On ἀνδροτῆτα: is there a metrical anomaly at Iliad 22.363?

This is an intriguing line; indeed the three-line complex of which it is a part is in my view among the most beautiful in the Iliad, dealing as they do with one of the poem’s central and most venerable themes, the life and death of young men, in this case of three principals Patroklos, Hektor and Achilles.

῝Ως ἄρα μιν εἰπόντα τέλος θανάτοιο κάλυψε.

ψυχὴ δ’ ἐκ ῥεθέων πταμένη ῎Αϊδος δὲ βεβήκει

ὃν πότμον γοόωσα λιποῦσ’ ἀνδροτῆτα καὶ ἥβην.    363

Thus then the finality that is death covered him over.

The breath of his life flitted from his limbs, and off to Hades’

place it went, bewailing his fate as it left the manliness of his youth.

As you can see from the three excerpts below, the passage at Iliad 22.361-363 is virtually identical to the one at 16.855-857 (appropriately so) except that the latter refer to dying Patroklos’ foreboding to Hektor that he will meet a similar fate at the hand of Achilles, and the lines at 22.361-363 speak analogously (and certainly with an intentional thematic linkage) to dying Hektor’s similar warning to Achilles that Paris and Apollo will one day in turn kill him — a kind of unending cycle of deaths and dying.  And then at the opening of the final book of the poem, we find the same word repeated, but in a somewhat different context.  Consider the three passages in sequence:

Iliad 16.855-863

῝Ως ἄρα μιν εἰπόντα τέλος θανάτοιο κάλυψε•

ψυχὴ δ’ ἐκ ῥεθέων πταμένη ῎Αϊδος δὲ βεβήκει

ὃν πότμον γοόωσα λιποῦσ’ ἀνδροτῆτα καὶ ἥβην.

τὸν καὶ τεθνηῶτα προσηύδα φαίδιμος ῞Εκτωρ•

Πατρόκλεις τί νύ μοι μαντεύεαι αἰπὺν ὄλεθρον;

τίς δ’ οἶδ’ εἴ κ’ ᾿Αχιλεὺς Θέτιδος πάϊς ἠϋκόμοιο

φθήῃ ἐμῷ ὑπὸ δουρὶ τυπεὶς ἀπὸ θυμὸν ὀλέσσαι;

῝Ως ἄρα φωνήσας δόρυ χάλκεον ἐξ ὠτειλῆς

εἴρυσε λὰξ προσβάς, τὸν δ’ ὕπτιον ὦσ’ ἀπὸ δουρός.


Iliad 22.361-366

῝Ως ἄρα μιν εἰπόντα τέλος θανάτοιο κάλυψε,

ψυχὴ δ’ ἐκ ῥεθέων πταμένη ῎Αϊδος δὲ βεβήκει

ὃν πότμον γοόωσα λιποῦσ’ ἀνδροτῆτα καὶ ἥβην.

τὸν καὶ τεθνηῶτα προσηύδα δῖος ᾿Αχιλλεύς•

τέθναθι• κῆρα δ’ ἐγὼ τότε δέξομαι ὁππότε κεν δὴ

Ζεὺς ἐθέλῃ τελέσαι ἠδ’ ἀθάνατοι θεοὶ ἄλλοι.

Iliad 24.1-8

Λῦτο δ’ ἀγών, λαοὶ δὲ θοὰς ἐπὶ νῆας ἕκαστοι

ἐσκίδναντ’ ἰέναι. τοὶ μὲν δόρποιο μέδοντο

ὕπνου τε γλυκεροῦ ταρπήμεναι• αὐτὰρ ᾿Αχιλλεὺς

κλαῖε φίλου ἑτάρου μεμνημένος, οὐδέ μιν ὕπνος

ᾕρει πανδαμάτωρ, ἀλλ’ ἐστρέφετ’ ἔνθα καὶ ἔνθα

Πατρόκλου ποθέων ἀνδροτῆτά τε καὶ μένος ἠΰ,

ἠδ’ ὁπόσα τολύπευσε σὺν αὐτῷ καὶ πάθεν ἄλγεα

ἀνδρῶν τε πτολέμους ἀλεγεινά τε κύματα πείρων•

Thus, in summary:

ὃν πότμον γοόωσα λιποῦσ’ ἀνδροτῆτα καὶ ἥβην. 

ὃν πότμον γοόωσα λιποῦσ’ ἀνδροτῆτα καὶ ἥβην.

Πατρόκλου ποθέων ἀνδροτῆτά τε καὶ μένος ἠΰ,

What is anomalous in all three lines is the fact that ἀνδροτῆτά does not scan ‘properly’!  Let’s look more closely at the two different lines: in the first, the meter ‘implodes’ or ‘falls apart’ in the fourth metron;  in the second, in the third and fourth metra things collapse metrically:

ὃν πότμον γοόωσα λιποῦσ’ ἀνδροτῆτα καὶ ἥβην.

_     _ | _   ˘ ˘| _˘  ˘ | _   _     ˘  _  ˘  ˘ |_ _

_     _ | _    ˘  ˘| _  _ | ˘   _ ˘  ˘ | _  ˘  ˘ |_ _

Πατρόκλου ποθέων ἀνδροτῆτά τε καὶ μένος ἠΰ,

Certainly the line in question [22.363] and in all likelihood the whole complex is of great antiquity – pre-Mycenaean — demonstrably so on linguistic grounds.  And this brings us to the metrical ‘problem’.

It seems at first sight that ἀνδροτῆτά androtēta has to scan the initial alpha as long ‘by position’ (before the three consonants νδρ), but in fact it must be short to fit the metrical strictures of the dactylic hexameter.  The proto-Greek form of the semantic carrier (ἀνδρ[ο]) would be something like *anR[o]- (where R stands for syllabic liquid r – normally shown as a lower-case r with a small circle underneath it; and ditto for the other syllabic liquids m (M) and n (N) and l (L) – but of course I don’t have that character on my computer!).  In any event, R (syllabic liquid) reflexes into historical Greek as –ar/ra- or –or/ro-  depending on dialect, and this phonological shift had taken place by the time of Mycenaean Greek (c. 1400 B.C.).  A phonological analogue is classical trapeza ‘table’ < Myc torpeza (documented in Mycenaean, where it transliterates as topeza < proto-Greek *tRpeza – ‘three-footer/legger’).  Thus the original pre-Mycenaean form of the extended (by t) semantic stem in *androtāta was anRt- and as such the syllabic liquid (R) was realized vocalically as o between the two consonants n and t, and therefore the initial alpha in anRt- > anot- was not in a closed syllable and thus was short, a-no-t …, giving the correct scansion for the hexameter (the matter of the dental [d] in historical andr- will be taken up shortly).  What this word – and the context of these lines — demonstrates, then, is the great (pre-Greek i.e., Indo-European) antiquity of the formulas, i. e. part of a larger Indo-European rather than a localized Greek poetics dealing with the life and death of man.

What happened after the phonological change of pre Mycenaean anRt– to post-Mycenaean anrot-  had taken place was that the alpha became metrically long as part of a closed syllable (an-), which it is synchronically so to speak, of course, but this ignores the diachronic fact that this formula was presumably fossilized in the pre-Greek form with its pre-Greek metrical structure intact, and the metrical structure endured (so deeply, I assume, was this central theme embedded in the tradition of the epic language) in spite of the phonological change when it went into the specifically Greek tradition.  A marvelous example of the extreme conservatism of the tradition!  As noted above, the word occurs only three times in the Iliad (16.857, 22.163 and 24.6);  it does not appear in the Odyssey.  (The importance of Patroklos in the tradition is perhaps suggested by the appearance of androtēta in a slightly different but nonetheless related context at 24.6.)

A crucially relevant question is where the dental [d] comes from.  It is not really a matter of Greek phonology at all and not part of the word’s etymology;  it is more a question of a not uncommon phonetic tendency.  Sometimes the combination nasal (here, n) plus liquid (here, r) has a propensity to develop a homorganic epenthetic voiced stop.  I can illustrate from English.  The word thimble < OE Þýmel  < Þúma  ‘thumb’ (cf Ger Daumen, Sw tumme, etc. [no b as in Eng ‘thumb’]) plus an instrumental suffix -le (cf. hand-le lad-le etc) contains an original collocation of m (labial nasal) plus  liquid l in Þým-le and inserts a homorganic voiced labial stop [b];  a later reanalysis of *thumble yielded thumb-le > thumb rather than the more ‘correct’ *thum-[b]-le > *thum- , mistakenly taking the b as part of the etymological stem rather than a phonetically motivated epenthetic stop.  Cf. E thunder <*thun-d-R < *thunR < OE Þunor  [no epenthetic d] ~ L ton-āre [no epenthetic d].  Cf. also the Eng (dialectal) type family > *famly > ?fambly.

And, closer to home so to speak, consider Gr ambrotos ‘immortal’.  Here the b is non-etymological but strictly a phonetic epenthesis between nasal m and liquid r < *N-mR-to- (N and R are both syllabic [N being privative, producing vocalic alpha in Gr and >*en > in-  in Latin]), cf *mR- > L mor-t- .  Thus, in Greek, the phonological development is *N-mR-to > *a-mro-to- > [with labial epenthesis] a-mbro-to-  (cf. L calque-like formation [with assimilation of n in privative in- > m ] im-mor-t-al-).  This produces a later reanalysis of ambrotos as (falsely) derived from unassimilated an-brotos (where an- is ‘understood’ as the privative allomorph of a-) and resegments the word as am-brot-os , and the ‘new’ word brotos ‘mortal’ (common in Homer, e.g. Iliad 22.31) arises – the same identical diachronic phonetic mechanisms that are at work with Eng ‘thimble’ above!

Similarly, in androtēta , the Greek collocation of n (dental nasal) and r after the phonological shift from anR- to anro- inserts the homorganic voiced stop d between –nr- .  Thus, the d in andro- is an innovative phonetic phenomenon — utterly without relevance to the word’s strict etymology — that operates in some dialects (e. g, Attic, where we have the stem andr-) but not in others (e. g. stem aner- , clearly older, appears in Homer, as at Iliad 22.38 [alongside andr- , clearly more ‘recent’, as at Iliad 22.379]), where there is no epenthesis.

The point, then, is that in the pre-Mycenaean form it is not anr but anR (the ‘r’ is not consonantal ‘r’ but syllabic ‘r’), which means that like all syllabics it is descriptively a chamaeleon segment but must of course realize either vocalically or consonantally according to context in any given actuality:  vocalic interconsonantally and consonantal intervocalically.   What is sometimes hard to get your head around is that R is a theoretical construct with no ontological status as it were, and in no one particular example of a reconstructed pre-Mycenaean ‘real’ word is R just R but either non-syllabic ‘r’ or syllabic ‘a/o’  [that’s the whole point of R – it is either ‘r’ or ‘a/o’] — it can’t in any actual realization (i.e., in a given word) be an ambiguous segment but must be either vowel or consonant (think of the glides ‘i/y’ and ‘u/w’ – which share certain features with syllabics).

And in the root anR (which may well be H2nR – but let me not get bogged down in the murky byways of laryngeal theory, which really isn’t relevant to the problem we’re dealing with here), once the root is extended with the stem-marker -t-, you have anRt- which realizes in actuality as anot- (or a-no-t- ) and makes the first two syllables open and thus short.  In fact, in the pre-Mycenaean reconstruction you never do have the initial alpha followed by two consonants but by one consonant (n) and one vocalically realized syllabic (R) — and thus the alpha was not long.  This is the metrical structure from pre-Mycenaean that perdures into historical Greek.

But the phonological shift that has taken place by the time of Mycenaean changes original anRt into anrot- in historical Greek (Mycenaean) which in turn triggers the nasal-liquid epenthesis of d between nr, and yields unscannable androt-.

The answer, then, to the title question (“On ἀνδροτῆτα: is there a metrical anomaly at Iliad 22.363?”) is a resounding ‘NO’!

This entry was posted in CLASSICA, LANGUAGE. Bookmark the permalink.

3 Responses to On ἀνδροτῆτα: is there a metrical anomaly at Iliad 22.363?

  1. heather says:

    How did you come to know so much about linguistics? 😉

  2. Yes, historical linguistics often makes for riveting tales. You explain things very eloquently.
    One thing though, if I’m not mistaken some authorities today claim that word-inital *mr- would develop into *br- anyway, so that βροτός wouldn’t necessarily have to have arisen from ἄμβροτος through reanalysis. I can’t at the moment give any specific references though. What do you think?

  3. Aurélius says:

    Cf. δρώψ • άνθρωπος (Hésychius): the short alpha could build a short syllable before the cluster DR < *NR, δρώψ being an original *NR- (zero grade of anér ‘man’) + ōp- ‘face’.

    The line

    ἠδ’ ὁπόσα τολύπευσε σὺν αὐτῷ καὶ πάθεν ἄλγεα

    is more curiuos. Can you explain the long alpha in ὁπόσα? I can. But I would be glad to know your opinion.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s