The Gemara (Shabbos 103a) gives a list of letters that should be written very carefully to avoid being mixed up with other letters:

"וכתבתם שתהא כתיבה תמה שלא יכתוב אלפין עיינין עיינין אלפין ביתין כפין כפין ביתין גמין צדין צדין גמין דלתין רישין רישין דלתין היהין חיתין חיתין היהין ווין יודין יודין ווין זיינין נונין נונין זיינין טיתין פיפין פיפין טיתין כפופין פשוטין פשוטין כפופין מימין סמכין סמכין מימין סתומין פתוחין פתוחין סתומין פרשה פתוחה לא יעשנה סתומה סתומה לא יעשנה פתוחה"

While some of these letter pairs look alike (or in some cases according to Rashi, sound the same), others are extremely confusing. So my question is, how do Pei and Tes look like each other? (This could be asked about Gimmel and Tzadi as well, feel free to comment on that pair if you with.)

(Note: I heard a Chiddush in the name of Rav Yaakov Kamenetsky,but I want to hear what others suggest first, as well as to see if others have heard this Chiddush before.)

  • (Also, I realize that many Rishonim deal with it, I just have a very hard time understanding their explanations.) – רבות מחשבות Dec 20 '17 at 19:42
  • The two main theories about gimmel/Tzadi that I remember are that a Gimmel can look like an upside down final Tzadi (if you make the right leg thick enough or the left leg long enough), and that if you point the Yud of the medial Tzadi downwards you get a Gimmel (since Gimmel and medial Tazdi are both just a medial Nun with a thing sticking off the back). – Double AA Dec 20 '17 at 19:43

hand drawn letter tet with colored segments

Here are three ways this Gemara is understood in Rishonim.

  1. The Yereim argues that the problem is if you add the green segment, then the non-blue parts make up a Pei, and that's a problem even though it's only part of a letter looking like another letter. This opinion gets quoted a bunch through to the Acharonim, but it's quite a difficult position since a) lots of letters have parts that look like other letters (eg. Hei has a Dalet in it), and b) there are lots of very old traditions of writing a special "curly" Tet (לפופה) which clearly looks Pei-like.

  2. The Rosh quoting Rabbenu Tam argues that the problem is if you were to rotate the letter counter-clockwise 90 degrees, then even without the green segment the letter would be invalid since it would be a Pei. According to this the pink segment is essential to the letter (assuming a Pei needs its dot to be bent inwards as usual).

  3. The Ramban argues that the problem is to be careful to attach the orange segment to the blue segment since if they split then you end up with two separate Kosher letters: Zayin and Pei. According to this the green segment might be an essential part of the form of a Tet (assuming a Pei needs its dot to be bent inwards as usual).

In short, everyone requires the blue and orange sections in a Tet. Option one allows the pink section but prohibits the green section. Option two requires the pink section and allows the green section. Option three requires all sections.

You are right that none of these answers are super convincing. One argument I've seen defending option 1 from the obvious questions it presents (and I think this argument can be extended interestingly to option 3), is that the concern isn't just that a part of the Tet will look like a Pei, but rather the letter will be written in a way (perhaps with an extended base) such that a child will think it's a Pei which is bumping into something else. We use this sort of metric in other contexts such as to determine if two touching letters can be split without violating Chak Tokhot or if a Lamed poking into a Dalet has formed a Hei. According to this, even the green part is acceptable according to option 1 as long as overall the letter is clearly a Tet. (See R' Kalman Hershlag's discussion of this in an appendix to his book Hasetumah Shebestam.)

Option 2 is also a bit odd since although we have examples of rigidly transformed letters in the traditions of "strange letters" (such as Gen 3:24 or Gen 11:32), they are certainly not common mistakes that the Gemara would need to warn us about them. This option might make more sense if we consider a slightly tipped letter which could occur in somewhat sloppy handwriting. Consider if this image courtesy of @Loewian is a Tet or a Pei:

tipped letter

Finally, it's worth considering simply what could happen if the scribe doesn't pull the left side of the Tet up far enough or alternatively if he tips the end of the Pei up with a bit of flourish. Consider which of the following letters a child would read as Tet and which as Pei:

series of letters with slightly differently sized left sides

  • Thorough answer. However, there are some issues with all of them. As you mentioned, #1 has a number of issues. #2 is a strange answer, because why would we be concerned about flipping letters, the Gemara seems to be giving practical advice to Sofrim about possible pitfalls. #3 is also quite strange, as other not-mentioned letters if not properly attached will become other letters (ches can become 2 zayins). Also troubling for #3 is that why would the Gemara not mention the zayin formed? – רבות מחשבות Dec 20 '17 at 21:00
  • @רבותמחשבות I agree #2 is a bit strange. Some think the Rosh means this because if not for it looking like another letter a sideways letter would be Kosher (think the upside down ן in חרן or upside down ך in דרך). Others think the Rosh is saying this is just an example meant to indicate that all sideways letters are no good. I don't see what's the big problem with option 3. It's by far the most common answer in Rishonim. Your questions can be answered simply that it's obvious to be careful not to split letters to get the basic building blocks ו/ז/י/נ . IAE these are the positions out there. – Double AA Dec 20 '17 at 21:07
  • Comments are not for extended discussion; this conversation has been moved to chat. – msh210 Dec 20 '17 at 22:51

Rav Yaakov Kamenetsky (http://hebrewbooks.org/pdfpager.aspx?req=47587&st=&pgnum=179) addresses the above question. Here is an organized summary of his points:

  • Tes and Pei don't seem to look alike
  • In these cases, the words are written incorrectly in the Gemara. What should be פאין is spelled פיפין. (Also, what should be גמלין are גמין. Rav Yaakov argues that צדין should be צדיקין as well, but let's leave that aside.)

The Rambam quotes this Gemara in Hilchos Tefillin 1:19 as follows:

אין כותבין תפילין ומזוזה אלא בכתב אשורית. והתירו בספרים לכתוב אף ביוני בלבד. וכבר נשקע יוני מן העולם ונשתבש ואבד לפיכך אין כותבין היום שלשתן אלא אשורית. וצריך להזהר בכתיבתן כדי שלא תדבק אות באות שכל אות שאין העור מקיף לה מארבע רוחותיה פסול. וכל אות שאין התינוק שאינו לא חכם ולא סכל יכול לקרותה פסול. לפיכך צריך להזהר בצורת האותיות שלא תדמה היו"ד לוא"ו ולא וא"ו ליו"ד ולא כ"ף לבי"ת ולא בי"ת לכ"ף ולא דל"ת לרי"ש ולא רי"ש לדל"ת וכן כל כיוצא בזה עד שירוץ כל הקורא בהן:

Of note is that Rambam quotes this Halacha along with two other related Halachos:

  1. Writing S"T, Tefillin and Mezuzos in Ashuris script, not any other language (including writing Sifrei Torah in Greek, which was originally allowed, but the Rambam notes that since Greek is now "lost", we cannot use it anymore).
  2. Making sure letters do not stick to each other.

Why would Rambam include point 1 in this Halacha, it seems to be a totally seperate topic? Also, note that Rambam chooses only ו/י, כ/ב, ד/ר, but not the others, and expects us to know from those examples. Rav Yaakov feels that based on Rambam's style he should have quoted ג/צ as well (but not א/ע because those don't look the same, only sound the same).

Based on all of the above, he comes to the following conclusion (as hinted to by the Rambam). The Hebrew letters Tes and Pei are not what the Gemara is referring to! The Gemara is referring to the Greek letters Theta and Phi, which do looked very similar back then, and are still quite similar even in today's version of Greek letters (they are basically a circle with some line through it, as shown in this chart).

Chart of Greek Alphabet

(He also suggests that Gamin and Tzadin are Greek letters, but again, this question focuses on Tes and Pei.)

Therefore, the Rambam begins by talking about that Greek is lost, and as a result does not quote the Greek letters that were in the Gemara. The word פיפין must be used because a Hebrew word always begins with a Dagesh, so in order to signify the soft "PH" sound, it doubles the Pei. (Similar answers are given for Gamin and Tzadin as well. See the actual piece inside for more details, as well as other proofs, notes, and attempted answers.)

ראוי לאמרו, ושפתיים ישק!

For the sake of Emes, I will be posting another answer below with my understanding of the Rambam (what I believe to be the simple Peshat), as well as some other points worth considering.

  • Some Rishonim do think that Alef and Ayin look alike, and it's not just a sound thing. (If you think about it an Ayin is like a dilated Alef that's missing one of the legs.) – Double AA Dec 21 '17 at 18:57
  • Why would the Gemara throw in exactly two examples of Greek letters and stick them right in the middle of the list not even next to each other?? This is a cute idea in the Rambam but I don't see how it really fits in the sources. – Double AA Dec 21 '17 at 18:59
  • @DoubleAA He is basing his opinion on Rashi. & Because Sifrei Torah were written in Greek, it was totally ordinary to be talking about them, just like any other Hebrew letters, so they wouldn't have to be grouped together. – רבות מחשבות Dec 21 '17 at 19:02
  • 1
    Why not also warn us of the dangers of nu, mu and upsilon? Lamda and chi? Sigma, alpha and omicron? Why does it being normal mean they wouldn't be grouped by language? If it was so normal why don't we have any evidence of these writings or see Chazal quoting and Darshining them more? How does he even know that פאין is the "proper" plural? Why would doubling the פ indicate a softer sound when it usually indicates harder sound? Again you can give Terutzim to anything, but it's all really Dachuk. It's much simpler to say the Rambam just picked the easiest to understand for his list of examples – Double AA Dec 21 '17 at 19:06
  • Since I don't know what the old Greek Osiyos looked like, I cannot tell you why some of them were not darshened. I'm not sure that Chazal would have Darshened them, but I would have expected other mentions. He gets פאין from hebrewbooks.org/…, and the doubled Pei would be softer, because the second one makes a soft sound, so you get a Pf sound, which is similar to an F. – רבות מחשבות Dec 21 '17 at 20:50

The Rambam quotes this Gemara in Hilchos Tefillin 1:19 (shown in the previous answer with a link). There he neglects to mention Pei and Tes, because he assumes that you will understand the idea from the other examples given. However, in 7:8, Rambam speaks about being careful when writing the curly Pei letter, likely due to this Gemara, lest it get stretched to look like a Tes, something which was suggested without a source in @DoubleAA's comprehensive answer above. There is a image in his answer which illustrates this point well. (It is also possible that this Halacha in 7:8 is just intuition, and not related to that Gemara as well.)

I strongly argue that the letters, despite their slightly different names, must be Hebrew, as we do not find other places where Greek letters are used in similar contexts, rather, they are referred to as "Yevani" letters (e.g. Kerisus 5b). Also, this Beraisa would presumably be discussing Tefillin and/or Mezuzos as well, which according to Rambam there could not be written in Greek anyways. The list of Hearos on this goes on and on, so if you're still not convinced, check the comment threads.

Feel free to edit this answer for clarity.

  • I thought about the order thing too, but it's no proof because the Greek alphabet parallels the Hebrew ones in terms of order and the Theta is in between Eta and Kappa just like how Tet is between Chet and Kaf. Similarly Gamma is right after Beta like how Gimmel is right after Bet. So the Beraita is in alphabetical order even if one or both of those are Greek. – Double AA Dec 21 '17 at 21:35
  • @DoubleAA You're absolutely right. I missed that. – רבות מחשבות Dec 21 '17 at 21:50

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .