Monday, September 17, 2007

Unidentified Manuscript Fragment 01

Here is a mystery.

I've been working again on some new manuscripts and this small fragment still has me stumped.

It came from a collection of small pieces which turn out to be mostly biblical (1 OT; 5 NT so far identified); but other than some ideas about the letters and a possible colour match with one of the NT pieces, I can't find a fit.
So feel free to solve the mystery. I shall give you full credit. You may "discover" a new piece of the Bible.














26 Comments:

Timo Flink said...

Peter, perhaps it would help to see them in ultraviolet light? It might reveal more letters.

Ulrich said...

It's two columns isn't it, the central part of the fragment being an intercolumn. In that case ultraviolet light might hardly provide promising results.

Ulrich Schmid

Peter M. Head said...

Thanks Timo,

These already are. 40 min UV exposure. Without that there is not very much to see.

Timo Flink said...

Thanks Peter and Ulrich. Well, it's a though one (for me at least). Would a color image be any more information? I'm a total novice in deciphering fragments that badly preserved.

Timo Flink said...

I meant "more informative?"

Eric Rowe said...

I just did a Bibleworks search of the BGT (LXX+NT) for the string *psin* followed by *as* within 25 words (pardon the transliteration--I'm not on a computer with Unicode at the moment). As best as I can tell, those are the strings of letters you can see on the right side of the bottom image. The search brought up 27 verses. So then I looked through them to see if any had a letter with two vertical strokes (pi, eta, nun) following the sigma. Only two verses still fit the criteria, Gen 42:21, where the vertical strokes would belong to an eta but with a column width that is unreasonably big, and Judith 4:13, where they would belong to a pi with a column width of ~30 letters.

Unfortunately, I thought the search command I used would not be limited to within single verses, but the way the results came up makes be believe I was wrong about that.

Eric Rowe said...

The letters on the left side of the lower image are not very helpful. It looks like an epsilon on top, preceded by something I can't identify, and nun kappa alpha on the bottom, which would most likely be a word ending in nun, followed by a word beginning with kappa alpha. Since nun is the last letter of several very common noun and participle conjugations, and two extremely common words begin with kappa alpha, I don't think that part of the image will narrow anything down. Is there anything somewhat distinctive on the top image that can be identified well enough to use as a test for the hypothesis that the right side of the bottom image is from Judith 4:13?

Timo Flink said...

Eric, I think you did it. I counted the letters and if it is Judith 4:13 LXX, then the line width is approx. 30 letters. If you look closely, I think there is a dim vertical line revealing Π after ΑΣ. So yes, IMO, it is Judith 4:13.

... θλιΨΙΝ ... ημερΑΣ Πλειους ...

Makes sense. Good work, Eric.

Timo Flink said...

btw, Peter, what are these new fragments on NT? Any text-critical things of interest?

Peter M. Head said...

Thanks Eric,

I'm not fully convinced by your starting point though: PS I N
a) your proposed PSI could be a RHO (cf. the rho on the upper photo) with something else to the left (damaged) [I'm about 40-60 on this with preference for rho on the basis of the other example]
b) your proposed NU lacks the diagonal downstroke (cf. the clear nu in the left-hand column), evidence of the horizontal suggests HTA.

Peter M. Head said...

Two other problems with Jud 4.13:
c) Assuming for the moment that PSI is correct, then is there not a circular letter before it? But Jud 4.13 is thlipsin.
d) A line width of 30 letters seems unlikely in a multi-column layout.

Eric Rowe said...

If I (or somebody) can find a way to do the same Bibleworks search without limiting the results to single verses, that would definitely be an improvement. There are probably other options out there that didn't fit the criteria I used because the "AS" is in a different verse than the "PSIN". So if we can do that, then we can try that search with both rho and psi and see what we get. Unfortunately, I think I'd have to fiddle around with Bibleworks for awhile to figure out how to do that search properly, and I don't have time at the moment. But I'm sure there's a Bibleworks or Accordance guru out there who can do it in a flash.

On the column width, I was also thinking that 30 letters seems to be on the wide end of the spectrum. But since we don't know anything else about the codex, it's definitely within the realm of possibility, even for a two-column codex. Am I wrong?

Eric Rowe said...

I have to admit, Peter, your hypothesis of rho instead of psi is probably right. If it is a psi, then it's awfully asymmetrical. I wasn't aware that that was a way to write rho, and I didn't know what to make of that rho on the top image.

Timo Flink said...

Ok. Peter has a good point. Here's another suggestion. If it has ΡΙΝ and ΑΣΠ, then the verse could be Judith 1:6, where Bibleworks LXX has

ΤιγΡΙΝ και τον ΥδΑΣΠην

Now the line length is only 11 letters. Is this any better?

If the last letter is HTA instead of NU, Bibleworks does not find any match. If it is GAMMA, the matches have the words too far apart (the same applies to TAU).

I can see why this one is "a nut to be cracked" as we say in Finnish :)

Peter M. Head said...

The basic reason is that there are not enough securely identified letter strings in close proximity. And if you leave it too broad you get too many results.

Peter M. Head said...

Eric said:
"On the column width, I was also thinking that 30 letters seems to be on the wide end of the spectrum. But since we don't know anything else about the codex, it's definitely within the realm of possibility, even for a two-column codex. Am I wrong?"

I don't think you are wrong necessarily. But it would be unusual enough to make me wonder if the identification was correct.

Getting a secure identification will have to involve fitting in the observable traces, esp in the left hand column.

Peter M. Head said...

Timo asked:
"Would a color image be any more informative?"

I don't really know what a colour UV image would look like; since UV is obviously below the wavelength of visible colour.

I suspect that some proper variable frequency or multi-spectral imaging might show up a bit more. And even a letter or two could be potentially decisive.

Peter M. Head said...

Eric said:
"If I (or somebody) can find a way to do the same Bibleworks search without limiting the results to single verses, that would definitely be an improvement."

I generally use the TLG as the proximity is not limited to particular verses. If BW can't do this then it is definitely going to miss some potential results.

jonathanclarkborland said...

My speculative and highly entertaining guess is:

Top image, left col.: Mt 8.17
Ln 1: ending in [PL]HRW ...
Ln 2: ending in [RHQEN DIA] HS

Top image, right col.: Mt 9.2
Ln 1: ...LU... (from PARALUTIKON OR PARALUTIKW)
Ln 2: (you figure out this one, hahaha)

Bottom image, left col.: Mt 9.17
Ln 1: ending in [MH]GE
Ln 2: ending in [ASKO]I KA[I]

Bottom image, right col.: Mt 9.35
Ln 1: beginning with [E]RIH[GEN]
Ln 2: beginning with AS P[OLEIS]

Thanks for the diversion.

Jonathan C. Borland

Timo Flink said...

Peter, I was thinking a multi-spectral image. With a proper computer software, it turns to be a color image. It might reveal some more.

Jonathan, very interesting. I made a typical mistake for me by not separating ΑΣ from Π, when looking for other possibilities.

Peter M. Head said...

Jonathan,

Thanks for this. You could be on to something here. I don't buy all of it. E.g.

Bottom image, left col.: Mt 9.17
Ln 1: ending in [MH]GE
Ln 2: ending in [ASKO]I KA[I]

I don't think iota is most likely for that letter. AND, this leaves 33 lines of NA26 text in between the top of the two columns while each line has only about 18 letters (well different line lengths is another inconsistency here). So two columns with around 70 lines of text: not an obvious format.

BUT I really like:
Bottom image, right col.: Mt 9.35
Ln 1: beginning with [E]RIH[GEN]
Ln 2: beginning with AS P[OLEIS]

Because this fits with the most natural reading of each letter (subjective I know). I am gonig to try to work back from this to check out other bits of Matt 9.

Peter M. Head said...

So, taking as a starting point:
Bottom image, right col.: Mt 9.35
Ln 1: beginning with [E]RIH[GEN]
Ln 2: beginning with AS P[OLEIS]

This gives a line length of 11 letters (assuming NS).

There is a reasonable match for the left hand column:
Mt 9.29
Ln 1: ending with UMWN] GE
Ln 2: ending with UMI]N KA[I

Line length: 12 letters.

Peter M. Head said...

This could be getting somewhere. There is a bit of an issue with somewhat unusual word breaks (e.g. needed for PERIHGEN at 9.35 and for KAI at 9.29), but in terms of letter identification both of these seem pretty good.

Ulrich said...

Peter Head:

"There is a bit of an issue with somewhat unusual word breaks (e.g. needed for PERIHGEN at 9.35 and for KAI at 9.29)..."

I would add T/AS to the unusual word breaks, too.

Ulrich Schmid

Peter M. Head said...

I'm still not happy with these proposed word divisions (thanks Ulrich by the way).

There is a possible match for the something-rho-omicron letters in Matt 9.17:
SUNT]HRO
UNTAI]
but I can't make much of the second line, and this is also an odd way to divide a word.

Peter M. Head said...

Lacking any further inspiration my current plan is to publish this piece with the rest but as unidentified. Put some features of the Borland [rev by Head] proposal into a footnote. Recommend that the owner might try to get some MSI, as even one or two additional certain letters would mean a lot.
Sound reasonable to you?