A catalogue of citations of scripture in the canonical Christian gospels, by way of a note to myself. Some caveats, terms, and conditions:
- The gospels don’t quote the Hebrew Bible. First, because they’re quoting Greek texts. They don’t exactly quote the Septuagint either, because there were multiple ‘Septuagint’ recensions. But we can call it ‘Septuagint’ so long as we’re clear that it’s a Septuagint and not the Septuagint.
- Second, because there was no Bible at the time. Just lots of texts, and of them, some were considered sacred or sacred-ish. But no canon. That’s why New Testament authors are happy to quote texts like Enoch, 1 Maccabees, and the Psalms of Solomon, some of which are omitted in some modern canons.
- Only citations, not echoes. It isn’t possible to enumerate or quantify all kinds of echoes, so I’m only listing places where the texts directly quote older texts, or explicitly state that they’re citing an older text.
An illustration, to show why we can’t catalogue or count inexplicit echoes. In Luke 2.22–24, Joseph and Mary present Jesus at the Temple. The narrator makes it sound like this is a well established unified custom, but it’s nothing of the kind. In fact it’s a mash-up. The author grabs three unrelated scriptural passages, and presses the purée button. The sources are: Hannah dedicating Samuel (1 Samuel 1.22–28); the offering of two turtle-doves as part of a ritual purification (Leviticus 12.2–8); and the consecration of first-born animals and children (Exodus 13.11–14). Is this one citation or three citations? None, of course: the passage doesn’t mention any of these three passages, it just echoes them.
That’s why we’re not counting echoes. The list below is just the explicit citations.
Key:
- LXX : Septuagint
- Bold type : passage with direct quotation or explicit citation
- Q : quotation, either verbatim or approximate
- QM : misquotation with major divergence from source
- C : citation of named source
- CM : misattributed citation — a source is named, but incorrectly
- G : generic citation of unnamed ‘prophet’ or ‘writings’
- small italics : no LXX citation/quotation, but parallel to another gospel which does have a citation
- MT : Masoretic Text (Hebrew Bible), where different from LXX
| LXX passage cited | Matthew | Mark | Luke | John |
|---|---|---|---|---|
| Isaiah 7.14 | 1.22–23QG | 2.1–7 | ||
| Micah 5.1 + 2 Kingdoms 5.2 (MT 2 Samuel 5.2) |
2.5–6QMG | 2.8–20 | ||
| Hosea 11.1 | 2.15QG | |||
| ?Judges 16.17 ?Isaiah 4.3 |
2.23G | 2.39–40 | ||
| Exodus 13.12 (≈ Exodus 13.2) |
2.23G (2.24 goes on to allude to Lev 5.11, 12.8) | |||
| Malachi 3.1 | 11.10QG | 1.2QCM (joined to following) | 7.27QG | |
| Isaiah 40.3–5 | 3.3QC | 1.3QC | 3.4–6QC | |
| Deuteronomy 8.3 | 4.4QG | 4.4QG | ||
| Psalm 90.11–12 (MT 91.11–12) | 4.5–6QG | 4.9–11QG | ||
| Deuteronomy 6.16 | 4.7QG | 4.12QG | ||
| Deuteronomy 6.13 | 4.10QG | 4.8QG | ||
| Isaiah 8.23–9.1 | 4.13–16QC | 1.14 | 4.14 | 4.43 |
| Isaiah 61.1 + 58.6 + 61.2 | 4.17–21QC | |||
| Isaiah 53.4 | 8.17QC (MT only) | 1.32–34 | 4.40–41 | |
| Isaiah 42.1–4 | 12.17–21QC | 3.7–12 | 6.17–19 | |
| ‘it is written’ (≈ MT Psalm 78.24, Exodus 16.4–5) | 6.31G | |||
| Isaiah 54.13 | 6.45QG | |||
| Deuteronomy 19.15 | 6.45QG | |||
| Isaiah 6.9–10 | 13.14–15QC | 4.10–12Q | 12.39–41QC | |
| Psalm 77.2 (MT 78.2) | 13.35QG | 4.34 | ||
| Isaiah 29.13 | 7.6–7QC (LXX only) | |||
| ‘as it is written’ | 11.14 | 9.13G | ||
| unnamed ‘prophets’ | 20.17–19 | 10.32–34 | 18.31G | |
| Isaiah 53.12 | 15.28 (spurious verse) | 22.37G | ||
| Isaiah 62.11 + Zechariah 9.9 |
21.4–5QMG (LXX only) | 11.1–7 | 19.29–35 | 12.14–15QG |
| Psalm of Solomon 7.1 (≈ Psalm 34.19, 68.5 / MT 35.19, 69.5) |
15.24–25QG | |||
| Isaiah 56.7 + Jeremiah 7.11 |
21.13QG | 11.17QG | 19.46QG | |
| Psalm 8.3 | 21.16QG (LXX only) | |||
| unnamed ‘writings’ (e.g. Enoch 104.4–6) | 22.29–30G | |||
| Daniel 9.27 and/or 1 Maccabees 1.54 | 24.15–16QC (Daniel) | 13.14Q | 21.20 | |
| Hosea 9.7 (cf. Isaiah 63.4, Jeremiah 5.29) |
21.22QG | |||
| Zechariah 13.7 | 26.31QMG | 14.27QMG | ||
| unnamed ‘writings’ | 26.56G | 14.49G | 22.47–53 | 18.2–12 |
| Zechariah 11.12–13 | 27.5–10QCM | |||
| Psalm 68.22 (MT 69.21) | 27.34Q (one word) | 15.23 | 19.28–29QG (one word) | |
| Psalm 21.2 / MT 22.1 | 27.46Q (Aramaic only) | 15.34Q (Aramaic only) | ||
| Psalm 30.6 (MT 31.5) | 27.50 | 15.37 | 23.46Q (LXX only) | 19.30 |
| unnamed prophets, psalms | 24.44G | |||
| ?Hosea 6.2 | 24.46–47QG |
Notes
1. Matthew cites scripture and names his sources more than the other gospels. There are 25 quotations/citations in Matthew, almost twice as many as in the runner-up, Luke. Perhaps that’s one reason why Papias thought Matthew was originally written in Hebrew (it wasn’t), and why Matthew was imagined in antiquity to be the earliest of the canonical gospels (it isn’t). The nativity story is especially densely packed with quotations; also notable is the string of unattributed quotations from Psalm 90 and Deuteronomy 6–8 in Satan’s testing of Jesus in Matthew 4 (largely copied in Luke 4).
2. Isaiah is the only scriptural source regularly mentioned by name. Isaiah is cited by name 6× in Matthew, 2× Mark, 2× Luke, 1× John. Two other sources get isolated citations: (1) A reference to Daniel in Matthew 24.15; but the text is a little odd (see the next point). (2) Two references to Jeremiah in Matthew 2.17 and 27.9; of these, the latter is a misattribution — the quotation is actually Zechariah.
3. The citation of Daniel in Matthew 24.15. The source for this quotation would be ambiguous, if not for the explicit citation of Daniel. The form used, βδέλυγμα τῆς ἐρημώσεως (‘sacrilege of desolation’), matches 1 Maccabees 1.54; Daniel uses the plural ἐρημώσεων (‘sacrilege of desolations’). And it turns out textual support for the citation of Daniel is less consistent than we might prefer. The modern text διὰ Δανιὴλ τοῦ προφήτου is the reading in the codex Vaticanus (δανιηλ) and the codex Bezae (δανιηλου), and the Latin version of Origen’s Commentary on Matthew includes the Daniel reference. However, the Sinaiticus instead has διὰ ιηλ τοῦ προφήτου, ‘through the prophet (of) Israel’, a fact omitted in the Nestle-Aland apparatus: δαν appears as a supplement above the line in a second hand. The part of Matthew 24.15 that cites Daniel is not extant in any ancient papyri, the codex Alexandrinus, or the Ephrem palimpsest.
Some factors make me hesitate over the Daniel attribution: (1) ιηλ in the Sinaiticus; (2) the fact that the wording is that of 1 Maccabees; (3) the fact that the gospels do not usually cite sources other than Isaiah by name.
LXX Daniel’s plural ἐρημώσεων is presumably intended to convery the MT reading šiqqūṣîm mešōmêm, a plural noun with a singular participle. The Sinaiticus reading ιηλ would most normally be read as a nomen sacrum for Ἰσραήλ (‘the prophet (of) Israel’). I wonder if it could be meant to invoke the anonymous author of 1 Maccabees. The Hebrew title of 1 Maccabees is not certainly known; Eusebius gives it as Σαρβηθ Σαβαναιελ (Ecclesiastical history 6.25).
By itself this isn’t enough to overturn the combined testimony of the Vaticanus and the Bezae. At the same time, the Sinaiticus reading ιηλ doesn’t look like a mistake. There’s something odd going on behind the scenes here.
4. Spliced quotations. In some quotations the gospels combine passages from multiple sources as if they’re a single quotation.
| Sources (LXX, my translation) | Gospel (NRSVue) |
|---|---|
| Micah 5.1 (MT Micah 5.2) And you, Bethlehem, house of Ephratha: you are the least that is to be among the throngs of Judah; from you shall come forth for me someone who is to be ruler in Israel: their exoduses are of old, from ancient days. 2 Kingdoms 5.2 (MT 2 Samuel 5.2) And once upon a time, when Saul was king over us, you were the one who led Israel out and led in Israel, and the Lord said to you, ‘You shall shepherd my people Israel, and you are to be leader over Israel.’ |
Matthew 2.5–6 5 They told him, ‘In Bethlehem of Judea, for so it has been written by the prophet: 6 “And you, Bethlehem, in the land of Judah, are by no means least among the rulers of Judah; for from you shall come a ruler who is to shepherd my people Israel.”’ |
| Malachi 3.1 See, I am sending out my messenger, and he will attend to a way ahead of me, and the Lord whom you seek will suddenly arrive in his temple, and the messenger of the covenant whom you desire. See, he is coming, says the Lord Almighty. Isaiah 40.3 A voice of one crying out in the desert: ‘Prepare the way of the Lord, make straight the paths of our God.’ |
Mark 1.2–3 As it is written in the prophet Isaiah, ‘See, I am sending my messenger ahead of you, who will prepare your way, 3 the voice of one crying out in the wilderness: “Prepare the way of the Lord; make his paths straight” ’... |
| Isaiah 62.11 For see, the Lord made it to be heard to the end of the earth: ‘Tell the daughter of Sion, Look, the saviour is at hand for you. He has received his payment, and his work is in his past.’ Zechariah 9.9 Rejoice greatly, daughter of Sion! Announce it, daughter of Jerusalem! Look, your king is coming to you, just and rescuing, humble and mounted on a donkey and on a young colt. |
Matthew 21.4–5 This took place to fulfill what had been spoken through the prophet: 5 ‘Tell the daughter of Zion, Look, your king is coming to you, humble and mounted on a donkey, and on a colt, the foal of a donkey.’ |
Each of these is an egregious misquotation, even aside from the splicing. Matthew 2.6 adds in a ‘not’ that reverses the meaning of Micah 5.1 (you are by no means the least); Mark 1.2 changes the identity of the one coming behind the messenger (I am sending my messenger ahead of you).
The third is the most notorious. The Septuagint has an erroneous ‘and’ added into the text, with the result that Matthew has Jesus riding on two animals simultaneously. The Hebrew text reads:
Behold: your king is coming to you, a just savior is he, humble, and riding on a donkey, on a colt, the foal of a donkey.
Zechariah 9.9 (NABRE)
In the Hebrew, the ‘donkey’ and the ‘colt’ are a case of parallelism: one animal is mentioned twice, phrased differently. This is a standard Hebrew poetic device. But the Septuagint (and so also Matthew) misunderstands that, and treats them as two separate animals.
Mark 11.1–7, Luke 19.29–35, and John 12.12–15 have Jesus riding a single colt (and John quotes a version of Zechariah 9.9, more correctly than Matthew’s version), so the oddity here comes solely from the Greek text used by Matthew. (The Marcionic Evangelion omitted this episode altogether.)
5. Quotations of Isaiah 40.3. This is one of the handful of scriptural passages that get quoted in all three synoptic gospels. As noted above, Mark joins it to Malachi 3.1 and attributes the whole thing to Isaiah. Matthew and Luke do use the Malachi quotation as well, but in completely separate contexts: evidently they corrected Mark, aware that the attribution to Isaiah was wrong.
(For reference, the scriptural passages that do appear in all three synoptics are: Malachi 3.1; Isaiah 40.3; Isaiah 56.7 + Jeremiah 7.11. Also notable is Isaiah 6.9–10, which appears in Mark, Matthew and John, but not Luke.)
Luke goes further, though, and provides a longer Isaiah quotation. Here’s how it lines up against the LXX text that we have today; the highlighting indicates places where the match is exact.
| LXX Isaiah 40.3–5 | Luke 3.4–6 |
|---|---|
| 3 φωνὴ βοῶντος ἐν τῇ ἐρήμῳ ῾Ετοιμάσατε τὴν ὁδὸν κυρίου, εὐθείας ποιεῖτε τὰς τρίβους τοῦ θεοῦ ἡμῶν· | 4 ὡς γέγραπται ἐν βίβλῳ λόγων Ἡσαΐου τοῦ προφήτου· φωνὴ βοῶντος ἐν τῇ ἐρήμῳ ἑτοιμάσατε τὴν ὁδὸν κυρίου, εὐθείας ποιεῖτε τὰς τρίβους αὐτοῦ. |
| 4 πᾶσα φάραγξ πληρωθήσεται καὶ πᾶν ὄρος καὶ βουνὸς ταπεινωθήσεται, καὶ ἔσται πάντα τὰ σκολιὰ εἰς εὐθεῖαν καὶ ἡ τραχεῖα εἰς πεδία· | 5 πᾶσα φάραγξ πληρωθήσεται καὶ πᾶν ὄρος καὶ βουνὸς ταπεινωθήσεται· καὶ ἔσται τὰ σκολιὰ εἰς εὐθείαν καὶ αἱ τραχεῖαι εἰς ὁδοὺς λείας· |
| 5 καὶ ὀφθήσεται ἡ δόξα κυρίου, καὶ ὄψεται πᾶσα σὰρξ τὸ σωτήριον τοῦ θεοῦ· ὅτι κύριος ἐλάλησεν. | 6 καὶ ὄψεται πᾶσα σὰρξ τὸ σωτήριον τοῦ θεοῦ. |
Notably, Luke 3.4 preserves αὐτοῦ from Mark 1.3: that word doesn’t appear in our LXX text of Isaiah 40.3.
6. Luke and the Evangelion. How many of Luke’s citations appeared in the Evangelion, the gospel that appeared in Marcion’s canon?
Background: the Evangelion and Luke are variant recensions of one another. The Evangelion was lost in antiquity, but is moderately well-attested because of its inclusion in Marcion’s canon. Marcion made the earliest effort to compile a canon of Christian texts in the early second century, and included ten Pauline letters and the Evangelion. Later anti-Marcionite writers claimed polemically that Marcion concocted the Evangelion by trimming bits out of Luke. However, we now know that Marcion compiled the texts, he didn’t write them; and in recent years it’s become standard to conclude that it was exactly the other way round — the Evangelion was the older recension (but not necessarily the ‘original’), and Luke is a later expanded form.
Remarkably, we know that most of Luke’s quotations/citations were not present in the Evangelion. Here’s a tabulation, primarily based on the work of Dieter T. Roth (2015: 49–78): this means some inconsistency with other recent work on the Evangelion, but Roth’s presentation of the evidence is by far the clearest currently available.
| Luke | Source | Present in Evangelion? | Testimony |
|---|---|---|---|
| 2.23 | Exod 13.12 | not present | Tertullian, Epiphanius, Origen |
| 3.4–6 | Isa 40.3–5 | probably not present | Tertullian, Epiphanius |
| 4.4, 4.8, 4.9–11, 4.12 | Deut 8.3, 6.13; Ps 90.11–12; Deut 6.16 | not present | Tertullian |
| 4.17–21 | Isa 61.1 + 58.6 + 61.2 | unknown | |
| 7.27 | Mal 3.1 | present | Tertullian, Epiphanius, Adamantius Dialogue |
| 18.31 | (generic) | not present | Tertullian, Epiphanius |
| 19.46 | Isa 56.7 + Jer 7.11 | not present | Tertullian, Epiphanius |
| 21.22 | Hos 9.7 | not present | Tertullian, Epiphanius |
| 22.37 | Isa 62.11 + Zech 9.9 | not present (NB text in Luke interpolated into Mark 15.28 | Tertullian, Epiphanius |
| 23.46 | Ps 30.6 | present | Tertullian, Epiphanius |
| 24.44 | generic | unknown | |
| 24.46–47 | Hos 6.2 | 46 unknown, 47 present | Tertullian |
A couple of notes on this tabulation:
- 3.4–6: see Roth 2015: 76 n. 69; Klinghardt 2021: 519–520. Roth thinks Tertullian’s comments on John the Baptist’s arrival, Adv. Marcionem 4.11.4, implies the absence of the Johannine narrative in Luke 3; Klinghardt thinks Epiphanius’ statement, Panarion 42.11.4–6, that the genealogy and the baptism were absent implies the absence of the whole of Luke 3.1b onwards. A given scholar’s verdict on this point is likely to be linked to their view on whether the Evangelion was pre-Mark or post-Mark. If it were post-Mark, and 3.4–6 were absent, then we would have to suppose that the Evangelion copied Mark but excised the quotation, then later on Luke restored it. Conversely, my opinion is that the timeline Evangelion > Mark > Luke involves some crazy textual acrobatics; then again, no one has to take my opinion seriously.
- 22.37: it is striking that, although Luke 22.37 was definitely not present in the Evangelion, it was considered important enough to copy it and interpolate it into the text of Mark (the spurious verse Mark 15.28).
7. Matthew and John. Some of John’s citations/quotations are striking for the fact that they have counterparts in Matthew: Mt 13.14–15 ≈ Jn 12.39–41 (both adding an explicit citation not present in Mark); Mt 21.4–5 ≈ Jn 12.14–15; Mt 27.34 ≈ Jn. 19.28–29 (both quoting just a single word). I find this supports the case that John knew Matthew.
Matthew and John are uneven in how faithfully they reproduce the older texts.
- Matthew 13.14–15 ≈ John 12.39–41. Matthew quotes the Septuagint text very exactly; John has only a few keywords in common. Both of them explicitly cite Isaiah. Mark echoes Isaiah, but only faintly, and doesn’t name the source.
- Matthew 21.4–5 ≈ John 12.14–15. This is the 'riding on a donkey and a colt’ passage that we looked at above, where the Septuagint and Matthew distort the meaning. Mark, Luke, and John have Jesus riding on a single animal, as in the Hebrew text. Only Matthew and John include a citation of Isaiah alongside the quotation.
- Matthew 27.34 ≈ John 19.28–29. Offering a drink to Jesus on the cross: wine, bile, myrrh, and/or vinegar? This one needs some explanation.
On the last item: Mark, Matthew, and John all take different approaches to interpreting the Septuagint text of the relevant passage. In the Masoretic Text, Psalm 69.21 reads (NABRE):
they gave me poison [rōš] for my food; and for my thirst they gave me vinegar [ḥōmeṣ].
The Septuagint changes this to (my translation):
For food they gave me bile [χολή], and for my thirst they watered me with vinegar [ὄξος].
Devouring bile makes no sense, of course. That’s probably why Mark 15.23 removes ‘bile’ altogether, and replaces it with ‘wine mixed with myrrh’. That disconnects the passage from the Psalm altogether (my catalogue at the start doesn’t count Mark 15.23 as a quotation/citation).
| Note. Why does Mark choose myrrh? Difficult to say. Certainly not as a painkiller, as the Oxford Annotated Bible weirdly alleges. In ancient sources wine perfumed with myrrh is normally a symbol of luxuriousness. Another possibility is to do with the medicinal use of myrrh to aid digestion and stimulate bile production. |
Matthew abandons Mark’s invention and brings it back a bit closer to the Psalm: ‘wine mixed with bile [χολή]’. Bible translations traditionally obfuscate this by using ‘gall’, an archaic word for bile.
John instead fixes the problem by going back to the other unpleasant foodstuff, the ὄξος ‘vinegar’. Modern translations usually mistranslate this word, probably self-consciously, as ‘sour wine’ (e.g. ESV, NRSV, NASB) or ‘cheap wine’ (GNB; NABRE: ‘common wine’). This is very strange, given that the same translations have no hesitation in translating Psalm 69.21 ḥōmeṣ/ὄξος as ‘vinegar’. A corrected translation of John 19.21 reads:
A jar full of vinegar [ὄξος] was standing there. So they put a sponge full of the vinegar [ὄξος] on a branch of hyssop and held it to his mouth.
When the chain of mistranslations and paraphrases is already hard to work out, modern Bible translations do us no favours by actively obfuscating the relationships between the texts with inconsistent or archaic translation choices. This is one of those rare cases where the NIV (‘wine vinegar’) is more accurate than the NRSV!
References
- Klinghardt, M. 2021. The oldest gospel and the formation of the canonical gospels (2 vols.). Leuven/Paris/Bristol (CT).
- Roth, D. T. 2015. The text of Marcion’s gospel. Leiden/Boston.
