Numerical composition and reading Luke

I recently wrote about hidden codes in Tom Wright‘s latest tome and similar patterns in the New Testament. In the former, this took the form of a acrostic code, using the letters of the text, and in the latter took the form of numerical features of the text.

In the comments, there developed an interesting discussion about how one might detect elements of numerical composition, and whether some people are attributing significance to numbers which are just coincidental. I highlighted there an interesting feature of Peter’s speech in Acts 2:

Here’s another remarkable feature pointed out by Menken (though not discovered by him):

There are also several instances of isopsephia in Acts, where the number of syllables of an episode or speech is equal to the numerical value of an important name or word occurring in or related to the passage in question (such as we found concerning John 1.1-18, where both the number of syllables and the numerical value of monogenes are 496). Peter’s speech in Acts 2.14-b-36 is made up of two equal halves: 444 syllables in 2.14b-24, and again 444 syllables in 2.25-36. Their sum, 888, is the numerical value of the name Iesous, a number which was famous in this quality in the second century, witness Irenaeus’ Aversus Haereses 1.15.2.

Richard Fellows raised a methodological objection to this:

The NT is about 300 times longer than this passage, so it is not so surprising that one passage would have 888 syllables, is it?

To which I offered a rough-and-ready response:

There are 24 speeches in Acts (that in itself is interesting!). I haven’t counted their length, but suppose they are between 500 and 1000 words or syllables. I guess there might be five numbers that could be seen as ‘significant’ such as 888, 1260, 1000, 153 and so on–but not many more. At a rough calculation, that would suggest that there is only 1% chance of a speech have a special length ‘by accident’. So there is a prima facia case for taking note of this.

51E73K7VQGLBut is there a better case for taking numerical composition seriously? I came across another fascinating example of Menken’s  (Novum Testamentum XXX, 2 (1988) pp 107–114.) by means of Joel Green’s excellent commentary, when preaching on Luke 7.11–17, the story of Jesus raising the son of the widow of Nain. A careful reading of this story, alert to its narrative features, would highlight the importance of v 13 as the turning point of the story. Jesus comes to the town, approaches the gate, sees the funeral procession, then sees the widow, and ‘his heart went out to her’—which then leads to the miracle. The verb is splagchnizomai, meaning ‘literally’ ‘his bowels were moved.’ It is Jesus’ compassion for the grieving woman which motivates him to perform the miracle.

But Menken goes one step further. Noting that the compassion of Jesus is an important theme in Luke, but also that this word occurs only three times, he analyses these occurrences, and finds something striking. The occurrences are:

  • Luke 7.13    The raising of the widow’s son
  • Luke 10.33    The parable of the man who fell among thieves
  • Luke 15.20   The parable of the two sons and the forgiving father

In each case, the verb is the turning point of the narrative unit from a literary point of view. But Luke has also written in such a way as to place the verb (or verb phrase) at the numerical centre as well.

In the first, the narrative unit has 208 syllables; the phrase ‘he had compassion on her’ is 7 syllables; the unit has 106 syllables preceding it and 105 following it (106 + 7 + 105 = 208). But considering the words, he argues that the composition is even more carefully structured:

Screen Shot 2014-05-09 at 10.09.59


In Luke 10, the verb splagchnizomai is preceded by 68 words and followed by 67 in the story told by Jesus. It is the central (17th) of the 33 verbal forms in Jesus’ words.

In Luke 15, the same verb acting as the turning point of the narrative is also numerically central to the indicative verbs in the story (16th of 31) and of the aorist (simple past) indicatives (12th of 23).

I think two things follow from this, one relating to study of NT texts, the other relating to our reading of and preaching from these texts.

The first is that, if any of this evidence stands up to scrutiny, then it makes a strong case for seeing it elsewhere too. In this example, Menken has started with an important word, and found numerical structure build around it. Because of the direction of the argument, this is much less susceptible to the kind of ‘it is just a coincidence’ critique of the 888 syllables in the Acts 2 speech. But finding numerical composition in one part of a writer’s work strengthens the probability that this is happening elsewhere too; it is a cumulative case.

The second is that, in reading these texts and preaching on them, it might well be that the actual shape and structure of the text is communicating something of importance. In my sermon, it offered a powerful, pastoral point. The centre feature of God’s response to tragedy is one of compassion, and Luke communicates this by placing this term at the numerical centre of the respective narrative units. Rather than seeing time spent with the text as a distraction from our pastoral application of the text, we perhaps should see the two as more closely related, and the second deriving from the first.

Signup to get email updates of new posts
We promise not to spam you. Unsubscribe at any time.
Invalid email address

If you enjoyed this, do share it on social media (Facebook or Twitter) using the buttons on the left. Follow me on Twitter @psephizo. Like my page on Facebook.

Much of my work is done on a freelance basis. If you have valued this post, you can make a single or repeat donation through PayPal:

For other ways to support this ministry, visit my Support page.

Comments policy: Do engage with the subject. Please don't turn this into a private discussion board. Do challenge others in the debate; please don't attack them personally. I no longer allow anonymous comments; if there are very good reasons, you may publish under a pseudonym; otherwise please include your full name, both first and surnames.

6 thoughts on “Numerical composition and reading Luke”

  1. Thanks again Ian. I have often preached pastorally from the Parables bearing in mind material from Ken Bailey’s works. Structural issues in the NT fascinate me as often we see the NT writers as backward or inferior but fail to see the poetry and ingenuity…..

  2. Ian, I think the statistics is the wrong way round. It is not the chance of it occurring by accident we should be looking at but the chance that it didn’t occur by accident. I am not explaining this very well but statistics is funny in this respect.

    There was a famous case in America where a witness saw a dark haired criminal escape in a red car with a blonde woman. The police found a red car, with a blonde woman and a dark haired man. The prosecutor said there was only a 1 in a million chance of there being a red car with a blonde woman and a dark haired man. the man was sent to prison. However the true question is given that they match the description of the guilty pair, what is the chance that the couple are the criminals. The odds turn out to be 40% or even less. the couple were released on appeal.

    I think you need to ask how many classical greek speeches reported in books were of approximately this length? Or what are the lengths of all the other 23 speeches? If 800-900 is a common length in classical Greek or for other speeches then the odds of it happening by chance are much higher.

  3. I’m no mathematician, but this article correlates with many of the concerns that I have about Biblical numerology carried out by people who aren’t statisticians:

    Would be interesting to get an opinion on this from Peter Ould, who works professionally with both the Bible and statistics.

    • Yes, that is not an uncommon kind of illustration. But we are talking about something quite different here—is there significance to a speech having a length, the number of which is known (from other sources) to have significance, and which is used in other, similar literature.

      The statistics themselves are not as important as the cultural issue—in reading the NT we are entering a foreign culture, where these things were given a significance which is instinctively strange to us.

  4. “In Luke 10, the verb splagchnizomai is preceded by 68 words and followed by 67 in the story told by Jesus.” I’m really struggling to see this. Depending on the text-critical decision in verse 32, counting backwards 68 words from ἐσπλαγχνίσθη lands you at εἶπεν (“[wanting to justify himself,] he said”) or ἑαυτὸν (“[wanting to justify] himself”). I’d have said the story Jesus tells begins at ἄνθρωπός (“A man”), 55/54 words before ἐσπλαγχνίσθη. Counting 67 words from ἐσπλαγχνίσθη lands me at the direct article with which the answer given by the Torah expert opens.

    On my count, there are 54 or 55 words in the story before ἐσπλαγχνίσθη and 50 afterwards (63, if we add the question appended to the story).

    The best I can do is to count 50 words (including γενόμενος in verse 32) from the beginning of the story until “came near him; and when he saw him, he was moved with compassion” and 50 words from there to the end of the story before we get to the question Jesus adds.


Leave a comment