The Identity of a Get
By Rav Moshe Taragin
Shiur #12: The Identity of a Get
The mishna in Gittin (19a) records a machloket between R. Yossi Ha-Glili and the Chakhamim regarding the validity of a get written on edible material or on living organisms (people, slaves). In theory, the get should be valid, since it was written with ink upon a solid substance (in the case of a living animal, it may have been tattooed). R. Yossi Ha-Glili argues, however, that since the Torah refers to a get as a sefer keritut, it should mimic qualities of a formal sefer, a book. Typically, books are not written on foods or animals, and a get written on these items is similarly pasul.
The response of the Chakhamim to R. Yossi Ha-Glili is quite enigmatic. The gemara records that the Rabbanan read the word "sefer" less literally than R. Yossi Ha-Glili and did not demand the extreme "sefer" profile that he assumed. Instead, it merely alludes to "sefirat devarim" the get must record and retell the story of the divorce. This description may or may not entail unique requirements, but regardless, the Rabbanan clearly did not maintain formality for a get based upon a literal reading of the word "sefer" as a book.
What is unclear is whether the Chakhamim ultimately agree that a get retains SOME elements of a sefer. While they clearly did not require that it be written on non-living and non-edible matter, perhaps a get possesses SOME formal elements of a sefer. This is very unclear in the gemara (Gittin 21b) and seems to be a point of contention amongst several Rishonim.
For example, Tosafot (20b) claim that although a get may be written on edible or living items, it must be written upon something durable. The mishna allows writing a get upon the leaves of an olive tree, but only because these leaves as less likely to wilt and do so less rapidly. In contrast, fig leaves, which quickly dry and shrivel, would not be a valid option for a get. Tosafot cite the Tosefta, which does not provide any rationale for this demand, but when the Sifri cites this requirement, it bases it upon the term sefer. Since a get must possess SOME formal elements similar to a book even according to the Chakhamim who disagree with R. Yossi Ha-Glili, it must be written on something durable. Accordingly, even though the Rabbanan reject the extreme position of R. Yossi Ha-Glili in disqualifying edible or living matter, they DO acknowledge the notion of sefer for a get.
Perhaps the clearest indication that even the Rabbanan who allow drafting a get upon edible items still maintain SOME sefer requirements stems from an interesting gemara in Gittin (21b) that describes a get delivered in fragments instead of as one integrated document. The gemara disqualifies this get because the Torah demands a sefer ONE document and not MULTIPLE documents. Apparently, even the Rabbanan adopt some parallel to a sefer for a get.
Interestingly, the Tosafot Ha-Rosh sensed this issue and diverted attention away from any sefer features of the get. He insists that the Rabbanans requirement of a single document reflects THEIR dissenting of the word sefer that the get must NARRATE the STORY of the divorce, NOT that it is a formal book. The term "sefer" does not refer to the quality or halakhic nature of the get, but rather to its TEXT and STORY, and it can therefore be written on any material. A get written on different pieces of paper, however, does not constitute one holistic sefira, or narration, but rather a series of partial installments. Thus, even if we do not assign the identity of sefer to a get, it must still be written on one piece of paper to assure integrity of the story.
In contrast to the Roshs approach, there are three comments of Rashi that may indicate that HE assumed that a get must reflect the nature of a formal sefer even according to the Rabbanan.
In his comments to the gemara in Gittin (20b), Rashi addresses a case in which a man delivers a get to his wife but withholds transfer of the parchment upon which the get is written (Harei zeh gitteich ve-neyar sheli). Rashi explains that this get is invalid because a get consists of letters and paper; if only ONE ASPECT of a get has been delivered (the letters) while the other has not (the paper), the entire get has not been transferred. By claiming that the paper is also an integral part of the get, it seems that Rashi is defining a get as a sefer. Since a get must resemble a formal book, the paper it is written on is indispensible. If the get requirements were less formal (more like a letter than a book), transferring possession of the text without the paper might not have been problematic.
Second, Rashi comments on an interesting gemara in Menachot (34a) which states that a mezuza may not be written upon stone; he claims that this is based on a gezeira shava of the term ketiva. Rashi claims that the source of this halakha and the gezeira shava is a get, which cannot be written upon a stone because the Torah refers to a get as a sefer keritut and it must therefore resemble a sefer. Evidently, according to Rashi, even the Chakhamim agreed that a get is considered a sefer and must resemble a document/book, and it therefore cannot be engraved upon stone.
Tosafot argue with Rashi, claiming that the sefer requirement does not apply at all to a get according to the Chakhamim; accordingly, a get may indeed be written upon stone. Rashi, however, seems consistent with his opinion in Gittin (20b) that a get maintains partial sefer status even according to the Rabbanan.
A third relevant comment of Rashi appears in Menachot (32b). The gemara invalidates a mezuza that was written like a letter. Rashi explains that this refers to a mezuza written without sirtut (engraved markings outlining the different lines of text). A mezuza must be written more formally, like a sefer. Rashi again claims that this formality is learned from a get; just as a get must be written with sirtut, a mezuza must be as well. Again, Rashi notes that the requirement for writing a get with sirtut stems from its definition as a sefer. Rashi consistently invokes the sefer terminology in detailing halakhic requirements for a get, even according to the Rabbanan. A get cannot be written upon a stone, requires sirtut, and includes the paper because of its status as sefer. (In this context as well, Tosafot reject Rashi's position, arguing that the Chakhamim reject the formal sefer-like status of a get. The laws for mezuza must accordingly be derived from a different model.)
This position of Rashi may help explain a related concept found in various Rishonim. Both the Ramban and Tosafot claim (Gittin 20b) that a get does not require mukaf gvil (letters completely surrounded by parchment). Unlike a sefer Torah or mezuzot, a gets letters may be partially attached (as long as the letter is still identifiable). This is quite logical; the strict laws that govern stam should not logically apply to get.
Yet many Rishonim claim that a get must, in fact, be written with the same style as a sefer Torah and that its letters must be mukaf gvil. Similarly, the Rishonim debate whether letters of a get must be crowned with taggin in the same manner that letters of a sefer Torah must be adorned. (See, for example, the Hagahot Ha-Ashri in the beginning of Gittin.) Why should a get require these same structural and formal requirements?
Perhaps these Rishonim agree with Rashi that a get DOES possess a limited status as a sefer. Accordingly, its text should be prepared in the same formal manner that the ultimate sefer a sefer Torah is prepared.
Although the question of whether a get must share the characteristics of a formal sefer remains unclear, there is an unmistakable statement of the Yerushalmi affirming this status. The mishnah (Gittin 21b) disqualifies a get written upon something attached to the ground. The Bavli explains this based on the need to detach the item before delivery. Since the item is not ready for immediate delivery, as it still requires severing, the get is invalid. The Yerushalmi, however, disqualifies this get because anything attached to the ground cannot be considered a sefer. The Yerushalmi is thus quite clear that a get DOES possess a sefer identity.