The Contents of Codex P46

P⁴⁶ is a papyrus codex of many of the Pauline epistles, most likely written around the year 200 (based on paleographic examination), currently consisting of 86 leaves containing portions of Rom 5:17 to 1 Thes 5:28, and also Hebrews (following Romans). The original size of the codex is uncertain due to the fact that all pages are eroded to a greater or lesser degree, however a reasonable estimate is that when folded it was approximately 16cm by 28cm (6.3in by 11in), containing text written as a single column varying from approximately 10 to 12.5cm (3.9 to 4.9in) in width on each page. All the pages are missing some text, mainly at the bottom corners of the leaves, with the most damage at the end of the codex.

The extant portion of P⁴⁶ is in the form of a single quire, i.e. it was created by laying sheets of papyrus on top of each other and then folding them in the middle, much like a modern magazine. The outer (verso) sides of the sheets have vertical fibers, and the inner (recto) sides have horizontal ones. The text also flows as in a magazine, from verso to recto, until at the center (which consists of two recto pages from the same sheet of papyrus) it switches to flowing from recto to verso. The edges of the leaves were trimmed so that they were aligned when folded, thus making the inner leaves smaller than the outer ones.

However, there are two big differences between a magazine and a codex like P⁴⁶. The first difference is that the pages of a magazine are printed before it is assembled and bound, whereas P⁴⁶ was bound and the leaves trimmed before the scribe wrote on them.[1] The second big difference is the thickness. P⁴⁶ originally contained 52 sheets of papyrus, folded in the middle to form 104 leaves. Estimates of the size of the outermost and innermost sheets suggest a width difference of 2.6 cm, so indicating that the innermost pages were 1.3 cm narrower than the outermost ones, which in turn suggests that the thickness of the binding was also 1.3 cm. As the codex originally contained 52 sheets of papyrus this would indicate that each sheet is 0.25 mm thick, or approximately five times the thickness of the paper used in a typical modern magazine. This resulted in a codex approximately one inch thick, in which the outer leaves would have curved significantly towards the fold.[2] The seven outer sheets and two others (18 leaves, or 36 pages) have been lost.

The Characteristics of the Text

It has been suggested that the whole manuscript appears to have been written by a single professional scribe, as there are stichoi notations at the end of several of the epistles, and also blank lines and centered headings before each one. The scribe himself made a number of corrections, and these were later added to by several other correctors. The text of P⁴⁶ is similar to that of both B and P¹³. However, this article does not discuss in detail the characteristics of the text itself (e.g. variants), but rather how much text was present, and how the text was presented. There are many easily accessible online discussions of the text itself, for example from The University of Michigan (U-M) and Robert Waltz.

The Current Contents of P⁴⁶

The surviving leaves contain most of nine of the Pauline epistles, in the order given below. All the leaves are damaged at the bottom to a greater or lesser extent, with several missing one or more complete lines of text [marked below by square brackets]. Lacunae (gaps) affecting less than a whole verse are not indicated, and for details of missing characters on many of the pages see the images provided by U-M, The Center for the Study of New Testament Manuscripts (CSNTM), and Peter M. Head.

Leaves 8v-21r: Romans (Rom)

[Leaves 1-7 missing], 5:17-6:3 [6:4], 6:5-14, [Leaves 9-10 missing], 8:15-25 [8:26], 8:27:35 [8:36], 8:37-9:32 [9:33], 10:1-11:22 [11:23], 11:24-33 [11:34], 11:35-15:10a [15:10b-11a], 15:11b-33, 16:25-27, 16:1-23 (v. 16:24 not in ms)

Leaves 21r-38v: Hebrews (Heb)

1:1-9:16 [9:17], 9:18-10:20 [10:21], 10:22-30 [10:31], 10:32-13:25

Leaves 38v-60v: 1 Corinthians (1 Cor)

1:1-9:2 [9:3], 9:4-14:14 [14:15], 14:16-15:15 [15:16], 15:17-16:23 [16:24]

Leaves 61r-74v: 2 Corinthians (2 Cor)

1:1-11:10 [11:11], 11:12-21 [12:22], 11:23-13:13 (vv. 8:19a, 19c-20, 13:14 not in ms)

Leaves 74r-81r: Ephesians (Eph)

1:1-2:7 [2:8-9], 2:10-5:6 [5:7-8], 5:9-6:6 [6:7], 6:8-18 [6:19], 6:20-24

Leaves 81r-86r: Galatians (Gal)

1:1-8, 1:10-2:9 [2:10-11], 2:12-21 [3:1], 3:2-29 [4:1], 4:2-17 [4:18-19], 4:20-5:17 [5:18-19], 5:20-6:8 [6:9], 6:10-18 (v. 1.9 possibly shorter or not present in ms)

Leaves 86r-90r: Philippians (Php)

1:1 [1:2-4], 1:5-15 [1:16], 1:17-28 [1:29], 1:30-2:12 [2:13], 2:14-27 [2:28], 2:29-3:8 [3:9], 3:10-21 [4:1], 4:2-12 [4:13], 4:14-23

Leaves 90r-94r: Colossians (Col)

1:1-2 [1:3-4], 1:5-13 [1:14-15], 1:16-24 [1:25-26], 1:27-2:19 [1:20-22], 2:23-3:11 [3:12], 3:13-24 [3:24-4:2], 4:3-12 [4:13-15], 4:16-18

Leaves 94r-97v: 1 Thessalonians (1 Thes)

1:1 [1:2-8] 1:9-2:3 [2:4-??], [Leaves 95-96 missing], 5:5-9 [5:10-22], 5:23-28 [Lines missing], [Leaves 98-104 missing]

This order differs slightly from what we see in modern bibles, which have: Romans, 1 Corinthians, 2 Corinthians, Galatians, Ephesians, Philippians, Colossians, and 1 Thessalonians, which are then followed by 2 Thessalonians, 1 Timothy, 2 Timothy, Titus, and Philemon, and finally Hebrews.

The Problem

There is really no dispute regarding what was present on the missing leaves at the beginning of P⁴⁶. Although there is no definite proof, the only thing that makes any sense is that they contained Rom 1:1-5:16. However, there has been much speculation about the original contents of the last seven leaves (98 to 104), and despite several theories, their contents have never really been determined to everyone's satisfaction.

Calculating how many characters could have been fitted onto the final missing leaves would appear to be a fairly simple task: Count how many are on each extant leaf, from that calculate the average number of characters per leaf, and then multiply by the number of missing leaves. However, in the case of P⁴⁶ this is complicated by a number of factors:

Due to the way this codex was constructed, the inner leaves are narrower than the outer ones, meaning that the scribe was able to write more characters per line on the outer leaves than on the inner ones. David Trobisch (Paul's Letter Collection) wrote that: "When several sheets of paper were folded in the middle, the inner leaves would stick out and would be cut in order to make the book look nicer. This is what was done to P⁴⁶ before the scribe started writing."
Because of the damage to the bottom of the leaves, many lines have to be re-constructed using assumptions based on the rest of the text, e.g. the use of nomina sacra.

The scribe started by writing around 26 lines per page, increasing to 28 in the middle of P⁴⁶, and then increasing further to around 31 lines per page in the final quarter[3].
In the final third of P⁴⁶ the scribe increased the number of characters per line by writing them closer together[4].
In the first third (roughly) of P⁴⁶ the scribe tended to write an average of around two more characters per line on one side of a leaf (the verso) than he did on other (the recto). In the middle third there is no clear distinction between recto and verso, but then in what is roughly the final third the scribe wrote more on the recto than on the verso, with the difference increasing to about 3 characters per line by the end of the extant portion of the manuscript.
The scribe left blank lines between each epistle in the extant portion of P⁴⁶. The number of blank lines that might have been left on the final missing leaves depends on which (and how many) epistles the missing leaves contained.

Because of the above issues, even if we make the (reasonable) assumption that the contents of leaves 98 to 104 were one or more of the ‘missing’ Paulines, then because of spelling differences, mistakes, and the use of various shortened forms of common words, we cannot know for sure exactly how many characters the scribe would have used when writing these epistles.

However, despite these issues, it is still possible to estimate both the number of characters that could have fitted onto the missing leaves, and also the number of characters the scribe would have used to write each of the possible missing Paulines. This is due in no small measure to the fact that the whole codex was written by a single scribe, so that averages from adjacent lines or pages can be used where necessary. However, a reasonable question to ask is: “Why does this matter?” The simple answer is that P⁴⁶ is the key manuscript witness to the early existence of most of the Paulines. Only 2 Thessalonians, the Pastorals (1 & 2 Timothy, and Titus), and Philemon are missing. If it could be shown for certain that P⁴⁶ once contained any of these then it could have a big effect on the question of the authenticity of these letters, or at least, how they were perceived in the 2nd century[5].

How Do We Know How Many Sheets are Missing?

The extant portion of the codex begins with Rom 5:17, written on the verso (in this case the outside) side of a leaf. This is not the same as in modern publishing, in which the verso usually marks a left hand page with the recto on the right. However, when a sheet of papyrus is used, it is oriented so that the recto has horizontal fibers and the verso has vertical ones. Typically the recto, where the writing is along the fibers, would be used in preference to the verso, but in a codex both sides are used, usually with the verso on the outside and the recto on the inside.

There are two different ways we can use to determine how many leaves preceded Rom 5:17. The first involves counting the number of Greek characters in Rom 1:1 - 5:16, and figuring out how many additional leaves would have been required for this number of characters. However, the second way is much easier: look at the page numbers! Many of the pages in P⁴⁶ have visible page numbers, and the first to do so is the verso of a leaf containing Rom 8:15-26, which is marked as page 20. By counting backwards from here it is easy to determine that the verso of the first extant leaf is page 14, and that page 1 would have been on the recto (inner) side of a leaf, with the verso side of the same leaf quite likely forming the outside cover of the whole codex. Estimating the number of characters in the missing verses of Romans confirms that it began on page 1, and that the number of missing outer leaves is therefore most likely seven (It is possible that there could have been one or more additional 'introductory' leaves, but there is no evidence for this).

The First Half of P⁴⁶

As stated above, the scribe did not write the same number of characters on each page of P⁴⁶. In part this is due to the differing widths of the Greek characters, and the use of additional spacing between characters (e.g. before OT quotations), but as these differences are common to all pages it will be assumed that, on average, these variations affect all pages to a similar degree, and will not be discussed further. In the extant portion of the first half the scribe begins by writing an average of around 850 letters per page (plus or minus 50 or so), dropping relatively steadily to around 650 in the middle as the number of characters per line drops from an average of 33 down to 24. The main reason for this approximately 27% decrease in characters per line is a similar decrease in the width of the portion of the page on which the scribe wrote (i.e. excluding the margins). However, acting against this is the fact that the scribe also increases the number of lines per page from around 26 (+ or -1), to around 28 (+ or -1) by the middle. Overall, the innermost pages of the first half of P⁴⁶ contain approximately 24% fewer characters than the outermost. Assuming that the missing leaves at the beginning of the ms followed this pattern, the first half of P⁴⁶ originally contained approximately 79,300 characters.

An additional, and somewhat curious point, is that in the first quarter (roughly) of P⁴⁶ the scribe tends to write an average of around two more characters per line on the outer (verso) side of a leaf than he did on the inner (recto). Then, from roughly the mid point onward, the situation is reversed. He begins to write more on the recto than on the verso, with the difference increasing to about 3 characters per line by the end of the extant portion of the ms. This can be seen as another indication that P⁴⁶ had been already bound before the scribe wrote on it. Because of the thickness of the codex and the extent of the bend towards the fold, the scribe would generally have had more difficulty writing on the outer leaves than those closer to the middle, where most of the codex would have been opened out flat. Assuming he was right-handed (the most likely scenario) then when writing on the recto of the outer leaves the bulk of the codex on the right-hand side would have interfered with his hand, and prevented him from writing as close to the fold as he was able to on the verso. This effect would diminish as he got closer to the center of the codex. When writing the second half the bulk of the codex would be to the left of his hand, so this would not affect him.

The Second Half of P⁴⁶

The second half of P⁴⁶ begins like a 'mirror' of the first, with the number of characters per line increasing fairly steadily as the page size grows. From the center point of the codex the number of words per page grows from around 650 to around 800 by the end of Galatians. If the second half of P⁴⁶ had continued to ‘mirror’ the first then the overall capacity of the ms would have been approximately 152,600 characters. However, something unusual happened at this point.

The scribe seems to have realized he was in trouble, and that he needed more space for what he wanted to include, and so he increased the number of characters per page to around 1,000 (by further increasing both the number of lines per page and characters per line) by the time he completed Ephesians. He then continued, dropping slightly to the low-900s by the end of Philippians, but then increasing again to around 1,050 characters per page by the end of the extant portion of P⁴⁶. Assuming the same number of lines per page (around 32 on the verso and 30 on the recto in the last quarter of the codex), this would have then increased to nearly 1,100 by the original end of the quire, so providing space for approximately 171,500 (13,000 more) characters.

In ‘P46 and the Pastorals: A Misleading Consensus,’ Jeremy Duff graphs the number of letters in each group of five pages in the whole of P⁴⁶, and adds a trend line that: “... depicts the general trend once the variations between individual pages are smoothed out [that] is a sixth-order polynomial approximation: its r-square value (0.91) indicates that it a good fit.” Duff may well be correct, but his comment regarding the fit is irrelevant, as well as being misleading. This is because it is an attempt to graph the effect of two completely different mechanisms affecting the number of letters per page, one of which is not under the control of the scribe:

1. The inner leaves of the codex are smaller than the outer ones because they have been trimmed, so causing the number of letters per line to vary;
2. The scribe was increasing the number of lines per page in the second half of the codex.

In order to understand what the scribe was trying to achieve the effect of the first variation has to be first taken into account.

Why Was the Scribe in Trouble?

When a codex is created in the form of a single quire that is bound before writing, there is a more-or-less fixed area on which to write. Once the scribe has got past the half-way point in the codex (i.e. has started using the right hand side of sheets that already have text on the left hand side) there is then no easy way to create more space if needed. It is possible to increase the size of the quire by adding extra sheets to the outside (and so leaving the first leaves blank), or by inserting additional folded leaves between existing ones near to or at the end, but with an already thick codex like P⁴⁶ this seems unlikely. In particular, adding a small number of additional leaves at the end (where the outer leaves would have already been very curved close to the fold) would have been very unwieldy), and so most likely not even practical. Alternatively, the codex could have been un-bound, the sheets with writing on both left and right thrown away, enough new sheets added, and the codex re-bound, but this is more unlikely still.

It is worth pointing out that if the scribe of P⁴⁶ knew that any of these (or any other ways of increasing the space available for writing) were options then there would be no reason for him to make his writing smaller the closer he got to the end of the quire. Instead, it appears that he knew that his only option was to squeeze more in at the end of the quire because he could not increase the number of pages in the codex. The problem is that even at 32 lines (and approximately 1,100 characters) per page there is no way to complete 1 Thessalonians, and then write 2 Thessalonians, 1 & 2 Timothy, Titus, and Philemon, on the leaves that have been lost at the end (13 pages, assuming that the verso of the final leaf was left blank). This would have required around 25,000 characters, much more than the approximately 14,300 that would have been available. As P⁴⁶ includes some of 1 Thessalonians, it is a reasonable assumption that it used to include the whole of both 1 & 2 Thessalonians, but if so the four remaining epistles would have still required around 6 pages more than were actually available. So, why did the scribe apparently miscalculate so badly, and what was he trying to do?

Single or Multiple Source Manuscripts?

If the scribe of P⁴⁶ was working from a single source manuscript (either scroll or codex) containing all the epistles he was trying to copy then it would have been fairly easy for him to find the mid-point of that ms. Consequently he would have been able to estimate how much he had to write on the ‘left hand side’ of P⁴⁶before starting. He should therefore have been able to include enough leaves (or write small enough if the number of leaves could not be changed) to fit the first half of the text on the left hand side, thus also making sure that he had room for the second half on the right hand side. Either way, he certainly would have known by no later than the mid-point of the codex whether he would be able to fit everything in, and adjusted his writing immediately. Even if the size of P⁴⁶ was fixed (i.e. more leaves could not be added to it), the scribe should have been able to select just a sub-set of the epistles before starting, or again by the mid-point at the latest. The fact that he did not start adjusting his writing until he had written approximately three-quarters of the codex strongly suggests that he did not know he had a problem (or did not know the size of the problem) until then.

It is reasonable to assume that if the scribe was a professional then he would know how to organize his codex so that he would not run into the problem that we see manifested in P⁴⁶, and there does appear to be some evidence for that. In The Text of the Earliest New Testament Greek Manuscripts Comfort and Barrett write:

The scribe who produced this manuscript used an early, excellent exemplar. He was a professional scribe, because there are stichoi notations at the end of several books (see the conclusions of Romans, 2 Corinthians, Ephesians, and Philippians). The stichoi were used by professionals to note how many lines had been copied for commensurate pay. Most likely, an employee of the scriptorium (perhaps connected with a church library) paginated the codex and indicated the stichoi.

However, there are problems with this suggestion, because the stichoi do not correspond to the numbers of lines actually in P⁴⁶, as shown in the table below (Note: For Colossians only the first digit is readable):

The differences between the stichoi and the actual numbers of lines in P⁴⁶ suggest that Comfort and Barrett are incorrect in their suggestion that the stichoi indicate that the scribe was a professional. If anything, they point to the reverse:

There is no sticheron for 1 Corinthians;
The first group of stichoi (following Romans, Hebrews and 2 Corinthians respectively) are in multiples of 100, while the second group (following Ephesians, Galatians, Philippians, and Colossians) are exact numbers;
The great differences between the stichoi at the ends of Hebrews and 2 Corinthians and the actual number of lines in P⁴⁶ suggest a serious mis-calculation by the scribe;
Given that all but one sticheron under estimate the number of lines needed in P⁴⁶, the opposite in the case of Galatians is a significant anomaly (and suggests a possible reason for why we see Galatians before Ephesians in bibles today. See The Order of the Paulines).

These inconsistencies are very unlikely to have been created (or checked) by a professional scribe who used them to get paid for his work. However, if the scribe was not a professional, then there is no reason for him to have added stichoi to P⁴⁶, so why are they nevertheless present? Rather than the stichoi having been calculated either during the writing of P⁴⁶ or after its completion, the differences instead point to the scribe simply having worked from more than one exemplar that already contained the stichoi, with the scribe simply copying the numbers into P⁴⁶. The differences between the stichoi and the actual numbers of lines in P⁴⁶ can then be seen to reflect differences between the page widths of his exemplars and that of P⁴⁶, with (in most cases) the exemplars being wider than P⁴⁶. Another feature that points in this direction is that page 74v (containing the ending of 2 Corinthians) contains only 22 lines of text (fewer than any other page), while 74r (the other side of the same leaf) contains 28 lines of text, and 75v (the beginning of Ephesians) contains 27 lines of text plus the title line. The clear impression is that the scribe completed something important when he finished writing 2 Corinthians, and the most likely scenario is that he had copied all of one exemplar, which (from the sitchoi) possibly contained Hebrews, 1 Corinthians, and 2 Corinthians. This evidence for P⁴⁶ having been copied from more than one exemplar, with widths different to that of P⁴⁶, then provides a very simple explanation for why the scribe had such apparent difficulty calculating how many pages he needed: The different formats of his exemplars meant that he had to write more lines than he initially expected!

As just indicated, perhaps the most reasonable explanation for the scribe apparently not being able to calculate how much space he needed is that he was working from multiple exemplars, perhaps even one for each epistle. These mss may have been written on different width scrolls or codices, by different scribes or amanuenses, using different page widths and text sizes, thus making it very difficult to determine exactly how much space was required in the codex. Also, given what we know about what happened to Romans (i.e. that copies with different endings existed) it is not unreasonable to suppose that the scribe may have had more than one copy of this epistle to work from, of different lengths, making the calculation harder still. Additionally, the scribe may simply have been asked to include additional epistles (e.g. perhaps the Pastorals) after he had passed the mid-way point in the codex.

What Did The Scribe Most Likely Do?

From the above information, it is reasonable to believe that the scribe (or perhaps the person who gave him the assignment) did not know a 'correct' order for the epistles. We don't know how many different mss containing the exemplars of the epistles were given to the scribe, we don't know the order of the epistles in what he had been given, nor do we know what he had been asked or instructed to do with them. However, even if he had been given multiple exemplars in chronological order, the scribe is unlikely to have known the significance of that order. As a result, if he was in any way uncertain of the space required he would have felt it acceptable to re-order the epistles to make his task easier, and it would then make most sense to start with the longest letter (Romans), and work downwards in size. That would then minimize any space problem, by allowing him to omit one or more of the shortest letters at the end of his codex if needed. Duff comments that:

... it is the period of the production of P46 which is relevant in determining its contents, not the date of the composition of the various texts. There is little evidence that in, say, 200 A.D. the Pastorals were seen as different from the other Pauline epistles.

Actually, P⁴⁶ itself provides evidence that the Pastorals were seen as different. In the extant portion of P⁴⁶ the epistles are ordered in decreasing order of size (with the exception of Hebrews, but this can be seen as a wish to not split 1 & 2 Corinthians). Therefore, if the scribe of P⁴⁶ did not consider the Pastorals as being different, we would expect to find 1 Timothy before Philippians, and 2 Timothy (following the example of Hebrews) before 1 Thessalonians. While the position of 2 Timothy in P⁴⁶ is speculative (it is possible it might have followed 1 Thessalonians), it is clear that 1 Timothy is not where we would expect unless the scribe wished to distinguish the Pastorals (or perhaps the epistles addressed to specific individuals) from the others.

If the scribe had all of the Pauline epistles in front of him, and size alone was the reason for the order in P⁴⁶, then we would expect to see Hebrews between 1 & 2 Corinthians, and 1 & 2 Timothy come before 1 Thessalonians (or possibly Philippians). The decision not to split 1 & 2 Corinthians is understandable, but by the same token why do we not see 1 & 2 Timothy before 1 & 2 Thessalonians? Either the scribe did not have exemplars of 1 & 2 Timothy, or he did but they were considered to be of lesser importance. The latter is perhaps more likely, as 1 & 2 Timothy were known to Polycarp, most likely at a time before P⁴⁶was written.

As previously suggested, it is also possible that the scribe did correctly calculate the space he needed to copy his source text(s), but that at some point after starting writing the second half of P⁴⁶ he found (or was given, or was asked to include) one or more additional epistles. If he had not compressed his writing in the second half of the quire then it would have contained space for around 13,000 letters fewer than appears to have actually been in P⁴⁶, in which case P⁴⁶ would then have had approximately enough space to have contained everything it currently does, plus 2 Thessalonians, but nothing else.

On this hypothesis, the scribe did not originally know of Philemon and the Pastorals, and only came into contact with (or was asked to include) some or all of them after he had completed around three quarters of P⁴⁶. However, given the possible co-existence of P³²(Titus) and P⁸⁷(Philemon) and the likelihood that the scribe would have at least have known of Marcion’s Apostolicon (a collection of the Pauline epistles in the order Galatians, 1 & 2 Corinthians, Romans, 1 & 2 Thessalonians, Ephesians, Colossians, Philemon, and Philippians), it is possible that he would have known of at least Philemon, and possibly Titus. It is also possible that he was initially tasked with copying just the epistles written to churches or groups, and was not intending to include any epistles written to individuals. Duff comments:

Marcion, so we are informed, did distinguish 1 Timothy, 2 Timothy and Titus from the other Pauline letters.

However, in Adv Marcion Book V, Chapter XXI.-The Epistle to Philemon, Tertullian wrote:

To this epistle alone did its brevity avail to protect it against the falsifying hands of Marcion. I wonder, however, when he received (into his Apostolicon) this letter which was written but to one man, that he rejected the two epistles to Timothy and the one to Titus, which all treat of ecclesiastical discipline. His aim, was, I suppose, to carry out his interpolating process even to the number of (St. Paul's) epistles.

In the absence of any knowledge about why Marcion did not include the Pastorals (Tertullian’s comment that Marcion actually “rejected” them must be taken as supposition, as the final sentence quoted above shows), perhaps the most reasonable explanation is that Marcion simply didn’t have access to them.

Another point to consider is this: Why bother squeezing more onto the last pages unless you are sure that by doing so everything will fit? The implication is that the scribe did manage to include everything he was originally asked to, because he had planned the size of the codex accordingly. However, he had perhaps relied on the stichoi in his exemplars in the calculation, and so had to compress his writing to make everything fit. Unfortunately for him, at a late stage (when the size of the codex was already fixed) he was asked to include something extra that he had not planned for - perhaps Hebrews - and was only able to make adjustments when the full impact of the extra material became apparent towards the end of his task.

Conclusions

Without any of the changes in scribal practice we see in the second half of the manuscript, P⁴⁶ would have contained space for around 152,600 characters. This would not have allowed for what we currently see in P⁴⁶, as without compressing the text there would not have been room even for 1 Thessalonians after Colossians. However, by condensing his writing in the final quarter of the manuscript the scribe was able to include 1 & 2 Thessalonians and, almost certainly, Philemon, but not the Pastorals (1 & 2 Timothy and Titus):

Romans, Hebrews, 1 & 2 Corinthians, Ephesians, Galatians, Philippians, Colossians, 1 & 2 Thessalonians, Philemon

It is clear that the scribe started to run out of space for what he intended to include in the manuscript, but it is also clear that he never intended to include the Pastorals. If that had been his intent, it would have become obvious to him by the mid-point of P⁴⁶ that, instead of creating space for the 13,000 additional characters he actually achieved, he would have needed to add around 31,000 more instead. He could perhaps have added additional leaves in the center of the codex, or increased the number of characters per page throughout the whole of the second half of P⁴⁶, but he did neither, instead just adding enough for what he intended to include. However, this does not answer the question of why the scribe needed to add anything at all.

The content of P⁴⁶ is close to that of Marcion's earlier Apostolicon. It contains the same epistles, except that P⁴⁶ also contains Hebrews, and the order is significantly different from Marcion’s:

Galatians, 1 & 2 Corinthians, Romans, 1 & 2 Thessalonians, Ephesians, Colossians, Philemon, Philippians

Given the range of dates ascribed to P⁴⁶ (From late 2nd to early 3rd century), it is quite possible that it was written specifically to counteract Marcion, using ‘approved’ mss of the Paulines instead of the (in some places quite different) text found in the Apostolicon. For example, Marcion’s version of Romans omits both chapters 15 and 16, while that in P⁴⁶ omits just v. 16:24. This would then account for the scribe having several different mss as his exemplars (and so making his size calculations more complicated).

The re-ordering also can be seen as an anti-Marcionite reaction: a different order was wanted in order to avoid appearing to ‘approve’ of that used by Marcion, and, for reasons given above, size order was the most reasonable instead. However, if the scribe of P⁴⁶ had been asked to create an ‘approved’ collection of Pauline epistles containing just those that Marcion had, he would not have needed to include Hebrews, and would have had plenty of space for everything else. It therefore seems likely that P⁴⁶ (the blank codex) was created and bound under this assumption, and that it was only decided to add Hebrews after the size of P⁴⁶ was fixed. If the scribe had not had to include Hebrews he would have had some space to spare at the end of the ms, and so he waited until he knew how much more space he needed, and then compressed his writing enough to complete the modified task he had been set. We don't know why Hebrews was included in P⁴⁶, but it is conceivable that if it had not been included Hebrews would never have been thought of as a Pauline epistle.

Data

The data on which the calculations here are based can be found in this Excel workbook (or use the link at the bottom of this page). The raw data used to obtain the numbers found in the 'Detailed Data' worksheet were obtained by counting the characters shown in the transcript of P⁴⁶ in The Text of the Earliest New Testament Greek Manuscripts from Comfort and Barrett, and by extrapolation where necessary in the case of lacunae, e.g. in the case of the missing leaves (shown in magenta in the 'Detailed Data'). I have included their reconstructions of the lacunae when counting. For the missing leaves I estimated the line and character counts based on extrapolation from the extant leaves, with the estimated numbers shown in magenta. The other worksheets contain the following graphs:

Letters-Page H1 The number of letters/page in the first half, with two trend lines: A period-6 moving average, and a straight line.
Letters-Page H2 The same for the second half, except that the magenta trend line is a ‘reversed’ copy of the first half line.
Lines-Page The number of lines of text per page (excludes titles etc.).
Lines-Page V&R As per 3), but with Verso and Recto shown in different colors.
Lines-Page V-R The difference between the line counts on the verso and recto of each page.
Letters-Line The number of letters per line of text on each page, with separate straight-line trend lines for each half of the ms.
Letters-Line V&R Similar to 6), but with Verso and Recto in different colors, and period-6 moving average trend lines
Letters-Line V&R H1 Similar to 7), but for the first half only, and with straight trend lines added.
Letters-Line V&R H2 Similar to 8), but for the second half only.
Letters-Line V-R The difference between the letters/line counts on verso and recto of each page, with period-6 and straight trend lines.

To view the charts clearly it may be necessary to download the whole workbook as an Excel file. Even then, if anything is not clear please Contact me directly.

Notes

[1] We can tell this from the number of characters the scribe could fit on each line. The lines of text on the outer leaves of P⁴⁶ are approximately one third longer than those on the inner leaves, with a corresponding difference in the number of characters per line.

[2] For comparison, I have in front of me a (single quire) copy of Scientific American containing 29 folded sheets (116 pages), which is one-eighth of an inch thick. I also have an 81-sheet (324 page) magazine which is three-eighths of an inch thick. However, this latter one has multiple small quires and is bound like a large paperback book, so that the pages are nearly all the same size.

[3] These numbers are rough averages. Excluding pages containing the start of a new letter, several pages in the first half of P⁴⁶ only contain 25 lines of text, and one has as many as 30. In the second half of P⁴⁶ the numbers vary from 26 to 32 lines per page.

[4] Even more than with the number of lines per page, the average number of characters per line varies greatly, from as few as 23 in the center of the codex, to 36 in the final third. In some cases, there is a difference of 4 characters per line between the verso and recto of the same leaf.

[5] It is for this reason that the dating of P⁴⁶ is of great interest. In 1936 the MS was dated to 200 – 250 AD. This dating was largely unchallenged until in 1988 Young Kyu Kim proposed a date of the late first century. This radical re-dating has been strongly disputed, but a less radical re-dating to around 150 AD appears to be becoming accepted.