γυμνὸς or γυμνοὶ in Clement's Letter to Theodore?

[RandyHelzerman]

When I look at the image of the character from III 13 with the naked eye, I see what appears to be a slight indentation, if that's the word for it, and I think perhaps AdamKvanta perceives it too. It seems to be preserved in the the early stages of your analysis, both in some of your earlier posts and this one, but then disappears in the third and fourth stage. This makes me think the image manipulation might be manipulating the image a bit too much.

I'd like to do a better job of addressing this point. Lets take a look at the usual suspect:
.

low_vs_hi_rez..png (19.09 KiB) Viewed 485 times

.
On the left is the cropped version from Ken's post, on the right is the hi-quality image. As of old, our primordial bugaboo of bad quality images of dubious origin are hurting us here. The low-res image is *really* low res. You can see the giant pixels. And when we move to the high-quality image, the indent is much more subtle.

i'm trying to get some good way to visualize my next point, but my software is not cooperating......recall, the theory is that under the nib there should be the most ink, and therefore, the areas should be the darkest. So, to prove that the indent is due to ink diffusing, and not due to an extra stroke, we should be able to just progressively remove the lighter pixels, and see if they disappear before the main stroke does.

I made a very bad animated gif of the process here:

gymnastic_iota.gif (17.99 KiB) Viewed 485 times

.
But really, it needs more frames and to be slowed down to really pop.......but if you have photoshop or something, you can just try it yourself. find a way to progressively remove the faintest pixels, and see what you find. Here's what the still image looks like, when you've eliminated all pixels which are fainter than the path I reconstructed:

gymnastic_iota_final.png (6.92 KiB) Viewed 485 times

As you can see, if you delete the faintest pixels---presumably the ones due to ink diffusing away from the nib--you see that the final stroke doesn't have any indents or kinks or anything. Its just a very smooth stroke.

BTW, all of the images in this post are derived from the unprocessed images. I haven't filtered, shaked, or baked them. Just took the original image and started to remove faint pixels. If you have got a halfway sophisticated image processing application, you should be able to just reproduce it for yourself.

[RandyHelzerman]

Ok, courtesy of my very patient wife......who has the mad art skills needed to produce this. She was able to simulate how a pen held at the scribe used would look when writing out the letters.

This is a movie which illustrates why I think I can recover the paths that the scribe used to write the letters.

1. The first few frames overlays a semi-transparent letters as would be written today with a pen held at the same angle and same width as used by the scribe. You can see that the transparent outline lines up nicely with the post-processed image which shows the putative paths of the pen.

2. Next, she's animated the drawing of the letters. Claim is the scribe made very close to the same motions when he stroked out the letters. Since the scribe is using a flat-tipped nib, and holding the pen at a constant angle, the thickness of the letter varies throughout the strokes.

3. Then the opaque letters fade out again, to give you another chance to judge how well they match the reconstructed path. Not bad eh?

4. And finally---the piece-de-resistance!!! I did not show my wife the original images, because I wanted to see if I just threw the reconstructed paths over the wall to her, how well she would do just cold.

So, the original image is made transparent and overlaid with the semi-transparent writing, and then the opacity of increases. Not a bad match for not knowing what they looked like before!!

Also, as the original image fades back in, what you are seeing is the darkest pixels appearing first.....as more and more lighter pixels fade in, you see that the blobs expand---this can be viewed as an illustration of how the ink has diffused outwards of the nib through the paper, getting lighter as it gets more spread out.

EDIT: I should mention this is NOT the gymnastic example--it is one of the other known iotas Ken choose for comparison purposes.

lv_0_20240502130933.gif (8.23 MiB) Viewed 453 times

[RandyHelzerman]

Ok, this is a head-to-head comparison: The gymnastic example vs a known iota. The known iota is the same one in the previous animation.

1. Initial screen overlays translucent versions of the letters which my wife inked, using the putative reconstructed paths asa guide.

2. Then there is a demo of how the scribe would have stroked both of them.

3. The letters fade out again, for another chance at comparison.

4. And then the original images are faded in and out, for direct comparison of the reconstructed path with the original glyphs.

lv_0_20240502145522.gif (10.71 MiB) Viewed 435 times

[RandyHelzerman]

Fellow questers for the historical (or hysterical) Carpocratians, it is striking to me how beautiful the reconstructed omicron is. It was drawn by my wife, who has no particular skills in calligraphy. I saw my wife while she was stroking it in---she just did one take, one circular swoosh, and out popped that beautiful omicron.

I'm wondering if it is *too good to be true*. Is that shape anachronistic for 18-centure greek cursive? If any of ya'll have compiled some exemplars of period writing, can you compare?

You know, this reconstruction of the exact paths can cut both ways. We could do things like reconstruct all the examples which Agamemnon Tselikas gave, and see if his ductile critiques hold up. If we know the path the pen took, can we more accurately and conclusively detect "forger's tremor?"

One interesting thing is that the angle the scribe held the pen at is very consistent angle throughout the letter. Letters drawn on the first page show the same angle of the nib---very much the same. It leads me to suggest that whoever wrote this, wrote it in one sitting. If they would have even put down their pen and took it up again, surely the angle would be further off. And after 3 pages of longhand my hand gets pretty tired, and my handwriting really suffers. If the scribe's hands were hurting towards the end, it doesn't show at all in the angle of his nib. Steady as she goes.

[AdamKvanta]

I was inspired by Peter Kirby's set of iotas and did a set for sigmas. These are not all sigmas but only those that I find somewhat similar to the gymnastic character (40 = III.13 γυµνὸς/γυµνοὶ):

sigmas.jpg (194.19 KiB) Viewed 400 times

1 = I.1 Κλήµεντος
2 = I.5 εἰς
3 = I.12 παντελῶς
4 = I.16 πράξεις
5 = I.18 πίστεως
6 = I.19 εἰς
7 = I.20 εἰς
8 = I.22 εἰς
9 = I.22 ὅµως
10 = I.24 ταῖς
11 = I.26 ἑπτάκις
12 = I.27 φθονερῶς
13 = II.1 ἀσφαλῶς
14 = II.2 πρὸς
15 = II.4 διδαχθεὶς
16 = II.4 τέχναις
17 = II.8 ἀχράντοις
18 = II.9 κράµατος
19 = II.14 πρὸς
20 = II.15 φῶς
21 = II.16 ἔχοντος
22 = II.16 Ἡµεῖς
23 = II.17 φωτός
24 = II.17 πνεύµατος
25 = II.19 τοῖς
26 = II.19 καθαροῖς
27 = II.21 εἰς
28 = II.23 εἰς
29 = II.23 ἧς
30 = II.23 ἀδελφὸς
31 = II.23 αὐτῆς
32 = II.25 ὀργισθεὶς
33 = III.6 εἰς
34 = III.6 πλούσιος
35 = III.7 γενοµένης
36 = III.8 πρὸς
37 = III.11 εἰς
38 = III.11 τούτοις
39 = III.12 Ἰάκωβος
40 = III.13 γυµνὸς/γυµνοὶ
41 = III.14 εἰς
42 = III.18 ἀληθὴς

I also made a comparison of some sigmas with the III.13 character and my conclusion is that the character might be a sigma because there are similarities between them.

comparison.jpg (25.99 KiB) Viewed 400 times

Other arguments are that I think there is a slight indentation as Ken Olson said and the lower part is thinner than the upper part which is not something we see in iotas. Also if the iota is curved, it usually starts with a top left hook. But I see no indication of a hook there.

These arguments are not a proof though. It might be an iota too. My point is that I don't think it was proven (Q.E.D.) that it is an iota.

[RandyHelzerman]

[simas]

Thanks for compiling those, after a few days of this I know it’s hard work. I’m looking forward to some fun times analyzing them!

Did you experiment around any more with digital signal processing techniques? For sure there is no such thing as being for sure, but it would really be nice to at least come to some consensus.

If anybody else wants to try them out, the software I’m using (gimp) is free. After reading a tutorial or watching sone YouTube videos—which there is an endless amount of—you can easily reproduce what I have done. I’d be happy to help anyway I can, and would be glad to collaborate.

Those of you at universities, go find a colleague or two in the engineering departments. Every assistant professor on the planet is looking for some way to showcase the techniques they are working on.

BTW, the black-and-white images from Smiths photos are very high res—and were taken when the document hadn’t suffered from being moved all around and mutilated. They are complimentary to, but are not made otiose by the color photographs.

This kind of back-and-forth is in the best tradition of scholarship, and it’s pushed me to think and rethink. It gives me pause that Smith—one of the few who have seen the actual document—translated it as a sigma. Surely those two words would have caught his eyes, and he would have looked very closely at them.

*even if* there is no slam-dunk proof possible, well that cuts both ways. If you do find it so very uncertain what it us, that means you can’t really use it to support any other contentions you make. If it is REALLY important to the rest of your program exactly which letter this is, that should make you REALLY motivated to reach some consensus, because it doesn’t have to be slam-dunk proven ether way to render it’s true identity powerless to help you advance any argument that depends on it.

Please think about what would it take to reach consensus and let’s see if we can provide it.

[RandyHelzerman]

Last post I mentioned that the Black and White scans of Smith's photos are also very useful. They are black and white, but they are very high-res.

When I run my sober-operator + viewing he darkest pixels, I get much the same result with either set of images:

Screen Shot 2024-05-02 at 10.59.22 PM.png (62.57 KiB) Viewed 380 times

Left is from Smith's Photos, right is from the high-res images supplied by Peter Kirby. As you can see, I get much the same result for the best guess of the path of the pen. NOTE: in the trade, this is a very good sign. If your technique gives the same result no matter whether its from color, or black and white, high-res or low-res, that is a sign that it is really getting to the essence---what all those version have in common with each other.

But NOTE!!! You do not even have to do fancy image processing on Smith's photos to get the same result. Check this out:

smith_unprocessed.png (139.82 KiB) Viewed 380 times

On the left is Smiths B&W image of the gymnastic character, zoomed in so you can see the pixels. Even looking at that unprocessed, zoomed in image, you can see that a some areas are darker than others--and those areas are very similar to my reconstructed pen path.

On the right, All I have done is taken the darkest pixels, and turned them to black, and the rest of the pixels, and turned them to while. Out pops the same damn path.

[RandyHelzerman]

[deleted--i really hate tapatalk forums]

[RandyHelzerman]

Other arguments are that I think there is a slight indentation as Ken Olson said

Take a look at Smith's Black and white photoes--the indentation is not apparent. Also, zoom in on the HIGH RES unprocessed gymnastic character--you'll see that where the "indentation" is is very much lighter ink than the rest of the blob. Indicating, to me at least, that this is just an uneven diffusion of ink though the paper.

and the lower part is thinner than the upper part which is not something we see in iotas.

cf. the known iota in my previous animated post with the omicron-iota. That iota certainly was wider on top. Also, any stroke which is written with a flat-nosed pen ---which is roughly in the shape of a comma or a right parenthesis--is going to be fatter on the top than the bottom, that's just how those kind of pens work. Again, cf my omicron-iota animation.

Also if the iota is curved, it usually starts with a top left hook. But I see no indication of a hook there.

The glyphs are "multiform"--which I guess is a nice way of saying inconsistent, and all over the place

Anyways, in appreciation of your collection of sigmas, before I have to run off to work, let me post here an animation of your sigmas plus several image enhancement techniques I've been experimenting around with. Most of them have no problems revealing the sigma-y nature of them---unless they are just a complete ink blob.

sigmas_slideshow.gif (2.7 MiB) Viewed 363 times

[AdamKvanta]

Other arguments are that I think there is a slight indentation as Ken Olson said

Take a look at Smith's Black and white photoes--the indentation is not apparent. Also, zoom in on the HIGH RES unprocessed gymnastic character--you'll see that where the "indentation" is is very much lighter ink than the rest of the blob. Indicating, to me at least, that this is just an uneven diffusion of ink though the paper.

and the lower part is thinner than the upper part which is not something we see in iotas.

cf. the known iota in my previous animated post with the omicron-iota. That iota certainly was wider on top. Also, any stroke which is written with a flat-nosed pen ---which is roughly in the shape of a comma or a right parenthesis--is going to be fatter on the top than the bottom, that's just how those kind of pens work. Again, cf my omicron-iota animation.

I think the gymnastic character has an unusual top part for an iota. It has a slight vertical elongation which creates the indentation effect. I see this effect even without the lighter ink pixels. But then again, it's very slight and I think it doesn't prove anything.

And it's true that the I.9 iota in συµφωνοίη is slightly wider at the top but I think the thickness difference is greater in the gymnastic character. The other thing is the ratio of the dark/light pixels between the two parts. In the I.9 iota, the ratio of the upper part is comparable with the lower part. On the other hand, I see many more dark pixels in the upper part of the gymnastic character than in the lower part.

Anyway, I tried GIMP and paint.net edge-detection algorithms and I'm not sure if it can really help. I'm afraid it can rather distort the images. I think we both agree that a sigma that looks like a complete ink blob can't be deciphered by any of these methods. And if a scribe produced sigmas that look like blobs, is it really possible to rule out that the gymnastic character is a sigma consisting of a blob and a curve?