Recently, I have been working through David’s Armstrong’s (1981) article, “The Ancient Greek Aorist as the Aspect of Countable Action.” It’s an excellent piece of research and I have greatly enjoyed comparing his analysis of Homeric and Classical Greek to the Koine. The patterns that he observes play out consistently across both Classical and post-Classical Greek. Perfective aspect verbs (aorist) use cardinal count adverbs to modify predicates. Imperfective aspect verbs (present/imperfect) use frequency count adverbs to modify predicates. The pattern is simple and satisfyingly consistent.

This is a small complication. Across several hundred examples of quantifiers that consistently follow the pattern, I did discover one phrase that does not seem to adhere quite as well, at least at first glance. Armstrong notes that the phrase δὶς καὶ τρὶς is used with imperfective verbs. While there are cardinal numbers, taken together they idiomatically express the meaning “again and again” or repeatedly and thus function to convey frequency count rather than cardinal count (Armstrong notes this idiom appears in the 5th century and does not appear in Homer).

Things get more confusing in the post-Classical period, where we find the phrase ἅπαξ καὶ δίς used a number of times. In the texts I have available for searching, which include all of Perseus and then all the texts Logos has produced (Josephus, Philo, the LXX, Greek Pseudepigrapha, the NT, and the Apostolic Fathers), this new phrase only appears seven times: four times in the LXX, twice in the NT, and once in the 1 Clement (an LXX quote).

For the most part, it appears to have roughly the same meaning as the Classical δὶς καὶ τρίς. Thus in Deut 9:13 (Quoted in 1 Clem 53.4), we find:

1. καὶ εἶπεν κύριος πρός με Λελάληκα πρὸς σὲ ἅπαξ καὶ δὶς λέγων Ἑώρακα τὸν λαὸν τοῦτον, καὶ ἰδοὺ λαὸς σκληροτράχηλός ἐστιν
   And the Lord said to me, “I have spoken to you again and again, saying, “Look at this people and see that they are a stiff-necked people” (Deut 9:13 / 1 Clem 53.4)

Other examples are more complicated, however. There is an instance in 1 Thess 2:18, involving the ellipsis of the verb:

2. διότι ἠθελήσαμεν ἐλθεῖν πρὸς ὑμᾶς, ἐγὼ μὲν Παῦλος καὶ ἅπαξ καὶ δίς
   For we wanted to come to you—I, Paul, again and again (1 Thess 2:18).

But here we have an aorist ἠθελήσαμεν, which shouldn’t happen with this kind of frequency count. This example is relatively easily explained able if we accept that θέλω isn’t the missing verb and translated the clause as:

3. διότι ἠθελήσαμεν ἐλθεῖν πρὸς ὑμᾶς, ἐγὼ μὲν Παῦλος καὶ ἅπαξ καὶ δίς
   For we wanted to come to you—I, Paul, [tried] again and again (1 Thess 2:18).

This is acceptable since there’s really no rule that an elided verb needs to maintain the aspect of the previous clause. But what about Philippians 4:16? Things aren’t so simple there.

4. ὅτι καὶ ἐν Θεσσαλονίκῃ καὶ ἅπαξ καὶ δὶς εἰς τὴν χρείαν μοι ἐπέμψατε.
   Because even in Thessalonica you sent for my needs once and twice [again and again?] (Phil 4:16).

Here we have the phrase appearing with an aorist, and thus perfective, verb. Armstrong says this shouldn’t happen. But then, perhaps this isn’t frequency count. Perhaps it’s cardinal count and is actually referring to a specific number of times that the Philippian church helped Paul.

If this is true, then we have an idiom that isn’t an idiom here—something that no translation agrees with. An instance of the phrase in 2 Esdras 23:20 might support that view:

5. καὶ ηὐλίσθησαν πάντες καὶ ἐποίησαν πρᾶσιν ἔξω Ιερουσαλημ ἅπαξ καὶ δίς.
   And they all spent the night and did [their] selling outside Jerusalem once and twice (2 Esdras 23:20).

Here again we have an occurrence of the phrase with a perfective verb (ἐποίησαν) as in Philippians. This clause is directly followed by Nehemiah’s words:

6. And I [Nehemiah] warned them and said to them, “Why do you spend the night in front of the wall? If you do it again, I will stretch my hand against you.” From that time onward, they did not come on the Sabbath.

So here we see that at least in this instance, the example actually appears to refer to the countability of the event. Nehemiah is saying: You’ve done it twice already. Do not do it again. That seems to make sense. And the Hebrew supports the view.

7. פַּעַם וּשְׁתָּיִם
   once or twice

And that’s precisely how Van der Merwe translates the phrase in his grammar (Van der Merwe, Naudé and Kroeze 1999, 269). And this got me thinking. The phrase only occurs seven times in the texts I have access to (which is probably close to 15 million words) and none before the Septuagint. Is this a Hebrew idiom? No, it does not appear to be. For one, both the Greek and Hebrew are much more transparent in their meaning to its constituent parts than the occurrences of the phrase that mean “again and again”. This is the opposite of what we expect from an idiomatic expression. Additionally, Nehemiah 13:20 (the Hebrew equivalent of 2 Esdras 23:20) is the only time we see the words in the MT. Everywhere else, there is no documented source text.

This phrase appears to be able to function to express both indefinite frequency (again and again) event quantification, and also in other circumstances looks like cardinal count. It does not look like an instance of translation Greek, but also does not appear in Greek literature before the Septuagint. The closest we get is the similar phrase, “ἅπαξ ἢ δὶς” which is pretty clearly a cardinal count expression that means “a couple times”, not frequency count. It occurs once in Philo and also five times in Perseus corpus of Greek texts. The following example from Plato on how to deal with stubborn hiccups is representative.

8. ειʼ δʼ ἄρα πάνυ ἰσχυρά ἐστιν, ἀναλαβών τι τοιοῦτον οἵῳ κινήσαις ἂν τὴν ῥῖνα, πτάρε· καὶ ἐὰν τοῦτο ποιήσῃς ἅπαξ ἢ δίς, καὶ ειʼ πάνυ ἰσχυρά ἐστι, παύσεται.
   If, however, it is a very stubborn one [hiccup], take something that will tickle your nostrils, and sneeze: do this once or twice, and though it be one of the most stubborn, it will stop (Plato, Symposium 185e).

My best analysis, as it stands, is that the phrase ἅπαξ καὶ δίς has a idiomatic reading as a single construction with a meaning that is more than the sum of its parts (indefinite frequency) and also a non-idiomatic reading. The difference between the readings is decided by the aspect of the clause. On this view, Philippians 4:16 and 2 Esdras 23:20, with their perfective verbs, should be treated as cardinal count, while the other examples with imperfective aspect verbs are indefinite frequency count expressions. In both cases, the aspect chosen functions as part of the event quantification construction and contributes to the meaning of the whole.

Works cited:

Armstrong, David. 1981. “The Ancient Greek Aorist as the Aspect of Countable Action.” Pages 1-12 in Syntax and semantics, vol. 14: Tense and aspect, ed. P. Tedeschi and A. Zaenen. New York: Academic.

Van der Merwe, Christo, Jackie Naudé, and Jan Kroeze. 1999. A Biblical Hebrew Reference Grammar. Sheffield: Sheffield Academic Press.