Roman Bank Receipt διαγραφή

The word diagraphe (διαγραφή) applies to a payment made through a bank, and by extension to the receipt that constitutes proof of the payment. Such payments are recorded in registers, and abstracts can be produced to confirm the existence of the payment. This description concerns Roman period diagraphai; for Ptolemaic bank documents, such as receipts, and instructions to accept a payment (both referred to as diagraphai), cf. the Erbstreit Archive [TM Arch 81], the Archive of the Theban Bank (section 3) [TM Arch 205], and see [Vandorpe and Vleeming 2017 : no. 1, with commentary] and [UPZ I : no. 114, with commentary].

Diagraphai are sometimes mentioned within the text of private contracts drafted in the form of a letter [cheirographon] in Oxyrhynchus. They provide an indirect trace of the use of a diagraphe for making payments through a bank; they go back to the Ptolemaic era [5257 73 or 44 BCE, Oxyrhynchus] and continue well into the Roman era [15644, 65 CE; 22521 144 CE, both Oxyrhynchus]. Direct attestations of receipts issued by banks (i.e. not mere mentions within a contract) appear at the very beginning of the Augustan age [18644 27 BCE, Herakleopolis; 22701, 6 BCE (copy of a diagraphe)].

In the early Roman period, transactions such as loans or sales are recorded – and are recognized as legally binding – through the use of a bank receipt, which can be produced as evidence before a court in order to support the claim of the existence of a contractual relation.

Bank receipts are divided by modern specialists in two broad categories: a) those that are not considered as self-contained, i.e. they refer to another document produced by a separate authority to make the transaction legally binding; b) those that are considered as self-contained, i.e. they include all the clauses that make the transaction legally binding [Wolff 1978b : 96].

It seems that no official legislation was introduced to actively support such a development towards self-contained, legally binding bank receipts; apparently bankers themselves gradually gained acceptance as a body that could issue documents to be produced before a court [Wolff 1978b : 100-103].

In the I CE, bank receipts are abstracts from bank registers [Wolff 1978b : 97-98] where a transaction was recorded. From a legal point of view, their contents correspond to the receipts that are not self-contained, i.e. they refer to another document produced by a separate authority to make the transaction legally binding.

The praescript is characterized by the following elements:

• ἀντίγραφον διαγραφῆς [through διά bank <gen.>]
• [date@start]
• [from creditor <nom.>] [to debtor <dat.>]

The implied verb, which is not explicitly provided on the document, is διέγραψε, e.g. [11326 92 CE, Soknopaiou Nesos]. In the subsequent part of the document, however, the payment is expressed from the point of view of the person receiving the payment, usually a debtor, with the infinitive ἀπέχειν, but without a conjugated verb introducing the infinitive [21035 57 CE, Alexandria; 9448 I CE, Arsinoite nome; 8736 II CE, Herakleopolite nome]. This is explained by the fact that those documents are abstracts from registers.

A distinction can be made between documents from Alexandria and the Arsinoite and Herakleopolite nomes on the one hand, and documents from Hermopolis on the other.

In the first area, the format is quite unremarkable: a vertical sheet in pagina format, with a height of ca. 22 cm and a width of ca. 8 cm, with the writing along the fibres [8736 (Herakleopolite nome); 9448 (Arsinoite nome)]. In a receipt from Alexandria [21035], the height is 36.5 cm.

In the Hermopolite nome, the format of bank receipts can be more elaborate. In [22825 44 CE] the scribe writes on a horizontally oriented sheet (H. 24.8 x W. 108.6 cm) in wide columns following the direction of fibres; the relevant contract is drawn up and a copy of the corresponding bank payment is added to the document by another hand.

The layout of these I CE documents do not seem to follow a strict pattern. In [8736], apart from a slight ekthesis in line 1 highlighting the caption ἀντίγραφον, the text is written by a single hand in a single block. In [9448], the enlarged initial alpha of ἀντίγραφον serves the same purpose of highlighting the nature of the document; in line 3, the date is also marked by the symbol for (ἔτους) placed in ekthesis; moreover, the debtor has confirmed the transaction in his own hand, below the main text.

It appears that apart from the wording of the praescript, there is not much consistency in the way scribes prepare such receipts on behalf of banks. This may be explained by the fact that, in the first century CE, bank receipts are not yet considered valid self-contained contracts. The degree of formality increases in the subsequent period, as bank receipts gain acceptance as self-contained, legally binding contracts.

In the early second century CE, bank receipts become self-contained: instead of relying on a transaction supervised by another authority, they become per se legally binding contracts. This change is reflected in the structure of those receipts: most conspicuously, a new praescript appears, with a certain degreee of consistency between the Hermopolite and Arsinoite nomes.

The absence of a mention of a copy (ἀντίγραφον) suggests that those receipts are not abstracts from registers anymore, but receipts in their own right. The praescript follows a new pattern:

• [date@start]
• (διαγραφή) [through διά bank <gen.>]
• [from creditor <nom.>] [to debtor <dat.>]

This is followed by a verb in the infinitive, e.g. [ἔχειν] [9302 II CE, Arsinoite nome], [ἐσχηκέναι] [22718 211 CE, Hermopolis], [ἀπέχειν] [9194 II CE, Arsinoite nome], [πεπρακέναι] [9164II CE, Soknopaiou Nesos; 22324 III CE; 22810 226 CE, both Hermopolis], with no conjugated verb to introduce it; the infinitive may also be missing [19432 153 CE; 23525 188 CE, both Hermopolis]. This wording apparently continues that of the first century, although now the receipts are no longer abstracts from registers: they function as full-fledged contracts of loan, sale etc., as shown in the detailed description of the transaction in e.g. the bank receipt for a camel sale [9164], which contains a guarantee clause (lines 20-23). In a bank receipt rent of an oil-press [10938 II CE, Theadelphia], there is a clause confirming the validity of the transaction (lines 18-20). In another bank receipt for a camel sale [9194], the buyer pledges to declare his purchase in his declaration of property (lines 18-21).

A particular loan document is identified by [Wolff 1978b : 103] as a fictitious diagraphe [20157 102 CE]: in the guise of a copy from a bank register, its purpose is to confirm the repayment of a loan that was already mentioned in another document, also preserved [9091 102 CE, Arsinoite nome], written in the form of a cheirographon - here the creditor acknowledges the repayment of the loan and promises to confirm it with a bank receipt.

Bank receipts from the Arsinoite nome retain the previous pagina format, and the writing along horizontal fibres. The distinctive format found in the Hermopolite nome in the I CE remains in use in the two subsequent centuries, with a very wide horizontal sheet cut from a roll with a height of c. 24 cm and a width of approximately 60 to 100 cm. The text is written along the fibres, in two columns: one column contains the bank receipt itself, while the other records the agreement underlying the transaction; in most cases this happens on the same day. The bank receipt can be placed first [22324] or second [19654 I CE, Hermopolis].

The text is written in a single block, but the date stands out because of the large epsilon of ἔτους [9194; 9302; 13751 151 CE, Ptolemais Euergetis; 9164; 9825 II CE, Soknopaiou Nesos]. The presence of a date at the beginning of the document lends a quasi-official character to such bank receipts.

Two documents concerning a loan are pasted together in [9195/9196 139 CE, Karanis]: col. ii contains a loan contract dating from 139 CE, drafted by the notarial office in Karanis. When the loan was repaid in 141 CE, a bank recorded the payment with a receipt (col. i), which was pasted on the left side of the contract. The contract itself was crossed out and thus cancelled. This shows that two distinct types of documents (i.e. contract drafted by a notary, and bank receipt) were considered as appropriate to validate the two stages of the process, i.e. loan and repayment. Presumably, the copy of the loan contract belonged to the creditor; once his loan was paid back, he handed the contract to the banker, who crossed it out; then both the contract and the bank receipt were handed to the debtor, as proof that his loan had been repaid.