γυμνὸς or γυμνοὶ in Clement's Letter to Theodore?

[Peter Kirby]

These iotas once again here.

[Peter Kirby]

How about this?

I don't think untrained use of image manipulation programs is the answer.

[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.

[RandyHelzerman]

How about this?
sigma.png

Hey Adam, thanks for reading all those posts, and thanks for trying your hand at it. Recall, the path the pen takes is one of those things which is inferred from the data. But I didn't talk a lot about how you *do* find the most probable path. Let me try to fill the lacuna.

Lets zoom in on the unprocessed images:

usual_suspects.png (22.75 KiB) Viewed 397 times

Think of how the ink diffused from the top of the nib though the paper......so the darkest pixels will be the most likely ones where the nib went. Even in the unprocessed images, if you look at the general shapes the darkest pixels make, you can kinda discern something sigma-y.

Now lets look at the transformed images:

transformed_usual_suspects.png (28.94 KiB) Viewed 397 times

Same thing holds in the transformed images. The pixels are dark here and the colors are close together so there's not a lot of contrast. But lets take the sigma, and lets color in the darkest pixels and see what we get:

line.png (27.61 KiB) Viewed 397 times

So that's how I got the results before.

Like I said, I don't want to prejudice the issue, so I won'd do the same for the gymnastic example---try it yourself and see what you get. Also, I gave two different edge finding results in the previous post. The second one has a thicker boarder, so you can look at it for clues as to how the nib of the pen goes.

Thanks again...took me ages to write up, its nice to know somebody took the time to read it.

[Ken Olson]

Like I said, I don't want to prejudice the issue, so I won'd do the same for the gymnastic example---try it yourself and see what you get.

I do not think this answers the point AdamKvanta raised. If by 'gymnastic example' you mean the letter under discussion on III 18, I think you have to discuss that.

At present, I think what you've shown is that a letter everyone in the discussion accepts as a Sigma is indeed Sigma-y.

If you want to argue that the letter under discussion in III 18 is an Iota, I think you need to show that it's Iota-y, and more Iota-y than Sigma-y.

Also, I think you should discuss the Iota in συμφωνοίη on I 9 (which I believe is Peter Kirby's Iota #11), which seems to me to be the only Iota in Peter's list which is similar in appearance (or not dissimilar in appearance) to the disputed letter on III 18.

Best,

Ken

[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.

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 (32.45 KiB) Viewed 382 times

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

[RandyHelzerman]

I don't think untrained use of image manipulation programs is the answer.

Oh I would actually encourage people to play around with this. I’m just using an off the shelf image processing software (gimp) and applying edge detection filters.

Adam I don’t have the same image processing program that you do—it looks, though, that the filter you are using reverses brightness inside of the outline. So the brightest pixels would be most likely to be on the path of the pen.

You did get a nice outline/contour of the blob, but the filter seems to make a lot of the pixels inside to be almost the same color.

So, if you can you can enhance the contrast, it might help too. But ideally, if your software has a “sobel operator” filter, that would give firmer results, because the sobel operator was designed specifically for computerized vision.

if we use the exact same filter, we should get the exact same results from the exact same input, but---as I illustrated by providing two different results from two different outline filters--the results should be not so very sensitive to filter choice. However, filters designed to make me look like Kim Kardashian on my instagram page might make make it harder to extract the most likely path Use the sobel filter if at all possible.

[RandyHelzerman]

I do not think this answers the point AdamKvanta raised. If by 'gymnastic example' you mean the letter under discussion on III 18, I think you have to discuss that.

You are right:-) I hope my followup post filled that gap. If there is anything in particular you'd like me to address, point it out and I'll do my best.

At present, I think what you've shown is that a letter everyone in the discussion accepts as a Sigma is indeed Sigma-y.

Cool!!! It certainly wasn't obvious just from the initial photos, was it? But these days we have some cool algorithms to really help.

If you want to argue that the letter under discussion in III 18 is an Iota, I think you need to show that it's Iota-y, and more Iota-y than Sigma-y.

My goal is to explain this in enough detail that anybody who follows the same steps should find that it is pretty easy to pick out the most likely path. If I do the analysis myself on the gymnastic example, I won't know that I succeeded But if lots of people try it out, and they all get similar results, we can have some confidence that I explained it in enough detail.

I'm hoping to illustrate how they work on the sigma example, or even examples of other letters. Since we all agree that these steps have revealed the sigma-y-ness of the sigma example, we can have confidence that if any two people follow the same steps, it will reveal the identity of the gymnastic example as well.

Also, I think you should discuss the Iota in συμφωνοίη on I 9 (which I believe is Peter Kirby's Iota #11), which seems to me to be the only Iota which is similar in appearance (or not dissimilar in appearance) to the disputed letter on III 18.

You mean #11 on this chart?

oi-1.jpeg (79.63 KiB) Viewed 359 times

If not, if you could just mark the ones you want to compare on the high-resolution images from Hendrick et al., I'd be happy to.

[Ken Olson]

You mean #11 on this chart?
oi-1.jpeg

If not, if you could just mark the ones you want to compare on the high-resolution images from Hendrick et al., I'd be happy to.

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

Iota I 9.png (507.63 KiB) Viewed 350 times

Best,

Ken

[Ken Olson]

And the character in dispute from III 13:

Iota or Sigma in III 13.png (54.43 KiB) Viewed 341 times

Best,

Ken