I. This paper discusses some documents relevant for the registration of a hypallagma in the bibliotheke enkteseon. My main point, though, does not concern so much the documents themselves as the way they were explained by some of the founding fathers of Juristic Papyrology. From their time to our own, part of the discussion on real securities keeps turning around their «real effect», or even whether they were or not «real rights». The core of this paper is thus methodological: it aims at a reflection on the danger of using Roman Law categories such as «real rights» or «real effects» for the law of the papyri.

II. What hypallagma is was clarified by A.B. Schwarz¹. He fully proved what also Rabel had suspected, namely that a hypallagma is not a hypotheke (mortgage), not even a special kind of hypotheke, but a totally different type of security, in many aspects actually opposed to the old hypothecary model². It may be worth restating the differences between hypotheke and hypallagma³.

Hypallagma contracts lack a forfeit: failure to pay is not enough to make the debtor’s property fall to the creditor. This is confirmed by the executive procedure: the hypothecary creditor, becoming owner by forfeit, needs only to request ἐμβαδεία from the praefect, that is, the material possession of the pledge. The hypallagmatic creditor instead needs first to claim for ἐνεχυραϲία: only through it does he acquires ownership, and only then may he start a second procedure for ἐμβαδεία. This means that he needs to go through the very same two-step procedure he would use if no security at all had been given.

What is then the point of a hypallagma? What does the creditor acquire with it? The right answer is: nothing. The security here lies not in anything the creditor acquires, but in something the debtor renounces, i.e. his right to dispose of the property. Every hypallagma contains a non-alienation clause, by which the debtor promises not to sell or further mortgage the pledge until payment. But such a clause is also present in hypothekai, so that one wonders why would any creditor choose hypallagma if hypotheke provides the same with a much easier executive procedure.

The reason is to be found in yet another significant difference: in hypallagma-contracts, the execution of the pledge is mentioned only as part of the general πρᾶξιϲ, the right of execution on the person and the rest of the belongings of the debtor. In hypothecations, on the contrary, the creditor’s right is confined to the pledge: failure to pay turns the creditor into an owner, and he cannot claim anything else from the debtor. In other words, the debtor’s liability is totally absorbed by the hypotheke: this is the so-called «reine Sachhaftung», purely «real» liability. For the creditor, this means bearing the risk of the hypotheke being destroyed or damaged or successfully claimed by someone else (hence the frequent clauses that address those risks, ĸίνδυνοϲ – and βεβαίωϲιϲ – clauses). Here, in my opinion, lies the advantage of a hypallagma, and possibly the reason for its invention: one loses the forfeit of hypotheke, and is thus forced to a more complex executive procedure, but one keeps the general liability of the debtor, which would disappear with a hypotheke⁴.

III. The documents I will consider pertain to the registration of the hypallagma in the bibliotheke enkteseon. The registration was, as I have argued elsewhere, a key element for the effectiveness of the hypallagma, securing non-alienation better than any of the previous mechanisms devised for this purpose⁵. In the second century AD, the hypallagma starts being actually contracted «through the bibliotheke», as many documents imply and others prove directly⁶; the two become so closely intertwined that the eclipse of the bibliotheke in the fourth century AD seems to have brought about the extinction of the hypallagma as well⁷.

Of all the documents relevant for the registration of the hypallagma, I will consider only the registration requests, and from this group only those already known at the time of Schwarz and Mitteis: it is precisely Mitteis’ explanation of these texts that concerns us here; and those edited later (P.Vars. 10 III, P.Kron. 18 and P.Wisc. II 54) all belong to a different model. They are not requests for registration of an already constituted hypallagma, but hypallagmata contracted directly through the bibliotheke.

The first document to be considered is the well known P.Lips. I 8 (= M.Chr. 210; Hermopolis, AD 220). It was edited by Mitteis himself in 1906, under the title «Apographe über eine Hypothek». The document is indeed an apographe (3: ἀπογρά (φομαι)), i.e. the declaration of a right to the bibliophylakes enkteseon, in this case to those of the Hermopolite nome (1). This right, though, is not a hypotheke but a hypallagma (7-9: δίκαιον ὧν ὑπήλλαξέν μοι |… κλήρων κατο [ι]κικῶν | [(ἀρουρῶν) τριῶ] ṿ). The notion of the hypallagma as something other than a simple type of hypotheke was in fact introduced by Rabel, Schwarz and Mitteis himself only some five years later.

The apographe is presented by Aurelius Tithoetion also called Sarapammon (in spite of the lacuna in l. 2, see the subscription in l. 15), from Hermopolis, in AD 220/221⁸. He declares his right arising – according to a diagraphe of the same year – from the hypallagma of three arouras of catoecic land at the village of Moirai, as well as – according to the same diagraphe – the right to recover with an interest rate of 12 % a capital of 840 drachmai, to be paid in the month of Thoth of the following year. The debtor, who gives her consent to the apographe (4: μετ̕ εὐδοκήϲεωϲ Αὐρηλίαϲ Κολλύχ[ιοϲ]) by means of subscription is Aurelia Kollauchis, from the village of Moirai in the nomos of Kusai⁹.

The whole scheme was designed for a short duration: the following year, the debtor was supposed to pay back the loan, and this seemed indeed well secured, with a hypallagma and also a collateral, Aurelius Castor, mentioned in line 6. And yet, thirteen years later, the loan was still unpaid; what is more surprising, no measures at all were taken against the debtor or the collateral, no attempt was made for the execution of the hypallagma¹⁰. This we know because we encounter the same business in P.Lips. I 9 (= M.Chr. 211; AD 233). Aurelius Tithoetion, the loan giver, is now dead, and in his testament – a Roman testament, as is made clear in this document – his instituted heirs are his three children, Achilleus, Heron and Eudaimonis, who are all still underage. Acting through their mother, Aurelia Aretous, they now submit again to the bibliophylakes of Hermopolis an apographe for the very same right, acquired as part of their inheritance, namely: the credit arising from the loan itself and the right on the land under hypallagma. Together with the petition, they submit a copy of the testament and of the initial loan-diagraphe, the original of which had already been deposited by their father.

These documents were the first pieces of evidence for the registration of the hypallagma: that is, the first conclusive proof that to the registrable rights mentioned in the Edict of Mettius Rufus – namely kteseis, hypothekai and preferential rights of wives and children – hypallagmata had to be added¹¹.

A third document was published by Grenfell and Hunt just one year later, in 1907: P.Tebt. II 318 (= M.Chr. 218; Tebtynis, AD 166). An unnamed creditor has made in AD 162 two loans, documented in two public homologiai, one jointly to the brothers Serenus and Didymus for 560 drachmai, the other to the former alone for an unknown amount. Each loan is secured by the declaration that a plot of land is to be kept unalienated and intact for the lender (8-9 and 14-15: φυλάξειν μοι ἀνεξαλλοτρίωτα καὶ ἀκαταχρη-μάτιсτα). Since the hypallagma is nothing but a surrender of the right to alienate, one would expect to find this one also labeled as a hypallagma; surprisingly, the document seems to avoid the term (see below). The creditor, out of fear that his right might escape notice (18-20: [φο]|[βου]μένη δ[ὲ] μὴ λάθω [κατὰ] ṭò εἴϲ με δίκαι [ον]), submits it for parathesis (20-21: ἐπιδίδ[ω]μι εἰϲ τò τὴ[ν] | [παράθεϲιν γεν]έϲθαι) in order that the land may be registered in the class of property subject to a claim (the editors’ suggestion for the lacuna before τῆς ṭ[ά]ξεωϲ κ[ατο]χίμων). The document ends with the usual clause explicitly safe-guarding the rights of third parties: εἰ δὲ φ.[άνειαν ἑτέρ]ῳ προϲή[κο]υϲαι ἢ π̣ρ̣ο̣κ[α] τ̣[ε|ϲχημέναι] δ̣ι̣[ὰ] τοῦ βιβλιοφυλακίου μὴ ἔсεc̣[θαι] ἐμπόδι[ο]ν ἐ̣κ̣ [τῆс|δε τῆс] π[αρ] αθέсεωс ἀκ [ο]λούθωс οἷс παρεθ[έ]μην ἀντ[ιγρ]ά̣[φοιc] (22-24). Finally, we have the signature of the secretary of the record office for registration: «I, NN grammateus, have recorded»¹².

IV. One obvious singularity of the Tebtynis papyrus is this mysterious reluctance to use the term hypallagma for a security that, as far as we can judge, has exactly the same content as a hypallagma, i.e. the surrender of the debtor’s right to dispose of the object. If we compare the papyrus with P.Lips. I 8-10, it is also striking that the term ἀπογράφομαι, which we find there, is avoided in P.Tebt. II 318. This document does not present itself as an apographe, but merely as a request for parathesis. It is tempting to connect those two details, as Mitteis already did: «Denn überall, wo es sich um Verträge handelt, welche nicht selbstständige Rechte, sondem bloβe Verfügungsbeschränkungen bei solchen Rechten zum Gegenstand hatten, war eine eigentliche apographe ausgeschlossen – es wurde ja kein dingliches Recht begründet – und einfaches Gesuch um parathesis am Platze. Es ist möglich, daß schon bei [P.Tebt. II] 318 diese Auffassung zutrifft: hier handelt es sich scheinbar nicht um ein dingliches Recht, sondern um ein vertragsmäßiges Veräußerungsverbot zur Sicherung einer Förderung. Zwar ist diese Erklärung nicht zweifelsfrei; denn es kann dagegen gesagt werden, dab das Veräußerungsverbot stillschweigend die Bestellung eines Pfandrechts in der Form eines hypallagma enthielt: doch ist formell ein hypallagma nicht bestellt worden.»¹³ In short: there is formally no hypallagma, hence no real right, hence no place for an apographe but only for a parathesis.

V. The first objection that comes to mind was obvious to Mitteis himself: a hypallagma is nothing else but a surrender of power of alienation, so that one cannot see a reason for a difference in the registration process depending on whether the parties label it as a hypallagma or not, even if one accepts that they avoid the term in our document for some substantial reason¹⁴.

A yet more compelling objection arises, in my opinion, from the following consideration. Mitteis is here transferring to the registration of securities the difference he has just presented for the registration of ownership, between a definitive apographe and a provisional parathesis¹⁵: when the seller had no right to alienate (having surrendered this right as a real security) or simply, when he was not registered, the buyer, specifying that he did not yet submit the apographe, asked for a mere parathesis, very often adding, as in our P.Tebt. II 318, that «if it be found that any other person has ownership of or claims against the property, no obstacle shall arise from this registration».

This opposition between apographe and parathesis, however, understandable for the registration of ownership, does not exist when securities are concerned: the surrender of the right to alienate is equally temporary, whether it is labeled as a hypallagma or not¹⁶. In both cases it will disappear as soon as the debt is paid. That no such opposition exists concerning the hypallagma should have been clear to Mitteis himself from P.Lips. I 9: the document is an apographe of a hypallagma (7: ἀπογραφόμεθα), but the aim of such an apographe is no other than the parathesis (21-23: ἀκολούθωϲ τῇ διαγραφῇ ἧϲ τò ἀντίγραφον… νῦν ἐπηνέγκαμεν ὑμῖν ϲὺν τῷ τῆϲ διαθήκηϲ ἀντιγράφῳ εἰϲ τò τὴν παράθεϲιν γενέσθαι). As far as a hypallagma is concerned, it seems, parathesis and apographe are not contraries, the former being the aim of the latter. It is also worth noting that, contrary to the ownership documents that expressly underline that they are not yet apographai, P.Tebt. II 318 simply does not use the term ἀπογράφομαι: no difference seems to exist with apographai regarding the substance of the inscribed situations or the inscription technique; the document is, exactly like P.Lips. I 8 and 9, the request for the parathesis of the hypallagma, which was the typical act of the bibliophylakes for this kind of security.

VI. The main disturbing element in Mitteis’ construction, and what makes this discussion methodologically relevant, is his use of the category of «real» rights¹⁷. This transcends Mitteis’ case and is far from being a mere historical anecdote: true, legal papyrology has been for decades acutely aware of the necessity of an emancipation from the categories of the Roman Law tradition¹⁸; and yet the discussion on real securities in the papyri is still all too often focused on the question whether they constitute or not «real» rights, whether they have «real» effects or not¹⁹.

To show how misleading this apparently innocent terminology is, a brief explanation for non-lawyers of what real rights are may be necessary. This category, «real rights», iura in rem, in the sense of «rights on things», is in the Roman legal tradition opposed to iura in personam, the rights of creditors against their debtors. The opposition iura in rem / iura in personam structures our whole understanding of private law, and is so deeply engraved in the continental legal mind that it is often presented as if arising from the very nature of things. It is thus understandable that scholars like Mitteis or Rabel, who incarnated the tradition of Roman Law in the turn of the century, let the category slip in their analysis of the Law in the papyri. This is understandable, but also dangerously misleading, because it belongs to the Roman tradition. Mankind has known plenty of legal systems which completely ignored it, in particular the Graeco-Egyptian practice found in the papyri.

The category emerged in Roman Law when the type of protection characteristic of ownership was extended to other situations²⁰. Since ownership is a position of exclusive privilege over a thing, the owner is protected against anyone who challenges or usurps his position. The claim of the owner is what we call a vindicatio or actio in rem, a claim on the thing itself, that can be used erga omnes, against whoever challenges the owner’s position. In Roman mid-republican times, the same type of claim, an actio in rem or vindicatio was extended to other situations of privilege over a thing different from ownership. This result seems to have been initially reached by assimilating such situations to ownership itself.

Thus, for instance, if A has a right of way through B’s property, we have reasons to believe that, in archaic law, A was considered to be the owner of the path, so that he could vindicate it even against B, and against any subsequent owner of B’s property²¹. Later, the notion of a «right of way», different from ownership, emerged, and was brought into the wider category of the so-called servitudes, whereby a property provides a service to another. Yet, although by then the holders of a servitude are no more considered to be owners, they keep their actio in rem, their vindicatio – now a vindicatio servitutis –, to claim their right against anyone who challenges it, including the owner of the servient estate, or anyone who may become owner after him.

A similar process took place in other situations of legitimate interest in an object, including the right of the creditor on the pledge that secures his credit²². That also the creditor has an actio in rem means that he can claim the pledge from whoever possesses it, including the owner; and this not only if the owner is the debtor himself, but also if it is someone else, including someone who has bought it afterwards in good faith.

This was the Roman way to secure the object for the creditor: granting him a claim on the pledge like that of the owner, erga omnes, i.e. against anyone, new owner included. One could however imagine a very different strategy to protect the creditor, not raising him to a position stronger than that of any owner, old or new, but rather creating obstacles to the sale of the pledge, or making such sale altogether void, or weakening the position of the buyer so that he could not oppose his right to that of the creditor. Here one does not operate at the level of the creditor raising it, but at that of the buyer. This seems to have been the case in the law of the papyri, where we find all three strategies: in the earliest preserved hypallagmata, there is an attempt to prevent a sale by having the creditor keep the title-deeds until the debt is cancelled²³; in both hypallagmata and hypothekai, the registration in the bibliotheke enkteseon prevents alienation through public deed; further, in some hypothekai a clause declares void the attempted alienation or further hypothecation²⁴; finally, when the bibliophylakes exceptionally deliver an epistalma allowing the debtor to sell the pledge – because his declared intention is to satisfy the creditor –, we find the registration of the buyer not by means of a definitive apographe but by a mere provisional parathesis, with the usual clause stating that «if another person should be proved to have a right to the object or a hold upon it recorded through the record office» such as, obviously, the creditor until payment – «the present parathesis shall not stand in the way of his claim.»²⁵ Here the protection for the creditor rests upon the fact that the buyer will not be registered as the final owner, and has to accept that his registration has no value before the creditor. The effect is similar to that of a «real» right, but the way that leads to this result is different: it consists of depriving the buyer of a full ownership registration.

Whatever the method to prevent the sale or keep its effects in check, practically every single hypothecation, every single hypallagma, contains a non-alienation clause²⁶. The contrast with the Roman system could not be sharper: the Roman debtor, precisely because real securities are constructed as real rights, keeps his potestas alienandi. There is no harm in it for a creditor who can claim the pledge from anyone²⁷. Things were more problematic in the case of movables – the sale in that case constituted furtum; but our discussion here has been mostly confined to immovables, and for these the principle is clear: the buyer becomes owner, and is in any case exposed to the creditor’s actio in rem²⁸.

Things are so different in the papyri: the concern that the documents show to prevent the debtor from selling or further mortgaging the pledge has called for many conjectures and a rich literature, especially at the beginning of the twentieth century²⁹. I would suggest, also here, a less «Roman» perspective: instead of taking the Roman system as the «natural» one and seeking the reasons of a «Greek deviation», we could shift our perspective and ask ourselves, not «why could not the Greek debtor sell ?», but rather «why could the Roman debtor sell ?» The answer then seems obvious: because this did not harm the creditor, who had a claim erga omnes. It is thus quite likely that the concern that we find in the papyri to prevent a buyer from showing up comes from the fact that, in the Graeco-Egyptian tradition, it would be unconceivable to grant a non-owner a claim against an owner³⁰. Actually, the whole idea of iura in re aliena, the idea of extending the protection erga omnes – which is typical of the owner – to certain non-owners against the owner himself, was entirely Roman.

This takes us back to our methodological question, i.e. the dangers of using – even for the sake of easier communication with the wider community of legal historians – the notion of «real» rights or «real» effects. Historically, the construction of real securities as limited real rights has been only one of the possible ways to secure the pledge for the creditor. If the discussion on the protection of the creditor against a subsequent buyer is built around whether the Graeco-Egyptian real securities are «real» rights or not, and have or not «real» effects, then the aim – the protection of the creditor against third parties, essential in any system of real securities – – is being confused with the means – with one only of the possible means, namely granting the creditor a claim erga omnes. This confu-sion of aims and means can only create further confusion. This is made worse if, from the mechanisms devised to protect the creditor before a possible buyer, we conclude that the GraecoEgyptian real securities had a «real» effect; and worse again if we do it with the intention of making the papyrological practice accessible to lawyers trained in the roma-nistic tradition, since they will inevitably imagine a claim erga omnes³¹. We will also create confusion if, identifying protection of the creditor with the «real» right structure, we deny the former in order to deny the latter: no matter how far fTom the «real right» model a real securities system may be, one of its main purposes all the same is to protect the creditor from a subsequent sale³².