If anybody is confused by ictus, they can just read Goodwin’s definition of it here in Part V of his Grammar, the section on Greek Versification. It literally just means “beat” and has been used that way since Ars Poetica, at least. The hexameter has six beats. West’s (ugly, painful, to the crows!) terminology for this is “princeps” in the hexameter.
If vowel lengthening were a phenomenon only of dropped consonants (digammas, etc.), we should expect to see it on the second part of the foot (the “thesis” in the Latin usage, but “arsis” in the Greek) just as frequently as we see it on the first. That is, we would see it wherever the old consonant was lost. Instead, we see it on the first part of the foot the vast majority of the time, the ictus (“arsis” in Latin usage, but “thesis” in Greek). The beat or place where you tap your foot (θέσις).
Anyway, Paul misunderstands West and the general orthodoxy here. In fact, West describes this lengthening as a type of performance fudge, or “prosodic license” in the hexameter. See West’s discussion on pg. 17-18, and 21-22 of Intro to Greek Metre.
Where I do depart from West is his suggestion that this is a sort of performance-time consonant fudging. In fact that would be a violation of what I’d call the simplest Homeric Law: The vast majority of Homeric lines are easy to scan as you go, without knowing what word comes next. You can generally put your hand or a cue-card over the rest of the line as you read. This is actually the whole reason why this sort of fudge/license is always on the ictus. Off-ictus, you would have to know what was coming next.
–
Example: The 20 lines Χ301-320 contain the following variations from “vanilla” meter:
Χ301 οὐδ’ ἀλέη· ἦ γάρ ῥα πάλαι τό γε φίλτερον ἦεν
Missing correption, on the ictus: ἀλέη· ἦ
Χ303 πρόφρονες εἰρύατο· νῦν αὖτέ με μοῖρα κιχάνει.
Lengthening, on the ictus: πρόφρονες εἰρύατο
Χ305 ἀλλὰ μέγα ῥέξας τι καὶ ἐσσομένοισι πυθέσθαι.
Lengthening, on the ictus: μέγα ῥέξας
Χ307 τό οἱ ὑπὸ λαπάρην τέτατο μέγα τε στιβαρόν τε,
Lengthening, on the ictus: (x3) τό οἱ, ὑπὸ λαπάρην, τέτατο μέγα
Χ309 ὅς τ’ εἶσιν πεδίον δὲ διὰ νεφέων ἐρεβεννῶν
Lengthening, on the ictus: διὰ νεφέων
Χ310 ἁρπάξων ἢ ἄρν’ ἀμαλὴν ἤ πτῶκα λαγωόν·
Missing correption, NOT on ictus: ἢ ἄρν’
Χ314 καλὸν δαιδάλεον, κόρυθι δ’ ἐπένευε φαεινῇ
Lengthening, on the ictus: κόρυθι δ’ ἐπένευε
Only 310 is a violation of the Law given above. Every line except for 310 could be read with your hand obscuring the rest of the line as you read. But here, you have to pronounce ἢ without correption. This, of course, really is a fossil consonant. It is a dropped digamma. The word was Ϝάρνα, and I (again disagreeing with West) suspect that it really was pronounced that way by the man who composed most of the Ϝάρνα lines in the poem.
Compare the sort of violation of this rule that we almost immediately get from a post-Homeric poet, the more recent composer of the Astronautilia:
Α2: εὔχομαι ἐκπάγλως καὶ Ἀπόλλωνι ἄνακτι
Here, we have a skipped correption of καί and Ἀπόλλωνι coming off-ictus, forcing someone reading aloud to have to repeat the line after he screws it up on first read. Nothing about this line is impossible to say aloud, once you have read it or heard it once. But for whatever reason we don’t get so many lines like this in Homer (and of course I suspect that the reason is because Homer was composing an incredibly long epic meant to be easily performed aloud while reading).