwald, and Rosenmeyer (a great little student reference guide) p.5:

The first syllable of a word like πέτρος is counted either short or long; the treatment of the combination of a mute (γ, β, δ, κ, π, τ, χ, φ, θ), with a liquid (λ, ρ, more rarely μ, ν) varies depending on whether the mute is felt to close the preceding syllable (πετ-ρος) or, in association with the liquid, to begin the next syllable (πε-τρος).

    Example: Sophocles, Philotetes 296:
                    ἀλλ ἐν πέτροισι πέτρον ἐκτρίβων μόλις
                                  ˘            ¯

With πέτροισι, -έτρ- here scan short, while with πέτρον, -έτρ- scans long, both by the same author, in the same play, in the same line. Since the next word is theta-rho in your example, the second iota in εἰνὶ scans short by position.

They do note that lyric poets would have "always" (their emphasis) scanned that epsilon long, and Homer typically follows that pattern.

However, calling it Attic correption is a bit of a misnomer. West in his Introduction to Greek Metre says that it in early poetry it usually occurs only in the first syllable of a word (as you note in the comments) or in "words which the verse would not otherwise permit." I would say that θρόνῳ would be considered such a word.

Another word whose mute+liquid initial syllables do not always make the preceding vowel long is δράκων, as in the line:

ἔνθ’ ἐφά|νη μέγα | σῆμα· δρά|κων ἐπὶ | νῶτα δα|φοινὸς

So it's not a hard and fast rule.