Carving out a letter invalidates (ḥak tokhot).
1. We may further learn from the word u-khtavtam the requirement that every letter be made by writing, particularly, and not by means of scraping – that is, by means of ḥak tokhot. For example, if one scraped out the middle of a letter and its sides, and in so doing found that there remained a letter with the correct form, it is invalid because of ḥak tokhot. Therefore, if one erred and wrote dalet instead of reish, or beit instead of khaph, or vice versa, he cannot fix it by scraping off the protrusion, because this is ḥak tokhot.
2. Even if he only formed part of the letter by ḥak tokhot, it is still invalid (see Excursus 2). Therefore, if one started and formed part of the letter correctly, and then it became spoiled by his error or from a blot falling on it and a small part of the letter was defective – even if it was only a tiny point defective, if it was a point without which the letter was invalid – and he scraped off the erroneous part and below, and after that finished the letter by writing, it is also invalid. For since at the start the entire letter was spoiled by the damage, and only by scraping did the form of part of the letter remain, this is also called ḥak tokhot, and it is invalid – rather, he must remove it entirely, and write the whole letter properly. However, if the letter is composed of two pieces of writing, and it became invalid from the second piece, it is only necessary to scrape away the portion which made it invalid (see below in 8; there are some who are stringent about this). Scribes must be very conscientious about the laws of ḥak tokhot. We will explain some of them here, and the remainder follow logically from them.
3. If the upper yud of the aleph became joined to the middle line which goes along its diagonal there are different procedures. If he wrote the yud first and then the middle line, the yud was made correctly and he has only to scrape the middle line away entirely, because it was invalid before it was finished. But if he wrote the middle line first and after that the yud and it became joined, he scrapes away the yud and not the line, because the line was made correctly and the yud made it invalid. If he did not correct it until after he had made the underneath yud, he must scrape away that also. If the underneath yud became joined, and not the upper, he needs to scrape only it away and write it afresh.
4. And if the whole aleph was written correctly and a drop of ink subsequently fell onto it, even onto the upper yud, it is sufficient to scrape that alone away and write it correctly.
5. If a drop of ink fell – even if it is still wet – into the interior of beit such that it appears like peh, or such that it could no longer be called beit, or if it fell into the middle of a different letter and it is no longer recognisable, the letter is invalid, and it is of no avail to scrape away the ink-blot such that the letter looks right, but he must scrape off the whole letter. And if he erred and wrote hey instead of dalet, likewise it is of no avail to scrape away the leg to leave dalet – but he need not scrape away the entire letter; after he has scraped away the blot, he need only scrape away the roof, so that the form of dalet is nullified, and then he can complete the letter by writing. If one wrote khaph peshuta instead of hey, he must first scrape away the leg until he has made dalet, and afterwards he writes the left leg (see para. 10). If he wrote the left leg first, he cannot fix it until he has scraped away the entire [right] leg. If he erred thus in a divine Name – further consideration is required (Peri Megadim, 32, Eshel Avraham paragraph 26).
6. If the opening of mem petuha joined up and closed, it does not help to scrape the joined part and open it. What should one do? He should scrape the whole snout [the vav part] until what remains resembles khaph or nun kefufa, and afterwards complete it correctly. This is the rule for all letters which are made from two writings – if it was made invalid by one of the writings, it is only necessary to scrape away that one. But if one made reish like dalet [squared at the corner], one should be stringent and say that it is not sufficient to scrape away only the leg before rewriting it as reish, because it was made invalid between the roof and the leg equally, and therefore it is necessary to erase both. And if he wrote reish and afterwards erred and added a point like dalet, it is sufficient to erase either the roof or the leg, with the point, and then rewrite it as reish – but blessings are upon one who is stringent in this.
7. The case of yud whose left point extends as much as the right leg (or more), and hence resembles ḥet. The method of fixing seems to me thus: if the scribe first wrote the yud with the right leg and after that made the left point and extended it too far, in this case it is sufficient to scrape away this point and also the right leg, and then write it correctly. But if he made the left one first (as is the custom of many scribes when writing yud – they first make a thin line which forms the tag and the left point, and extend this to the body of the yud), he must erase the entire yud and write it afresh. In either case, if he is able to fix it with ink, that is good (for the yud in the Divine Name, see chapter 12 para. 12).
8. There are some great scholars who teach that also in the case of letters made from two parts, such as mem petuha which became closed up and suchlike, that it is not sufficient to scrape off only the second part – by which it became invalid – but he must scrape away the whole letter and write it afresh. One should ideally be stringent like this when it is possible (Ganzfried, 125:38; Responsa of the Zikaron Yosef, ch. 3).
9. If one extended the roof of khaph peshuta such that it resembled reish, he must scrape away the roof until only the form of a vav remains. If it was in the divine Name “eloheikha” – see below, 12:13.
10. If one erred and wrote hey instead of ḥet, it is better to scrape away the entire hey. Post facto, if one fixed it in this manner: he scraped away the leg and the roof until only a small part and the right leg remained, like vav or zayin, and afterwards wrote a zayin to its left and joined them with a hump, it is also valid. And since post facto ḥet is valid without a hump, if he wrote the dalet of the hey it was not made invalidly, because it could have been made into a flat ḥet, and since the dalet was not made invalidly, it is not necessary to scrape it all away, only as much as he needs to fix it.
11. If one made a ḥet in this manner: he wrote dalet and zayin and put a hump on top of them – if the right point of the hump was joined to the right side of the dalet, across from the right leg [image], it may be fixed: he may scrape at the dalet until it resembles zayin, and this is valid even for tefillin and mezuzot, even if he had continued on (and there is no issue with writing out of order since it did not need to be corrected by writing). But if the hump was joined on the left side of the dalet [image], he must scrape away it all, or to the opening of the hump and the roof of the dalet, until only the form of a zayin remains, and afterwards he writes the hump – and for tefillin and mezuzot, if he had continued writing, it cannot be fixed. It makes no difference to all this whether the dalet was made on the right side and the zayin on the left or vice versa.
12. If one wrote shin with four heads: even if a child recognised it, one still cannot fix it by scraping off one of the yuds so that three heads remain. Instead, he must nullify its form and complete it, since if he does not, it constitutes ḥak tokhot. With tefillin and mezuzot, if he had continued writing, it cannot be fixed because it will be out of order.
13. In any case where a letter has become invalid by ḥak tokhot, it does not help if he goes over it afterwards with a quill.
14. If wax or fat dropped onto the letters – even if the letters are obscured – in all cases, after the wax or fat has been removed, it is valid (see 11:4), and this does not constitute ḥak tokhot, because even when the wax or fat is on them, they are not nullified, just obscured.
15. If a letter became stuck to another and its form was not spoiled, but each had its form – if the join was made before the word was finished, it is permitted to scrape away just the join only in pressing circumstances, like if it was a letter in a divine Name. It seems to me that this is also so for tefillin and mezuzot if one had continued writing, so they should not be made to require geniza. But if there are not pressing circumstances, he must erase the whole second letter which is stuck to its fellow. If, however, the join was made after the second letter was completed, like if one wrote vav after nun and the bottom of the nun ran into the vav, he may scrape away just the join, and this does not constitute ḥak tokhot, because the form of the letter was completed before they became joined – and a blessing is upon one who is stringent and scrapes away the entire vav (see Excursus 6).
16. If the letter became stuck after he had already moved his hand and lifted the quill from the letter – whether the join was at the top or at the bottom, he can scrape and it is of no consequence whatsoever. This is if the forms of the letters were not spoiled, but if the forms were changed because of the join between the first and second letters, obviously scraping doesn’t help, and he must rectify it as explained above.