γυμνὸς or γυμνοὶ in Clement's Letter to Theodore?

[RandyHelzerman]

How about this?

Ok, Adam, this is what I mean by enhancing the contrast. I loaded your image into gimp, then brought up the colors->level tool, and then just cranked the "output levels" setting all the way to zero. It has the effect of making the brights brighter, and the darks darker (like all-temperature cheer!)

other_filer_adjusted.png (85.02 KiB) Viewed 239 times

If you zoom in, you'll find that the top line segment takes a route over some pixels which are darker, so I would assign this path a lower probability.

[Peter Kirby]

The manipulated images are capable of minimizing visual information that one's eye and brain can pick up to tell whether edges are created by a break in direction of the stroke or other factors, including especially the thickness of the line created by the ink of the pen, which in the case of the character under discussion appears to taper from wider at the top to thinner at the bottom, perhaps with some variation along the way.

The non-modified images can also be read incorrectly, but at least they are a better basis for discussion, especially when drawing on relevant training for understanding what it shows about the direction of the stroke of the character.

I have always agreed here that we would benefit from relevant insights from experts. Ultimately the desired outcome would essentially be to have these relevant insights informed by all the relevant data and then presented in a journal article outlining the relevant considerations and addressing any objections.

A discussion forum like this is capable of getting ahead of its skis and throwing up ideas that may not hold up to closer examination, such as that a paint program can produce true conclusions here and not merely illustrate those true conclusions that are already justified.

It's also capable of generating ideas that may be deemed inconclusive, at least based on the present state of discussion here in this thread. As I previously mentioned, I am willing to view the placement of the accent this way. But I should mention that the accent was considered relevant by those who have more experience and expertise than I do, and I agree that I should follow up.

I have heard from a couple experts here myself. They have noted (a) that the character here starts out going southeast/northwest [atypical for sigma], (b) that the stroke is unbroken [not observed in sigmas], and (c) that the accent placed directly over the character is indicative in this case.

To examine just (c) at the moment, why is the accent placed directly over the character indicative in this case?

Are there not other cases where the accent is placed to the right of the accented vowel? I have previously pointed to the accent being placed over the Sigma in εὐθὺς on III 2 (second to the last word), while the accent over the Nun in τὸν in the same line III 2 (second word) is perhaps an even clearer example.


Accent over Nun.png


Have your informant(s) been over the entire document and determined how the scribe decided whether to place the accent over the accented vowel or the letter to its right? And if not, why should we accept the claim that the accent placed directly over the character is indicative in this case when it is not in others?

Best,

Ken

A few people were asked for a few moments of their time. None of them thought it looked like a sigma, and anyone who commented thought it looked like an iota. In all, fifteen people have commented that it looks like an iota.

They were provided photographs of all 3 pages and given no guidance on how to comment. So they were not given a transcription with sigma, they were not informed of the opinion of Secret Mark authors, they were not provided with image manipulation program outputs, they were not instructed to focus on particular details or comparanda, etc.

On the three points, which were mentioned at their initiative, different points were mentioned or emphasized by different people who responded.

Tselikas performed a detailed study informed by relevant expertise not only on this point but of the entire transcription, being the only one with relevant expertise who has substantially investigated the accuracy of the received transcription that I know about. On the question of the transcription here, Tselikas found that it was an iota.

Tselikas doesn't go into much detail in writing on his conclusion here. I agree that what we want is an informed study, not only drawing on experts but also bringing in all the relevant considerations. In this thread we have seen some of the things that will be considered as throwing doubt on the subject. If everyone or at least most people with an interest in the subject have been working with a different assumption for a long time, then it's relatable that they would want to have the explanation of why it is to be understood as an iota to be thoroughly tested. It's also understandable that the facts that indicate in favor of this conclusion may not be readily understood that way, without a more developed apparatus here.

I don't expect everyone involved here to accept any of this right now. I believe we need the discussion to develop more fully before we have reached the point where everyone - even the Secret Mark experts - is capable of recognizing the point that it's an iota. Those who have offered publicly interpretations assuming it's a sigma may be the slowest to turn around.

I am appreciative of the counter-point on the accent you brought up on the thread. I would have to follow up with them to have a sense of why they think the accent placement is relevant and whether they think its relevance is (a) conditional on additional facts or (b) probabilistic and supplementary to the other two points mentioned.

I appreciate the objections that you're bringing up in this thread, Ken. They are exactly the kinds of points that should be considered.

[RandyHelzerman]

Here is perhaps a better quality image of the Iota from συμφωνοίη in I 9

Yeah that image was very high quality. Here's what I get:

ken_iota.png (22.11 KiB) Viewed 212 times

From left to right: original cropped image, sobel-transformed image, and the top few percent darkest pixels.

Recall, the darkest pixels will be awarded highest priority to be on the path of the pen. The sobel-transformed image unfortunately also has very similar colors, so its hard to pick out by eye. Not impossible, just hard. To get rid of the subjectivity in deciding which pixels are the darkest, I had gimp compute the top few percent darkest pixels. The path the pen took is pretty apparent, wouldn't you say?

[Ken Olson]

Here is perhaps a better quality image of the Iota from συμφωνοίη in I 9

Yeah that image was very high quality. Here's what I get:

ken_iota.png

From left to right: original cropped image, sobel-transformed image, and the top few percent darkest pixels.

Recall, the darkest pixels will be awarded highest priority to be on the path of the pen. The sobel-transformed image unfortunately also has very similar colors, so its hard to pick out by eye. Not impossible, just hard. To get rid of the subjectivity in deciding which pixels are the darkest, I had gimp compute the top few percent darkest pixels. The path the pen took is pretty apparent, wouldn't you say?

Could you please do the same thing with the higher quality image of the disputed character from III 18?

Best,

Ken

[RandyHelzerman]

Could you please do the same thing with the higher quality image of the disputed character from III 18?

Sure, however, the sample for III 18 you gave in the last post (viewtopic.php?p=172091#p172091) was a lower quality image than the one for 1 9. (zoom in, you'll see the pixels in II 18 are much bigger).

If you don't mind differences in quality, I've done that example in many previous posts. But if you want a true apples-to-apples comparison, you can send me a higher quality version. Or I can just cut them both out of the high resolution images, if you prefer. If I do that, I get:

head_to_head.png (36.67 KiB) Viewed 182 times

My laptop battery is dying, so I won't be able to reply in real time again until tonight

[Ken Olson]

Could you please do the same thing with the higher quality image of the disputed character from III 18?

Sure, however, the sample for III 18 you gave in the last post (viewtopic.php?p=172091#p172091) was a lower quality image than the one for 1 9. (zoom in, you'll see the pixels in II 18 are much bigger).

If you don't mind differences in quality, I've done that example in many previous posts. But if you want a true apples-to-apples comparison, you can send me a higher quality version. Or I can just cut them both out of the high resolution images, if you prefer. If I do that, I get:
head_to_head.png

My laptop battery is dying, so I won't be able to reply in real time again until tonight

My apologies. I must have attached the wrong image file. Here is (hopefully) a better quality image of the disputed character from III 13:

Iota or Sigma III 13.png (655.64 KiB) Viewed 171 times

Could you run it again with this image? I can certainly wait until tonight to see the results.

Thanks,

Ken

[Peter Kirby]

But if you want a true apples-to-apples comparison, you can send me a higher quality version.

For the purpose of apples-to-apples comparison, I have sent you a Google drive link with the (circa 300MB each) bitmap files of all three pages. I have additionally sent the same invitation link to Ken Olson, Andrew Criddle, and Adam Kvanta. If someone reading this also wants a link, please e-mail me at peterkirby@gmail.com for that.

These images do not come with a license (they're not mine). They come with some advice; they can be used privately for research purposes. Individual characters or small segments (and derivatives of the same) can be reproduced appropriately, for comment and analysis, in accordance with the "fair use" doctrine. What this means is that you don't have permission, for example, to publish the entire image. But every other use in this thread so far would almost certainly satisfy the "fair use" exception.

You will notice that I have conveniently segmented the images 5 lines at a time, for easier reference. These derivative cropped images were produced in a lossless way and saved to PNG files. Republication of these files of 5 lines at a time also might be an infringing use if not cropped further and accompanied by analysis.

A bit of the oral history of the images: there are two sets of images here. The common history of the images involves the photographer at the employ of Quesnell and a technician assisting Hedrick for a publication in The Fourth R in 2000. The principal divergence is in the transition from analog to digital. One set of images is published by Westar as a PDF of their 2000 issue that is currently available to anyone with a $10 subscription. I have used the pdftoppm software package to convert the three relevant PDF pages to PNG without any change in resolution, loss in quality, or any other change from the image information in the PDF. The other set of images - the bitmap files - was sent from Charles Hedrick to Scott Brown, from Scott Brown to Roger Viklund, and from Roger Viklund to me. The most obvious difference is color and lighting, but that isn't unusual based on my understanding of the process of image creation. The bitmap files eclipse the PDF images in image resolution entirely.

[Peter Kirby]

I have heard from a couple experts here myself. They have noted (a) that the character here starts out going southeast/northwest [atypical for sigma], (b) that the stroke is unbroken [not observed in sigmas], and (c) that the accent placed directly over the character is indicative in this case.

A few people were asked for a few moments of their time. None of them thought it looked like a sigma, and anyone who commented thought it looked like an iota. In all, fifteen people have commented that it looks like an iota.

They were provided photographs of all 3 pages and given no guidance on how to comment. So they were not given a transcription with sigma, they were not informed of the opinion of Secret Mark authors, they were not provided with image manipulation program outputs, they were not instructed to focus on particular details or comparanda, etc.

On the three points, which were mentioned at their initiative, different points were mentioned or emphasized by different people who responded.

I agree that I should follow up.

In the interests of full disclosure, one follow-up to the initial query, asking someone who initially demurred from offering an opinion, reveals that not everyone believes the accent placement is significant:

It could be "γυμνοί" (plural), but it might also be "γυμνός" (singular) if the accent was misplaced due to speed, a common occurrence with other words as well.

So far the query has been met with non-response, not providing an opinion that is in favor of one conclusion or the other, and the fifteen comments saying it is an iota. None so far have said that it is a sigma.

Obviously there are limitations of this kind of polling, and opinions even from the same people could evolve as they are introduced to more information, e.g. with the presentation of additional comparanda, examples of the placement of accents elsewhere, or possibly thoughts on the subject of ligatures (nobody commented on ligatures unprompted).

It does seem to provide support to the idea that the iota conclusion is clearly the most obvious one to those coming in (as much as is possible here) blind regarding the subject of SGM but informed regarding the subject of Greek and reading Greek manuscript texts. I will say that "most obvious" doesn't always mean "correct," but it also doesn't count for nothing.

[Ken Olson]

I have heard from a couple experts here myself. They have noted (a) that the character here starts out going southeast/northwest [atypical for sigma], (b) that the stroke is unbroken [not observed in sigmas], and (c) that the accent placed directly over the character is indicative in this case.

A few people were asked for a few moments of their time. None of them thought it looked like a sigma, and anyone who commented thought it looked like an iota. In all, fifteen people have commented that it looks like an iota.

They were provided photographs of all 3 pages and given no guidance on how to comment. So they were not given a transcription with sigma, they were not informed of the opinion of Secret Mark authors, they were not provided with image manipulation program outputs, they were not instructed to focus on particular details or comparanda, etc.

On the three points, which were mentioned at their initiative, different points were mentioned or emphasized by different people who responded.

I agree that I should follow up.

In the interests of full disclosure, one follow-up to the initial query, asking someone who initially demurred from offering an opinion, reveals that not everyone believes the accent placement is significant:

It could be "γυμνοί" (plural), but it might also be "γυμνός" (singular) if the accent was misplaced due to speed, a common occurrence with other words as well.

So far the query has been met with non-response, not providing an opinion that is in favor of one conclusion or the other, and the fifteen comments saying it is an iota. None so far have said that it is a sigma.

Obviously there are limitations of this kind of polling, and opinions even from the same people could evolve as they are introduced to more information, e.g. with the presentation of additional comparanda, examples of the placement of accents elsewhere, or possibly thoughts on the subject of ligatures (nobody commented on ligatures unprompted).

It does seem to provide support to the idea that the iota conclusion is clearly the most obvious one to those coming in (as much as is possible here) blind regarding the subject of SGM but informed regarding the subject of Greek and reading Greek manuscript texts. I will say that "most obvious" doesn't always mean "correct," but it also doesn't count for nothing.

I wrote earlier in this thread:

The problem is we have very few published works since Tselikas that consider the two possible readings. By that I mean scholars who publish under their own names their considered opinion along with a discussion of how the came to their conclusion (ligatures, comparanda, etc.). I would be very happy to consider such if they were available.

If I knew nothing else about The Letter to Theodore and you had sent me images of the three pages and asked me what the word in dispute on III 18 is, I would almost certainly have said it was γυμνοί with a final Iota because one of basic rules of Greek is that accents go over vowels. That's the quick and easy answer. If you sent me the same images of the three pages and the same question along with the the statement from Venetia Anastasapoulou's report that the accents in the manuscript are placed either over the letter or to the right of the letter in and provided me with comparanda such as εὐθὺς and τὸν on III 2, in which the accent is placed over the letter following the vowel, my immediate response would likely have been that I can't be sure what the disputed word is, I would need to look at it some more.

I know you wanted to simulate something analogous to a 'veil of ignorance' so as not to skew the answers in favor of Smith's opinion (as I take it you think has happened in the field), but I think you have skewed the results in the opposite direction. I would weigh the opinion of scholars who have addressed the entire manuscript (Smith, Tselikas, Pannanen and Viklund, and Adams) more heavily than those were asked for and/or gave an opinion on the single word under consideration or just a few words.

Best,

Ken

Added note: I once asked Ulrich Schmid to look at a reading I was interested in and justified it by saying I wanted the opinion of 'an expert on textual criticism'. I thought he'd be flattered that I had so high an opinion of his expertise. His reply was something like: 'You can't be an expert on textual criticism. You can only be an expert on a particular manuscript'. (He did give me his opinion on the reading though).

[RandyHelzerman]

Could you run it again with this image? I can certainly wait until tonight to see the results.

Wait no longer....feast your eyes upon:

ken_hi_res.png (365.15 KiB) Viewed 74 times

Ok, we have both requested blobs (can you tell which is which? Top is the gymnastic example, bottom is the other iota)

From left to right we have:

1. Unprocessed blob--from the hi-res images which Ken requested.

2. Sobel operator

3. Threshold the image to display the top few percent darkest pixels

4. With images this good, we can actually even detect the angle which the scribe held his pen at!! If you've done calligraphy with a squared-off nib--or better yet, cut yourself a quill pen from a feather--you will know, you hold the pen at an angle, and then as you move around, sometimes the line gets thicker, sometimes thinner.

The red lines are just the same red line cut and pasted over and over again. They represent the nib position and angle at various points during the stroke. As you can see, the scribe held the pen at a constant angle, and then made a motion like a large comma.

Both letters were drawn with the same pen, held at the same angle, in a single stoke. Time for me to stop pussy-footing around: if one of them is an iota, the other one is too. Q.E.D.