There are two distinct linguistic issues in the latter part of Matthew 28:17, which reads this way (NA28 = SBL GNT = UBS4) -

καὶ ἰδόντες αὐτὸν προσεκύνησαν,a οἱ δὲ ἐδίστασαν.
kai idontes auton prosekunēsan, hoi de edistasan.
and seeing him they worshiped, ?? ?? ??.

a TR + αὐτῷ, cf. 28:9.

One issue, relating to the semantics of διστάζω, already has a question on BH.SE.

My question has to do with the syntax of οἱ δὲ ἐδίστασαν. The vast majority of English translations render this by something like:

...but some doubted.

Another possible translation seems to be:

...but they doubted.

The former suggests some few of the Eleven harboured doubts, the latter implies that they all doubted. I have not been able to identify any "public" translations that go this way, however.

I'm interested primarily in a syntactic evaluation of the verse. Is the overwhelming preference for the "partitive" sense ("some of") justified?

If, tangentially, BH.SE participants are familiar with any "public" translations in any language that give an "inclusive" translation, that would be worth knowing.

2 Answers 2


Partitive or Switched Subject is Nearly Certain as Correct

K. Grayston makes an argument for the inclusive view,1 but is challenged by both K. L. McKay's brief reply,2 and P.W. van der Horst's more lengthy reply,3 both upholding a partitive view.

Grayston argues the inclusive view largely upon two points.

First, the inclusive is the case in the primary uses of the hoi de construction in "conversational exchange" passages (105-106). The only two non-conversation passages are Mt 26:67 and here in Mt 28:17. While he recognizes there are partitive uses of the construction, he believes those are because the context renders the partition. He gives two examples from Xenophon's Hell. 1.2.14 and Cyr. 3.212 (106). But he states of both the Matthew passages, "what reason is there for discovering separate groups?" (106).

Second, he believes it is the logical or theological problem of doubting worshipers that moves people to view it as partitive (106-107). He resolves this postulating "what the Eleven were doubtful about" (107), and ultimately concludes:

The Eleven found themselves confronted by the Lord who appeared to have returned from the world beyond death. If it was indeed the Lord, they were fearful of being condemned for their desertion (26.56), and they instantly performed proskynèsis—as the fearful women had done. They doubted whether even total submission would save them. The Risen Jesus increased their apprehension by announcing that all authority had been given to him in heaven and on earth. And then he met their fear by bestowing a commission and asserting his presence till the end of the age.

Hence any translation needs to imply: 'When they saw him, they threw themselves down in submission, though they doubted its effect' (108).

"Effect" meaning to save them from "being condemned for their desertion," as Grayston noted in the earlier part of the quote above.

Before I note McKay's and Van der Horst's critiques, let me note one critique of my own--the disciples had already had a few previous encounters with the resurrected Jesus by this time, so why would they be fearful now of Him condemning them for having deserted Him when He was captured and went to the cross? The logic does not hold.

McKay gives a brief critique, purely based on grammatical usage. Primarily that

the normal use of this phrase [hoi de] involves a change of subject [emphasis added by me] ... when there is no change of subject there is normally no pronoun (71).

McKay refers to the same two examples of Xenophon that Grayston noted. It is this normal usage of the article as a pronoun in this construction indicating a change of subject that argues for the partitive sense, McKay stating:

It is quite likely, in grammatical terms, that hoi de in Mt. 28.17 were, like the wounded in Cyr. 3.2.12, a minority of the main group mentioned specifically, here the eleven; but in view of the distance of the whole group subject from the main verbs it could be that hoi de were a minority of a larger group led by the eleven and did not necessarily contain any of the eleven themselves (71).

And finally concludes:

the proposition that all both worshipped [sic] and doubted is untenable without real evidence that hoi de was ever used without indicating some change of subject (71-72).

Van der Horst agrees with McKay, but specifically adds clarity in two ways.

First, the "change of subject" from the construction "may be complete or partial" (27). He asserts:

Let it be stated clearly that it is a well-known and frequently used syntactical device to indicate a division of a group of persons or things into two (or more) subgroups only in the second half of the sentence (oi μέν / τους μέν being omitted in the first half) (28).

Second, his main critique of both was that they only referred to the same two examples of the partitive use from Xenophon, when many other examples exist based on what he "stated clearly" above (27-28). He broadly gives the following:

Instances of this [partitive] usage can easily be found, e.g. in the grammar of Kühner-Gerth, in Denniston's classic work on the particles, or in Liddell and Scott (28).

He then gives some examples, of which I will give his references here from page 28, but link to online versions and translations as possible:

  1. Euripides, Hercules furens 635-636 ["χρήμασιν δὲ διάφοροι: ἔχουσιν, οἳ δ᾽ οὔ"] (trans. "it is wealth that makes distinctions among them; some have, others want")
  2. Inscriptiones Graecae 2.2, 652 A 45 [note: reference matching the words he quotes I found in [IG II² 1389 line 6 "δύο σφραγῖδε λιθίνω χρυσο̑ν ἔ]χοσα τ[ὸν δακτύλιον, ἡ δ’ ἑτέρα ἀργυρο̑ν"] (You will have to navigate to the location yourself within that link; I have not found an online translation, but Van der Horst gives "two stone seals, one having a golden ring, the other a silver one").
  3. Plato, Leges 828b-c ["ἀγῶνας μουσικούς, τοὺς δὲ γυμνικούς"] (trans. "musical contests, and also gymnastic contests")
  4. Andocides, De mysterüs 38 ["τοὺς δὲ ἀνὰ εἴκοσιν"] (trans. "some cases, twenty") and 105 ["εἰ αὐτοῖς ἐξέσται ἀδεῶς συκοφαντεῖν καὶ γράφεσθαι, τοὺς δὲ ἐνδεικνύναι, τοὺς δὲ ἀπάγειν"] (trans. "to find out whether they will be given complete licence to fill their pockets by indictments,or informations, maybe, or arrests")
  5. Lucían, Lexiphanes 2 ["κατέλαβον γὰρ τοὺς ἐργάτας λιγυρίζοντας τὴν θερινὴν ᾠδήν, τοὺς δὲ τάφον τῷ ἐμῷ πατρὶ κατασκευάζοντας"] (trans. "for I found the yokels caroling the harvest-home; some of them, too, were preparing a grave for my father")

He notes as well there are also "several other examples, from Homer to post-classical times" of the partitive use. Another biblical example he gives is Act 17:18, "where it is clear that there were different reactions to Paul's preaching" (29); he also notes there that Act 2:12-13 is a related example, but not the exact same construction:

The words έτεροι δέ after πάντες in ν. 12 indicate that the meaning is that all reacted to the phenomenon described in vv. 6-11, but that the reactions were divided: some were amazed and perplexed, others however mocked at the disciples (classical usage could have had oi δέ here) (29).

Van der Horst concludes:

All this implies that in Mt. 28.17 oi δέ cannot mean all of the disciples, can mean (from a strictly grammatical point of view) other persons than the disciples, but, since no other persons are involved here at all, must be part [emphasis added] of the disciples (29).

Conclusion on Partitive Concept

It does seem the evidence argues very heavily for the partitive idea, and at the very least demonstrates both the precedent to use the partitive, as well as why all the translations have taken that direction. None of the three noted above give an example of a Bible version that uses "they" in an inclusive sense other than a few individuals in their commentaries (such as Grayston in his article, and a few others he notes [108-109], which Van der Horst affirms [27 n.3]).

Van der Horst's conclusion may be correct that it is "part" of the disciples. If so, I think the most likely explanation of what they doubted was whether they were going to see Jesus in Galilee or not (recall that such a message was not expressed to them directly by Jesus, but by the women who saw him early on the first day, v.5-10).4

However, Van der Horst's conclusion is not without being challenged when he asserts "no other persons are involved here at all," which was his main reason for rejecting the idea that it refers to "other persons than the disciples."

Possible that None of the Eleven Doubted

To review, the grammatical construction uses the third person plural article οἱ as a pronoun, with δὲ and the finite verb ἐδίστασαν. Daniel Wallace notes this particular construction, which essentially relays what McKay and Van der Horst have noted in their articles:

The article is often used in the place of a third person personal pronoun in the nominative case. It is only used this way with the μὲν… δέ construction or with δέ alone. (Thus, ὁ μὲν… ὁ δέ or simply ὁ δέ.) These constructions occur frequently in the Gospels and Acts, almost never elsewhere. ...

The δέ is used to indicate that the subject has changed; the article is used to refer back to someone prior to the last-named subject. Most frequently, the subjects are speakers and the interchange is one of words, not action. ...

Typically, the ὁ δέ (or ὁ μέν) construction is immediately followed by a finite verb or circumstantial participle.5

Wallace did not give Mt 28:17 as an example, but it does match the parameters he gives exactly, and clearly from the previous articles noted, it falls within the same family of usages.

The typical view is that it is then partitive of the Eleven themselves. Now most of the clear examples Van der Horst referenced either (1) clearly indicated two or more groups being distinguished previously, often in the first part of the sentence, with the construction referencing just one of the others, or (2) in one case (Euripides, Hercules furens 635-636), it affirmed both a positive and a negative of the single previous group noted, which logically must mean part had and part had not.

This clarity is missing in the Mt 28:17 passage if the reference is to just the disciples, who are a singular group. Nevertheless, a couple of Van der Horst's examples did split a singular group, but in such cases it appears to always be an either/or split to the groups. That is (I am only basing this off of Van der Horst's given examples in the article), if the Elven are the split group, then it appears that a proper translation would have to be "some of them worshiped, but some doubted."

Yet recall that Van der Horst noted the partitive could be a whole other group with the construction, but he seems to indicate that for such to be so, another group would need to be named in the context. A group for οἱ to point back to, which as Wallace noted would be to "someone prior to the last-named subject."

Contra Van der Horst, there is in fact such a group in the immediate vicinity within the text of Matthew's narrative itself: the "Jews" (v.15) among whom the reports were circulating that the disciples stole the body (v.13). In fact, the purposeful insertion of v.11-15 of this story about what happened to Jesus' body and its effects "among the Jews" falls between the command to go to Galilee (v.10), with specific reference to the timing of this story's circulation related to "while they were going" (v.11a), and the narrative of the Eleven's arrival in Galilee to see Him (v.16-17). This deliberate splitting up of the command from the performance of going to Galilee by this particular story appears to purposefully set up another group to be referenced in opposition to (in an either/or relation to) answering what happened to the body of Jesus.

So "they doubted" ("doubted" means hesitancy, not out-right unbelief) can be argued as referring back to "the Jews" receiving the soldiers' reports, doubting that the disciples saw and worshiped Jesus in Galilee.

Such an understanding resolves the fact that it is unlikely the Eleven had any doubts (even hesitancy) at all at this point with respect to Jesus being resurrected. At the time Thomas sees Christ, the last of the Eleven to see Christ resurrected (Jn 20:28), it was at least the second time the other disciples saw Him (Jn 20:19-23). These appearances were before they go into Galilee (Jn 21:1), and thus before the events of Mt 28:17.


The partitive idea or preceding referenct is almost certainly the correct reading—whether that is parting the Eleven themselves or referring to another group all together, which if the latter would most likely by context be the Jews noted just prior. The construction would be hard to argue as being inclusive of the previously mentioned group based off the historical usage evidence.


1 K. Grayston, "The Translation of Matthew 28.17," Journal for the Study of the New Testament [JSNT] 21 (1984):105-109.

2 K.L. McKay, "The Use of hoi de in Matthew 28.17: A Response to K. Grayston," JSNT 24 (1985):71-72.

3 P.W. van der Horst, "Once More: The Translation of oi δέ in Matthew 28.17," JSNT 27 (1986):27-30.

4 It is also possible the disciples doubted whether they would see Him in Galilee or not, as it may be that "My brethren" Jesus refers to in v.10 are actually His half-brothers (Mary and Joseph's children), not a command for the disciples. He had already previously told the disciples He would meet them in Galilee after His resurrection (Mt 26:32), so perhaps they were hoping that Jesus' call to His half-brothers to meet Him in Galilee would also include them, but some may have doubted they would see Him.

5 Daniel Wallace, Greek Grammar Beyond the Basics - Exegetical Syntax of the New Testament (Zondervan Publishing House and Galaxie Software, 1999), 211.


I will limit my comments to the question is the “inclusive” reading of οἱ δὲ grammatically impossible rather than merely improbable which is the majority view:

Stephanie Black objects to Grayston's approach, observing that οἱ δὲ signals discontinuity and would be highly unlikely if there were continuity of subject with the previous sentence. (Stephanie Black, Sentence Conjunctions in the Gospel of Matthew: kai, de, tote, gar, oun and Asyndeton in Narrative Discourse (Sheffield Academic Press, 2002), pp. 160-162.)

Stanley E. Porter claims that there are three viable readings; partitive which is the most satisfactory, switched subject and the inclusive which is the least satisfactory. (Idioms of the Greek New Testament (Sheffield Academic Press, 1994) 2nd ed., p. 113). Note that both S. Porter and his former student Stephanie Black do not simply rule out the inclusive reading.

With regard to Attic Syntax, Guy Cooper says that the inclusive reading of οἱ δὲ is very rare. He gives three examples Sophocles Electra 711 , Thucyd., Xenophon Anabasis Guy Cooper, Attic Greek Syntax, vol. 1, 50.1.4, pp. 355-56.

I will only show Sophocles Electra 711 since I am familiar with that text.

Soph. Electra1

Κεῖνος γὰρ ἄλλης ἡμέρας, ὅθ' ἱππικῶν
ἦν ἡλίου τέλλοντος ὠκύπους ἀγών,
εἰσῆλθε πολλῶν ἁρματηλατῶν μέτα. [700]
Εἷς ἦν Ἀχαιός, εἷς ἀπὸ Σπάρτης, δύο
Λίβυες, ζυγωτῶν ἁρμάτων ἐπιστάται·
κἀκεῖνος ἐν τούτοισι θεσσαλὰς ἔχων
ἵππους, ὁ πέμπτος· ἕκτος ἐξ Αἰτωλίας
ξανθαῖσι πώλοις· ἕβδομος Μάγνης ἀνήρ· [705]
ὁ δ' ὄγδοος λεύκιππος, Αἰνιὰν γένος·
ἔνατος Ἀθηνῶν τῶν θεοδμήτων ἄπο·
Βοιωτὸς ἄλλος, δέκατον ἐκπληρῶν ὄχον.
Στάντες δ' ὅθ' αὐτοὺς οἱ τεταγμένοι βραβῆς
κλήροις ἔπηλαν καὶ κατέστησαν δίφρους, [710]
χαλκῆς ὑπαὶ σάλπιγγος ᾖξαν· οἱ δ' ἅμα
ἵπποις ὁμοκλήσαντες ἡνίας χεροῖν
ἔσεισαν· ἐν δὲ πᾶς ἐμεστώθη δρόμος
κτύπου κροτητῶν ἁρμάτων·

For on another day, when at sunrise there was the speedy contest of the chariot horses, he entered the lists with many charioteers. One was an Achaean, one from Sparta, two were Libyans, masters of yoked cars, another among them had Thessalian mares, the fifth; the sixth came from Aetolia, with chestnut colts; the seventh was Magnesian; the eighth had white horses, an Aenian; the ninth came from Athens, built by gods; another was Boeotian, filling the tenth chariot. They took their stand where the appointed judges had sorted them with lots and placed their chariots, and at the sound of the brazen trumpet darted off. Shouting to their horses, the drivers gripped the reins and shook them loose; the whole course resounded with the clash of rattling chariots;

Translation: Hugh Lloyd-Jones LCL Harvard 1994

1. Greek text and translation also available in the older Loeb edition.

Note that οἱ δ' line 711 refers to the charioteers πολλῶν ἁρματηλατῶν line 701:

Shouting to their horses, the drivers gripped the reins …

This example isn't simple. οἱ δ' refers to a group which has been described at considerable length. Never the less οἱ δ' is not used in this context to mark switched reference or discontinuity and the referent of οἱ δ' is inclusive of the charioteers, not partitive.


It's dangerous to cite grammarians without doing the exegesis! I will not try to read Guy Cooper's mind, I have a hard enough time just understanding him. Never the less I will attempt to explain what I see going on in Soph. Electra 710-713.

Στάντες δ' ὅθ' αὐτοὺς οἱ τεταγμένοι βραβῆς
κλήροις ἔπηλαν καὶ κατέστησαν δίφρους,
χαλκῆς ὑπαὶ σάλπιγγος ᾖξαν· οἱ δ' ἅμα
ἵπποις ὁμοκλήσαντες ἡνίας χεροῖν

Following J.H. Kells and Hugh Lloyd-Jones the charioteers are the subject of Στάντες and ᾖξαν. ὅθ' “where …” introduces background material about the judges and casting lots. The judges appear to be the subject of κατέστησαν which must mean something like assign a position not physically put some place. Kells: "Having taken up position (in the places) where the appointed umpires, drawing lots for them, stationed their chariots …" The background material ὅθ' αὐτοὺς οἱ τεταγμένοι βραβῆς κλήροις ἔπηλαν καὶ κατέστησαν δίφρους serves as a complex adverbial modifier for Στάντες the subject being the charioteers. In other words the judges assigning positions by lot is a side comment and the verbs with the charioteers as subject are mainline (foreground) in the narrative.

The main story line resumes χαλκῆς ὑπαὶ σάλπιγγος ᾖξαν where the charioteers are the subject. Thus we have continuity of subject in the next sentence οἱ δ' ἅμα ἵπποις ὁμοκλήσαντες ἡνίας χεροῖν ἔσεισαν·

Note that Gildersleeve also cites this text

519 ὃ δέ same as subject of preceding sentence:

ἐνταῦθ' ἔμενον ὡς τὸ ἄκρον κατέχοντες. οἳ δ' οὐ κατεῖχον, ἀλλὰ μαστὸς ἦν ὑπὲρ αὐτῶν, XEN. An. 4.2.5-6; There they remained fancying that they were in possession of the summit. Not they, but there was a knoll above them.

THUC.1.87.1-2: τοιαῦτα δὲ λέξας ἐπεψήφιζεν αὐτὸς ἔφορος ὤν. ὃ δέ, κρίνουσι γὰρ βοῇ καὶ οὐ ψήφῳ, οὐκ ἔφη διαγιγνώσκειν τὴν βοὴν ὁποτέρα μείζων.

HDT.1.17 HDT., 1.66 HDT., 1.107 HDT., 1.171 HDT., 1.196 HDT., 3.126 HDT., 5.35 HDT., 5.120 HDT., 6.3 HDT., 6.9: εἰ δὲ ταῦτα μὲν οὐ ποιήσουσι, οἳ δὲ πάντως διὰ μάχης ἐλεύσονται, τάδε ἤδη σφι λέγετε. 133. 7.13. 125. 163· Γέλων δὲ . . . ταύτην μὲν τὴν ὁδὸν ἠμέλησε, ὃ δὲ ἄλλης εἴχετο. 218. 8.40. 9.52. 108.

SOPH. El. 711-3: χαλκῆς ὑπαὶ σάλπιγγος ᾖξαν· οἳ δ' ἅμα | ἵπποις ὁμοκλήσαντες ἡνίας χεροῖν | ἔσεισαν.

HOM. Od. 9.398-9: τὸν μὲν (sc. μοχλόν) ἔπειτ' ἔρριψεν ἀπὸ ἕο . . . | αὐτὰρ ὃ Κύκλωπας μεγάλ' ἤπυεν. 13.219· τῶν μὲν ἄρ' οὔ τι πόθει· ὃ δ' ὀδύρετο πατρίδα γαῖαν.

Il. 1.189-91: διάνδιχα μερμήριξεν, | ἢ ὅ γε . . . | τοὺς μὲν ἀναστήσειεν, ὃ δ' Ἀτρεΐδην ἐναρίζοι. 4.491. 5.148· τοὺς μὲν ἔασ', ὃδ' Ἄβαντα μετῴχετο. 8.119. 126. 302. 11.80. 148. 426. 13.518. 15.127· ἣ δ (έ). 136. 16.467. 20.322. 21.115. 171.

Basil Lanneau Gildersleeve [1900], Syntax of Classical Greek, para. 519.

  • Thanks for these references + example. I find Cooper's account of this syntax fascinating: "In the very few places where the substantive article with δέ merely reiterates the subject there is strong hypercharacterisation... This is an effect which Attic authors picked up from Homer and Herodotus (cf. In those authors it is so common that it is not always felt as strongly as it is in the Attic imitations." (Greek Syntax, vol. 1, p. 355). For the banter, Smyth, para. 1112 is less full, and cautious, and without examples of "same subject".
    – Dɑvïd
    May 1, 2015 at 9:17
  • I disagree that it is inclusive and that it "is not used in this context to mark switched reference or discontinuity." The subject of the preceding sentence is "οἱ τεταγμένοι βραβῆς" ("the appointed judges"), so the οἱ δ' is switching back away from that subject to the prior subject of the charioteers. Thus, rather than being an inclusive statement (of either all the judges, or the judges and the charioteers), this is a case of Porter's "switched subject" use, back to "someone prior to the last-named subject" (as I quoted from Wallace). Cooper was wrong to include it as an example.
    – ScottS
    May 1, 2015 at 12:47
  • @Davïd: As a side note, the example given here from Soph. Electra actually matches the usage upon which I based my theory that it points to the Jews doubting, as it is an example of a switch back to the prior subject group.
    – ScottS
    May 1, 2015 at 12:53
  • Upon examination, Cooper also misidentifies the other examples, as both switch to prior subjects. Thucydides 1.87 does not refer to Σθενελαΐδας (Sthenelaidas, 1.85) the speaker of 1.86, but to Ἀρχίδαμος (Archidamus, 1.85) the king (1.79), who spoke in 1.80-85, part of which was a call to decide calmly ("βουλεύσωμεν, ἀλλὰ καθ᾽ ἡσυχίαν") about war. So the king is referred back to in 1.87.1, who goes on to declare a count of those for or against war over the uproar from Sthenelaidas's speech (1.87.2).
    – ScottS
    May 1, 2015 at 15:02
  • Then Xenophon 4.2.6 does not refer back to "οἱ δ᾽ ἔχοντες τὸν ἡγεμόνα" ("Meanwhile the party with the guide") of 2.5, but the volunteer forces to go before them (noted about taking the height in 2.1 and set off in 2.2). The οἱ δ᾽ ἔχοντες τὸν ἡγεμόνα group supposed their side controlled the height, but the other group of the volunteers (οἱ δ᾽) had not taken the height.
    – ScottS
    May 1, 2015 at 15:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.