Great works

Recruiting ἀγαθός as another potential example of w-stem hypercorrection in Ancient Greek

Following on from my last post on the origin of the greek word for horse, a perceptive reader pointed out that it should be possible to unearth at least a few more examples of the kind of my proposed paradigm-related hypercorrection.

de Vaan’s original paper on the subject gives us three which are potentially interesting here: δμώς, ἥρως, and πάτρως, which are posited as original u-stem nouns with an irregular nominative in -ōu, reflecting

• *sēr-ōu “raider, one who captures”

• *dm-ōu “belonging to the house(hold)”

• *ph₂tr-ōu “belonging to the father(‘s family)”

In the cases of δμώς and πάτρως, we apparently see cases where the stem of the noun, in the zero grade, has been generalised from the oblique stems of hysterokinetic/hysterodynamic nouns in -u, while the suffix has been generalised from the nominative insertion of a long ō between the stem and the —w suffix. In ἥρως, the long ē is presumed to be retained from an original root noun. (So presumably following an even older pattern *sēr-us, *seréus, *sruós, perhaps?)

The picture we get here is one where *-w or *-u as a suffix originally functioned to create a broad class of subordinated or relational nominals from other words. I say “subordinated” because they are not always exactly possessive, unless one construes *sēr-w as “with acquisitiveness” or “having loot”.

I might explore each of these forms in more detail later because I’m not sure I’m entirely happy with this explanation. For now, however, I’d like to briefly turn to another, slightly more exotic candidate: ἀγαθός.

For all the good it does you

ἀγαθός is an odd duck. Its expanded sense of “good” or “noble” generally takes comparatives and superlatives based on different roots, principally those in bel- (βελτίων, βέλτερος “better, superior”, βέλτιστος “best”) and *h₂ér- (ἀρείων, “superior”, ἄριστος “best”).

It is quite possible that the “base” form of both of these roots were lost from the language because of their close similarity to other words: a hypothetical *βελος or βελιος would have been close to βέλος “missile, dart” and βαλιός “swift” or “dappled”.

The ordinary adjectival form of *h₂er-, *h₂er-yos, would have been realised in Greek as *ἀιρος and would have been confused in most instances with a reflex of the verb αἴρω, ἀείρω “to raise up, praise, win, or remove”. We may see one instance of *ἀιρος in its original sense preserved in the personal name Αἴροπος (perhaps “of noble countenance”), but this is attested only once.

ἀγαθός must have arisen to fill the gap, and demands an explanation of some kind. Beekes suggests a pre-Greek origin. Beekes is, in my opinion, far too trigger-happy when it comes to declaring things pre-Greek, especially when an Indo-European explanation is available.

“Great Works”

Our tentative IE origin for ἀγαθός is supplied by Panagal: *m̥ǵh₂dʰh₁ós < *m̥éǵh₂- “great” + *dʰeh₁- “do/make”, with a sense “made great” or “one whose deeds are great”.

Beekes rejects this meaning on semantic grounds, arguing that “good” is more likely to be extended to mean “noble” than vice versa.

However, we can point to several examples where this is not the case, most obviously “noble” in English but also reflexes of *h₂er-yos which must necessarily extend the meaning of the root “fit” to mean “one who is fit”, or “a peer” in the sense of a member of the nobility (as reflected in its Germanic, Celtic, and Indo-Iranian reflexes) and comes by extension to mean “praiseworthy, hospitable” and “noble” in the purely aesthetic sense of “good”.

Such a progression would also be entirely in line with the value system of Homer’s characters, in which great achievements are prized over anything else.

The problem

*m̥ǵh₂dʰh₁ós as a form is, however, not without its issues. Endings in -dʰo- are fairly uncommon. But we can perhaps account for its odd form by assuming an original hysterodynamic u-stem *méǵh₂dʰh₁us with the following paradigm:

Nom. *méǵh₂dʰh₁us

Acc. *m̥ǵh₂dʰh₁éum

Gen./Abl. *m̥ǵh₂dʰh₁wós

This configuration would have yielded in Proto-Greek (following laryngeal loss):

Nom. **megatʰeus (accent unmarked, perhaps shifting to the a)

Acc. **agatʰéum

Gen. **agatʰwos

Obviously, this paradigm is highly irregular, and would have been intuitively resolved by assuming the genitive as a nominative, with the original nominative and accusative forms lost entirely. This process would have likely been sped along by analogies to other reflexes of the zero-grade form *m̥ǵh₂- such as the adverb ἄγαν, formed from the original accusative (*m̥ǵéh₂m or *m̥éǵh₂m̥) of *méǵh₂s.

You’ll note that I progress *m̥ǵh₂dʰh₁wós to *agatʰwós rather than **agatʰewos in this reconstruction because the h₁ would likely have experienced deletion in this sequence before being vocalised. A w before o would also have vanished without trace by the time of Homer.

This is of course entirely speculative

But it is, at the very least, falsifiable. If we happen to come across a Mycenaean reflex of this word, we should expect it to show a w (perhaps something like *𐀀𐀏𐀵𐀺) we will have confirmation. Until then—