Is the σου in καλή σου ἡμέρα out of place in Northern Greek

By: | Post date: 2017-11-01 | Comments: 1 Comment
Posted in categories: Linguistics, Mediaeval Greek

In a previous post, I mused that the use of καλή σου ἡμέρα “Good day to you” in Constantinople, in texts such as De Cerimoniis from the 10th century, was problematic—since by then the dialect split was meant to be in place, between genitive pronouns in the South (Southern Italy), and accusative pronouns in the North.

Looking at the expression more closely, I think I was wrong about that. I don’t think καλή σου ἡμέρα is at all out of place in 10th century Constantinople, or for that matter 20th century Constantinople.

For the first millennium AD, accusatives and genitives competed to replace the dative in papyri and inscriptions, as Humbert documents in his 1930 monograph. 1 Up until the 10th century, the genitive is more common, but the accusative turns up throughout the Greek-speaking world. The geographical division only sets in around the 10th century, with the monastic acts published by Trinchera and Cusa from Southern Italy on the one hand (and they really only start in the 11th century), and on the other inscriptions from Thrace and Macedonia—and texts like the De Administrando Imperio and the De Ceremoniis, both attributed to Constantine Porphyrogenitus.

The picture is muddied somewhat with nouns, where prepositional indirect objects are more common than inflected ones; and in plural personal pronouns (μας, σας, τους, τις)—which are derived from accusative forms throughout Greek, even if they are used as indirect objects or possessives.

In fact, the genitive plural pronouns, ἡμῶν, ὑμῶν, αὑτῶν, have vanished without a trace in Modern Greek. The regular developments *μῶ(ν), ἐσῶν > *σῶ(ν), *τῶ(ν) would have been insufficiently distinct from their singular counterparts μου, σου, του. Tsakonian preserved its cognates, ἁμῶν > νάμου, ὑμῶν > νύμου; but the metanalysis of their initial /n/ guaranteed that its genitive plural pronouns remained distinctive.

But with singular personal pronouns, the division is clear: δώσε μου το “give me.GEN it” south of Thessaly, δώσε με το “give me.ACC it” from Thessaly up.

Which brings us to expressions like καλή σου ἡμέρα/καλημέρα σου “Good day to you.”

Southern Greek has a suite of expressions expressing attitudes towards someone, and consisting of a nominal or interjection followed by a genitive pronoun—which we would by default interpret as indirect objects:

  • καλημέρα σου/καλή σου μέρα “good day to you”
  • καλό σου βράδυ “good evening to you”
  • καλησπέρα σου “good evening to you”
  • καληνύχτα σου “good night to you”
  • γεια σου “health to you = hello”
  • γεια χαρά σου “health and happiness to you = hello”
  • χρόνια σου πολλά “many years to you” (happy birthday, happy anniversary)
  • μπράβο σου “bravo to you = good for you!”
  • έννοια σου “worry to you = look out” (warning, threat)
  • αλί/αλοίμονό σου “woe to you = you’re in trouble (if)” (threat)
  • χαλάλι σου “it is halal (Islamically lawful) to you = I forgive you your minor transgression, because it was worth it/because I still like you”

I was going to add the Cretan ξ(ι)α σου (Cretan) “never mind, as you wish”, but that derives from εξουσία “power, right”, as in “it’s your right to think so.” In other words, that’s a straightforward possessive pronoun, and not an indirect object.

I was trying to translate those into Constantinopolitan, and it didn’t make sense to me that people up North say γεια σε, αλοίμονό σε, χαλάλι σε. I don’t think that pattern has generalised there. (If there’s an exception, it’ll be in Pontic.) And if it hasn’t, that would be consistent with the fact that Porphyrogenitus uses the accusative instead of the genitive in pronominal indirect objects—but uses the genitive in καλή σου ἡμέρα.

So why has the expected accusative in such nominal wishes or threats been avoided in Northern Greek?

I’m going to suggest three possibilities.


The first possibility is that the pronoun really isn’t an indirect object at all, but a possessor; just as the σου in ξια σου really means “that’s your right”, these expressions might have originally meant “καλή σου ἡμέρα “may your day be good”. The topicalising use of the pronoun would be consistent with that: καλή σου [ἔστω ἡ] ἡμέρα “good your [be the] day”. And γεια σου corresponds in English to the possessive in [to] your health.

Possible, but not likely. The expressions correspond to Koine expressions in which the dative is used explicitly. Two renowned instances occur in the New Testament: γεια σου “health to you” corresponds to εἰρήνη ὑμῖν “peace be with you”, and αλοίμονό σου “woe to you” to οὐαὶ ὑμῖν “woe unto you.” As it turns out, De Cerimoniis uses one of those datives itself: καλὴ ἡμέρα ὑμῖν, ἄρχοντες “good day to you, my lords”. It is more plausible that the newer expressions continued to be understood with indirect objects. And a possessive meaning would be impossible for more recent expressions like μπράβο σου “bravo to you” or χαλάλι σου “it is lawful to you”.


The second possibility is that the original expression, “good day to you”, was already well established with a genitive pronoun, before Greek dialects split along dative vs accusative indirect objects. The expression was frozen by the time the split happened, and Northern dialects kept it—and even allowed new expressions with genitive pronouns to be formed, by analogy.

That account is possible, though it’s not very satisfying linguistically. There are frozen expressions in Modern Greek based on the dative too (δόξα σοι ὁ Θεός “Glory to Thee, God” = “Thank God”), though those were preserved through the church; the preservation of the genitives in wishes doesn’t seem as well motivated.

The earliest example I can find of the expression “good day to you” however is old enough to suggest that the expression was well established. It is from the Vita of St Auxentius, which is dated to the fifth century AD; the Vita is supposed to have been written by Vendemianus, disciple of Auxentius of Bithynia, and Vendemianus died in 512. So says Google, though not Migne’s edition, which instead refers to a disciple called Sergius.

The text does not appear to have a modern edition, and who knows how reliable that date is, and whether scribes have tinkered with it. Migne’s edition says it’s from a 10th century manuscript; and it comes from the collections of saints’ lives by month due to Symeon Metaphrastes (late 10th century). (That is, “the translator”—from Koine back to a more respectable language.) Significantly in this case, the greeting with the genitive is put in the mouth of a “rustic”, and the Saint responds with a dative. Auxentius if anything shares Metaphrastes’ distaste for the vernacular, and that has made it likelier that the vernacular greeting preserved here is original to Vendemianus—and unlikelier that the vernacular greeting is an interpolation by a later scribe.

Μιᾷ δὲ τῶν ἡμερῶν πλήθους συνηγμένου τῶν ἐπιχωρίων τε καὶ πολιτικῶν καὶ τοῦ μακαρίου τὰ πρὸς σωτηρίαν διαλεγομένου, ἐπιστὰς ἄγροικος χαιρετίζων ἔλεγεν αὐτῷ· Καλή σου ἡμέρα, δέσποτά μου· τοῦ δὲ σεμνὸν ἐπιγελάσαντος, καὶ γὰρ ἦν τῷ ὄντι χαρίεις καὶ, Καλός σοι καιρὸς εἰρηκότος, οἱ παρόντες εἰς ἑαυτοὺς προσεμειδίων. ὁ δέ φησι πρὸς αὐτούς· Γραφικός ἐστι λόγος, ἀδελφοὶ, «Μὴ ἀποκρίνου ἄφρονι πρὸς τὴν ἀφροσύνην αὐτοῦ·» εἰ ἦν εἰρηκὼς, Εὐλόγησον, ἤκουσεν ἄν, ὁ Κύριος εὐλογήσει σε κατὰ τὸ συνῆθες ἡμῖν· αὐτὸς γάρ ἐστιν ὁ τὴν τιμὴν διὰ τῶν οἰκείων αὐτοῦ δούλων εἰς ἑαυτὸν ἀποδεχόμενος, καὶ ἑκάστῳ τὴν ἀμοιβὴν τῆς μισθαποδοσίας τῇ αὐτοῦ χάριτι ἀποδιδούς. Ἐπειδὴ δὲ τὴν καλήν μου ἡμέραν ἀπένειμεν, τὸν καλὸν καιρὸν ἀντέλαβεν. Οὕτως διορθωσάμενος, οὐ μόνον τὸν διαπταίσαντα, ἀλλὰ καὶ τοὺς λοιποὺς ἅπαντας, ἔνδον εἰσελθὼν πάλιν ἡσύχασεν.
One day, when a crowd had assembled of locals and citizens, and the blessed man was discoursing on matters of salvation, a rustic standing nearby greeting him saying “Good day to you (GEN), my lord”; and he solemnly laughed—for he truly was witty—and said: “A good season to you (DAT)”; whereupon those present smiled at him. But he said to them, “It is said in Scripture, brethren: ‘Answer not a fool according to his folly. (Prov 26:4)’ Had he said ‘Bless me,’ he would have heard back ‘The Lord will bless you’ as is our custom; for it is He who accepts honour from his servants, and grants each the reward of their desserts through his grace. But since he granted me a good day, he got a good season in return.” Correcting thus not just the person who had offended him, but all the others, he went back inside to meditate. (MPG 114 1429A.)

The story is retold by Michael Psellus, a century after Symeon Metaphrastes—with datives from both parties. But Psellus’ version makes it more explicit that those standing by found the rustic’s greeting amusing because it was colloquial:

Ἀγροῖκος γοῦν τις αὐτῷ συγγενόμενος, ὁμοῦ τε πρῶτον εἶδε καὶ ‘καλή σοι’ φησὶν ‘ἡμέρα, πάτερ Αὐξέντιε,’ τοῦτο δὴ τὸ ἐπιχώριον καὶ τοῖς πολλοῖς σύνηθες· ὁ δὲ μείλιχόν τι καὶ ἡδὺ μειδιάσας καὶ ὥσπερ εἰώθει χαριεντισάμενος, ‘καλός σοι καιρός,’ τῷ προσφθεγξαμένῳ ἀντείρηκε. τοῦ δὲ πλήθους διαχυθέντος, ‘δίκαιόν ἐστιν’ ἔφησεν ὁ πατὴρ ‘κατὰ τὴν ὁμοίαν λέξιν τοῖς διαλεγομένοις συνδιαλέγεσθαι, ἵνα καὶ μᾶλλον ὠφελοῖντο, τὰ φίλα καὶ συνήθη δεχόμενοι· […] πλὴν οὐδ’ ἐν ταῖς ἀγροικικαῖς ταύταις προσφωνήσεσί τε καὶ ἀντιφωνήσεσι τοὺς κρείττονας ἡμῶν τὸ ἔλαττον ἔχοντας ἀπηλλάχθαι δεῖ, ἀλλὰ νικᾶν τῷ πλείονι, ὥσπερ καὶ τοῦ προσειρηκότος τὴν ἡμέραν μοι δόντος αὐτὸς ἀντέδωκα τὸν καιρόν.
Now, a rustic happened to be with him, and when he saw him he said “Good day to you (DAT), father Auxentius.” This is indeed the local expression usual among the many. And he smiled gently and sweetly, and, joking as was his custom, he responded to the person who had addressed him: “A good season to you (DAT)!” When the crowd had scattered, the father said: “It is fair to converse with people with the same words that they converse with you, so that they may benefit all the more, accepting what is familiar and usual to them. […] Yet we should not let those stronger than us get away with having less, in their rustic addresses and responses; we should rather defeat them with something more. So when the person who addressed me offered me a day, I responded with a season.” (Orationes hagiographicae, 1,c 466ff)

The expression was clearly colloquial enough for St Auxentius to do some passive–aggressive ridiculing of it (“Have a good day” is the folly of a fool?!); and even in De Cerimoniis, it is a popular greeting by the assembled masses rather than a courtly one. That of course is all the more proof that it was a widely known and familiar expression.

Whether that was enough to prevent it being re-expressed as *καλή σε μέρα… well, maybe. But I’d like something more persuasive than that.


The third possibility is that, just as Southern Greek stopped short of using genitives in its plural pronouns, Northern Greek stopped short of using accusatives in its nominal wishes. Southern Greek was able to use the accusative instead, because both the accusative and the genitive had been in use there previously for indirect objects. Likewise, Northern Greek was able to use the genitive instead, because both the accusative and the genitive had been in use there.

Why would Northern Greek have hesitated to say καλημέρα σε but not δώσε με το? In the case of verbal indirect objects, Ancient Greek had precedent for both accusative- and genitive-coded objects of verbs, so introducing either as an indirect object was nothing syntactically unprecedented:

  • ζηλῶ σε τοῦ νοῦ I.envy you.ACC the.GEN mind.GEN “I envy you for your prudence”
  • ταῦθ’ ὑπέμνησ’ ὑμᾶς this.ACC I.reminded you.PL.ACC “I have reminded you of this”

On the other hand, there was no syntactic precedent for a bare predicate noun, like “peace!” or “good day!”, to be accompanied by a pronoun as a direct object: *εἰρήνη σε “peace you!” *καλὴ ἡμέρα σε “good day you!” (Or, to make it more obvious: *”peace him!” *”good day him!”)

What there was syntactic precedent for was a noun being accompanied by a genitive pronoun, as a possessive: εἰρήνη σου “your peace”, καλὴ ἡμέρα σου “your good day”. When time came to do away with the dative in those expressions, then, both the accusative and the genitive were available as alternatives, in both North and South; both were in fact current. Normally in the North the accusative prevailed, and the resulting ditransitives did not cause anyone to blink (although apparently they are now avoided in Northern Greek.) But when it came to such wishes, the accusative would result in a kind of expression noone had seen before; and the genitive, more familiar in that kind of combination, was used instead.


I like that argument. But it too has a difficulty.

The list of wishes/threats taking a genitive pronoun I gave does not involve only nouns. It involves interjections as well: μπράβο, αλοίμονο. And the distinction between interjections and nouns is going to be porous anyway, in wishes and threats. The noun “health”, for example (Ancient ὑγιεία /hyɡiéːa/), has been re-antiquated in Standard Modern Greek as υγεία /iˈʝi.a/, and is distinct from the greeting γεια /ʝa/, which can only be thought of as an interjection. Even before it was re-antiquated, I suspect it was kept distinct; the Cretan dialect I was exposed to pronounces the noun as υγειά /iˈʝa/.

Now, nouns may not take direct objects in Greek; which would explain why Northern Greek avoided καλὴ ἡμέρα σε. But interjections do take accusative indirect objects in Mediaeval Greek; and they do take accusative direct objects in Modern Greek.

For interjections taking accusative direct objects, the most obvious example in Standard Modern Greek is νά “behold! there it is!”: νά την “there she is!” νά με “here I am!” A second example, which is even more concerning for my argument, is the noun-turned-interjection anaθema “damn”: ανάθεμά τον “damn him”.

Interjections taking genitive indirect objects in Southern Modern Greek are also recorded as taking accusative indirect objects, at least in Middle and Early Modern Greek. Humbert (p. 189) gives a single instance, in the Vita Euthymii (written between 920 and 935):

(The courtier Stylian Zaoutzes, in response to Euthymius I predicting that he would perish in front of Emperor Leo VI): οὐαὶ τὴν ἡμέραν ταύτην καὶ ὅτ’ ἂν ἐγὼ ἠβουλήθην ἐνθάδε ἐλθεῖν. ἰδοὺ γὰρ ὁ πατὴρ ἐφωδίασεν ἡμῖν τῶν εὐχῶν αὐτοῦ, καὶ ἀπερχόμεθα.
Woe [to] this day and [to] when I wanted to come here. For see, the Father has supplied us with his blessing, and we are departing. (§3.15)

Zaoutses’ outburst is reminiscent of both Modern ανάθεμα την ώρα και τη στιγμή που ήρθα “damn the hour and the moment when I came”, and εγώ φταίω που είπα να έρθω “it’s my fault for coming” (literally, “for saying I’d come”—which here means “considering, deciding to come”.) The example is not necessarily clear, because the accusative could be a temporal adjunct (“Woe, on this day”), and in fact is coordinated with a temporal clause (“Woe, on this day and when I wished to come here”—with the hypothetical particle ἂν stressing that he should have done no such thing.)

There is a much older instance of οὐαί taking an accusative object, however, in the Questions of Bartholomew (3rd century AD), and it is much more clear cut:

οὐαὶ τὸν ὀμνύοντα κατὰ τῆς κεφαλῆς τοῦ θεοῦ, οὐδὲ  ἐπιορκοῦντι κατ’ αὐτοῦ ἀληθῶς
Woe unto him who swears against the head of God; nor shall the true violator of an oath against him [be forgiven] (§5.5)

And there are several instances of the more modern αλί “woe” taking accusative objects:

καὶ ἀλὶ ἐμέν, <ἀλὶ ἐμέν,> ἂν μάθουν οἱ Κασίσοι
And woe unto me (ACC), woe unto me, if the Hashashin find out
(Digenes Escorial ms. 245; 15th century manuscript, 12th century text?)

καὶ σείσῃ τὸ κοντάριν του καὶ εἰπῇ τὸ ἀλί σ’ ἀλί σε
And he’ll shake his lance and say “woe to you (ACC), woe to you”
(Ptochoprodromos §4.490; 1170s?)

τρεῖς ἀλί τον καὶ ἀλὶ εἰς τὴν ψυχήν του
Thrice woe unto him (ACC) and woe to his soul (preposition “to”)
(Papa-Synodinos, Chronicle of Serres §3.28; 1630s, Serres in Greek Macedonia)

In fact, George Chortatzis writing in Crete around 1600 uses both accusative and genitive after αλί in the same play:

Πράμά ’ναι ἀδυνατότατο, καὶ μὴν τὸ βάλη ὁ νοῦς σου,
γιατ’ ἡ κερά μου τὸ γρικᾶ, κι’ ἀλὶ τοῦ ριζικοῦ σου!
It’s a most impossible matter, and you should not even consider it,
for my mistress will hear it, and woe then to your fate! (GEN) (Katzourbos I.3 322)

κι’ ὡς τό ’κουσε ἐτρόμαξε κι’ ἐπῆγε στὸ σκολειόν του,
κι’ ἂν ἔναι κι’ εὕρη τον ἐκεῖ, ἀλὶ τὸ ριζικόν του.
And when he heard it, he was frightened, and went to his school,
and if he finds him there, woe to his fate. (ACC) (Katzourbos IV 2.32)

αλοίμονό τον “woe to him (ACC)” is not possible in Standard Modern Greek, but there is fluidity in using the accusative even in 1600 Cretan—solidly genitive indirect object territory. Between “woe” + Accusative in Northern Greek plus Chortatzis, “behold!” + Accusative in Standard Greek, and “anathema!” (noun > interjection) + Accusative also in Standard Greek, the impossibility I’m claiming for Noun + Accusative Pronoun is looking less compelling.


… Maybe. Interjections are still different from nouns. In particular, interjections in Greek are much closer to verbs (and imperatives in particular) than nouns are, so them taking an accusative pronoun as a direct object is not as strange as it would be for a noun.

A minor piece of evidence for the proximity of interjections to imperatives is that in Standard Greek, clitic pronouns follow imperatives, participles, and interjections: νά τον “there he is!”, βρες τον “find him!”, βρίζοντάς τον “cursing him”. When the verb is finite, the clitic in standard Greek precedes it: τον βρήκα “I found him”, τον έβρισα “I cursed him”.

That’s a minor piece of evidence, because the criterion at work is whether the clitic host is finite or not. An imperative, a participle, and in older Greek an infinitive, are more noun-like, since they are not finite. If either a noun or an interjection ends up being a clitic host for pronouns, for whatever reason, of course it’s going to pattern with non-finite verbs instead of finite verbs. So the similarity of νά τον and βρες τον is trivial, and it does not establish that nouns and interjections are different about their likelihood of having pronoun arguments at all.

The more interesting piece of evidence, which Brian Joseph brought up decades ago, is that interjections in Greek occasionally take second person plural imperative endings—meaning that the interjection without the plural ending is already being thought of as an imperative. νά has the dialectal plural νά-τε; άντε “go on!” < Turkish haydi, which does not take an object at all, has the plural άντεστε.

If Greeks are predisposed to think of interjections as imperatives, then they are predisposed to attach direct objects to them; and regarding interjections as imperatives is prior to attaching direct objects to them (as with άντε, which doesn’t take objects at all). So seeing an accusative after an interjection should not come as a surprise, whether the accusative is for a direct object, as in “behold! him”, or for an indirect object, as in “woe to him!” It is also not a surprise for a noun that has come to be seen as an interjection, such as “anathema”.

But in earlier stages of the language—such as the stage represented by St Auxentius of Bithynia—”good day” was still understood not as an interjection but as a noun phrase, to the extent of St Auxentius ridiculing the “rustic” for “offering him a good day”, and offering a good season in return. So even if the Questions of Bartholomew, a century before St Auxentius, were comfortable having the interjection οὐαί “woe” followed by an accusative (“woe him”), a phrase like *καλή σε ἡμέρα “good day thee” would still have sounded more odd than καλή σου ἡμέρα “good day thine”, as the vernacular counterpart to καλή σοι ἡμέρα “good day to thee”. And if that discomfort with direct objects of nouns was still in place by the 10th century, then it would explain why Northern Greek never came up with καλή σε ἡμέρα, even as it was saying ἀλὶ ἐμέν “woe me!”

That’s my thinking to date; I’m interested to hear from anyone who’s already tackled this issue. (For all I know, it’s already been resolved somewhere.)

One Comment

Leave a Reply

  • Subscribe to Blog via Email

    Join 299 other subscribers

  • January 2018
    M T W T F S S
    « Dec    
    1. Humbert, Jean. 1930. La disparition du datif en grec du Ier au Xe siècle. Collection de la Société de linguistique de Paris, t. XXXIII. Paris, Champion.

%d bloggers like this: