I. Language and Law[169]

An important feature of the Babatha and Salome Komaise archives is their bi- or even trilingualism.^[170] The documents are written in sev­eral languages, as could already be observed from the treatment of the documents in the different volumes of their editions.^[171] In the Babatha archive the first ten documents, with the exception of P.Yadin 5, are all written in Aramaic, while the later are all written in Greek.

Greek as language used in legal documents

To start with, it should be observed that the use of Greek in legal doc­uments within a Roman legal context constitutes a problem in itself. A problem that does not attract much scholarly attention as Wacke noticed:

Welche Sprache schrieb das antike romische Recht vor für die Gültig­keit von Rechtsgeschäften? War die Verwendung des Lateinischen dafür unerlässlich, dessen Kenntnis darum nebenzu unentbehrlich, oder wie weit ging man in der Diskriminierung von Nichtlateinern? Die Proble­matik gehort zu den Allgemeinen Lehren vom Rechtsgeschäft.

In his detailed article Wacke explains that basically Latin was the only language in which valid legal acts could be conducted. Few exceptions are made, for example in the case of the stipulatio, the act of promis­ing something to another party. The nature of the promise obviously prompted leniency: parties who did business with one another should be allowed to make the stipulatio in their own language. Therefore, the basic rule for the stipulatio is that both parties can understand the lan­guage in which the stipulatio is made. Interesting is the case, discussed by Wacke, where the question is put in one language and the answer given in another. If question and answer are compatible, the promise is binding and the legal act valid.^[175] Concerning other legal acts, Wacke derives conclusions whether other languages could be used or not from a comparison with the acceptance of dumb parties in a legal act. Wacke assumes that when a person who could not speak at all was allowed to make a certain legal act this would certainly go for someone who could speak albeit in another language than Latin. Acts that fall into this group include acts that are found in our archives like sale or marriage contract. Wacke also refers to the difference between ius civile and ius gentium: one would expect that acts that were considered part of the ius gentium could be conducted in other languages than Latin. Ignoring all the exact details of his discussion it serves to cite his conclusion here:

Rechtsgeschäfte unter Lebenden konnten, wie wir sahen, soweit sie dem ius gentium angehorten, ziemlich früh unter der sich seit 242 v. Chr. ent­wickelnden Gerichtsbarkeit des Fremdenprätors auch fremdsprachig for­muliert werden.^[176]

Wacke distinguishes the languages used in legal acts and the language used by the judge in the lawsuit (‘Gerichtssprache'): according to the legal texts, in the second century CE the ‘Gerichtssprache' was Latin and only in late antiquity Latin or Greek.^[177] ‘Gerichtssprache' then indi­cates the language used by the parties to bring their case and by the judge to formulate the verdict, not necessarily the language used by the parties during the hearing of their case.^[178] This is important as there are no verdicts among the papyri from the Judaean Desert, while there are documents that pertain to several phases of a suit.

While Wacke is foremost concerned with a discussion of the Roman legal sources to determine how Rome dealt with other languages than Latin, Werner Eck approached the issue from the angle of legal practice: what do inscriptions and papyri tell us about the use of languages by the Romans in contact with their subjects?^[180] In this context he also mentions several documents from the Babatha archive. His conclusions are that Greek was used in legal acts and official announcements as the language that government and population shared, if only in theory, for Eck notes that the fact that Babatha had legal acts in Greek in her archive does not say anything about her own knowledge of Greek. The documents show that local parties made and wrote statements in their own native language (Aramaic) which were sometimes included in the original, sometimes in a translation. Eck assumes that during a lawsuit the parties spoke Aramaic, while a translator interpreted their statements for the governor or his representative.^[181] Eck assumes that bilingual personnel was easy to find as in most places soldiers were recruited from among the populace. These recruited soldiers spoke their mother tongue and learned Latin, which made them the ideal translators to serve at suits such as the ones conducted at the governor's court.^[182]

Combining the conclusions of Wacke and Eck one can summarize that although Latin was the preferred language for legal acts, foreigners could use their own language for acts within the ius gentium from an early stage onwards. The reality as presented by the documents them­selves shows that Greek was used for the main text of documents, while subscriptions could be in Aramaic, with or without Greek translation. The language used during a lawsuit can reasonably be assumed to have been Greek when all parties understood this, and in cases like Babatha's where the parties only spoke a local language, one should assume that the parties spoke their own language and a translator was used to inter­pret their statements for the Roman judge.

The different approaches by Wacke and Eck serve to illustrate an impor­tant point. Do we approach the questions regarding the languages of the documents from a strict legal angle, asking what language should be used to make a legal act valid in a Roman court context, or do we approach them from reality as it presents itself from the documents? Within the strict limits of the first approach, as Wacke's article shows, Greek is as much a foreign language as Aramaic is: the only ‘real' lan­guage of Roman law was Latin. Only in specific cases is the use of other languages mentioned and sanctioned, but this rather confirms the idea that in general Latin was the norm. This conclusion is obviously rather at odds with the evidence from our documents: there is not a single Latin document among them, what is more, there is not a line in Latin in the texts. In P.Yadin 16 the statement of the praefect is given in a Greek translation, in P.Yadin 28-30 a Greek version of a Latin actio is found. The documents thus give the impression that Greek was the language that the Roman administration and the Roman judiciary system used. If, however, this impression leads to the conclusion that Greek was the language that made legal acts valid in a Roman court context, this can easily cloud our judgment of the part local languages could play in a Roman context, as the following short survey of opinions from other scholars shows.

In relation to the languages used in this archive it has been remarked frequently that the use of Greek should be related to the transition from Nabataean Kingdom to Roman province of Arabia. This means that the transition of Aramaic to Greek as language for legal documents is related to a Roman court context. The way in which these observations are phrased varies and merely factual observations sometimes turn into suggestions or even conclusions as to the relationship between a change in language and a possible change in law. In such cases it needs to be asked whether the conclusions are supported by the evidence as we have it.

Merely factual, and therefore completely acceptable, are observations like

As is well known, in their communications with the subject populations of the Roman Near East, the Romans too used the medium of the Greek language.^[183]

Roman administrators normally used Greek in dealing with the indig­enous populace.^[184]

and, argued the other way around

Although Greek was the lingua franca of government and business in the whole eastern Mediterranean area, it was not the mother tongue of the parties and the scribes of the Babatha documents.^[185]

More of a conclusive nature, but still firmly grounded in the material, is

Thus the intimate connection between provincialization and the use of Greek in legal documents from Nabataea/Arabia is firmly established... Whereas the use of Greek by Jews in legal documents in the province of Arabia is connected with the advent of the Romans, and Romanization filters through the Greek prism, no such association can be made for the use of Greek in Judaea whose provincialization dates to 6 CE...^[186] Obviously there is a link between the Roman conquest and the use of Greek in administrative and legal documents. However, views can differ as to how this relationship should be understood.

Cotton wrote about this on several occasions:

Indeed elsewhere I have argued that the use of Greek in legal documents in the Roman province of Arabia is one of the reasons for believing that these documents were intended for a Roman court of law.^[187]

While this is a valid assumption, as indeed it seems logical to suppose that the use of Greek is related to the desire to use the document in a Roman court context, the next quotes go a step beyond.

Having previously used Aramaic and Nabataean, they now resort to Greek in their legal documents, for no other reason, it seems, than to make them valid in a Roman court of law.^[188]

Nevertheless, I believe that identical reasons motivated the adoption of Greek in legal documents in both provinces [i.e.

Less peremptory but still quite strong is:

Factors other than the hellenization of the writers may well have made the use of Greek obligatory, or at least desirable, in documents of a legal nature, for example, the need to make them accessible (or valid?) in a non- Jewish court of law.^[190]

In these quotes writing the document in Greek is not just associated with a Roman court context, but with making the document valid in a Roman court context. The other side of the coin would then be that documents in the indigenous languages would be invalid in a Roman court context.

Indeed this is the implication of Wassersteins observations regarding P.Yadin 18, a marriage contract in Greek, as

no more than a further safeguard (additional to the kethubbah) for the pecuniary interests of the bride, enforceable in a non-Jewish secular court.^[191]

and regarding P.Yadin 7, a deed of gift, drawn up in Jewish Aramaic, after the Roman conquest:

It is further to be noted that the Babatha archive includes some Ara­maic documents such as sales contracts, certificates of deposit, and most important of all, a bequest of property. In the latter case certainly, and in the others probably, it seems reasonable to assume that enforceability in a provincial court would also have been in the mind of the testator, and it is therefore remarkable that the language used is not Greek but Aramaic.^[192]

Wassersteins interpretation of the alleged validity of Greek in a Roman court context is adduced as explanation for the language usage in the Babatha archive by Hezser in her important study about Jewish literacy in Roman Palestine:

How is the language usage of the Babatha documents to be explained? The Nabataean documents were mostly written before the region became a Roman province and therefore reflect the local custom of the inhabit­ants. The fact that the large majority of the later documents are written in Greek, although Aramaic and Nabataean were the native languages of the involved parties, most of the witnesses, and even the scribes, may per­haps be explained with Babatha and her family's desire to make the deeds enforcible in a Roman court, as Wasserstein has suggested in connection with Shelamzion's Greek marriage contract (P.Yadin 18). One may assume that a Roman court “would prefer or even insist on the use of the Greek language” as the lingua franca of all government and legal proceedings in the East. The possibility to register documents in public archives may have been another reason.^[193]

I note here that assumptions of various orders are combined without a clear explanation of their meaning from a legal point of view. Prefer­ence for a language is obviously not the same thing as insisting on its use. A mere preference could explain for the fact that the majority of documents, but not all, are written in Greek, while no inference can be drawn as to the legal validity of Aramaic versus Greek. Insisting on its use, on the other hand, could be at issue in cases of documents that are completely in Greek while the word ‘translation' indicates that cer­tain sections had been in other languages originally. However, it is clear from a cursory overview of the documents in the archives that in the majority of cases of Greek documents subscriptions are incorporated in Aramaic and not in a Greek translation. Furthermore, it needs to be observed that the documents that are completely in Greek are copies, implying that the original acts did contain the sections in the other lan- guages.^[194] This means that one can hardly infer the demand for use of Greek from the evidence in the archives, let alone, conclude that the use of Greek was necessary to make documents valid within a Roman court context. This may be argued on the basis of specific arguments for spe­cific cases, such as P.Yadin 18, where Wasserstein assumed that also an Aramaic marriage deed had been written, but those arguments should not be used to draw general conclusions seeing to all documents, in any case not without some more detailed discussion of the various problems connected with them.

The interpretation of the language issue, as outlined above, intimately connects language with court: Greek was chosen to ensure the docu­ment could be used in a certain type of court as against another.

This is also emphasized in the following quotes by Cotton: the fact that we have dozens of contracts written in Aramaic means that there was some sort of Jewish jurisdiction in civil cases^[195]

and:

the Jews continued to write contracts in Aramaic after 70 in Judaea and after 106 in Arabia, but they could not expect the Roman courts to enforce these contracts.^[196]

Concerning recourse to certain courts Cotton remarked in other articles:

There is nothing in the documents we have reviewed here to suggest that recourse to Roman law and Roman courts was anything but voluntarily adopted.... No other courts occur in this archive, and there is no good reason for assuming that Nabataean and Aramaic could not be used in a local court.^[197]

In other words it may not have been a matter of necessity to go to a Roman court, but a step taken out of choice. Rome's subjects could and would seek Roman justice whenever they believed that it would be more effec­tive, more advantageous and more just than the local one.^[198]

To summarize: documents in Nabataean or Aramaic could be used in a local court; there are no such courts mentioned in the archives; the people chose to use Greek and turn to a Roman court. Apparently this was a choice they could make.

These assumptions presuppose two divisions: one between Roman and indigenous courts and one between Greek and Aramaic documents. But do these divisions really exist, or, theoretically speaking, is it fruitful to assume they exist?

In this context it is good to recall that Eck accepted that the language used during a lawsuit was the indigenous language which was translated for the sake of the Roman judge.^[199] If translation of the proceedings dur­ing a lawsuit were necessary and accepted, why not accept a legal act in Aramaic as basis for the case to be decided? The same translator who translated the oral statements could read and interpret the document. Why would Greek be more valid than Aramaic, as Greek was, accord­ing to both Wacke and Eck, also not the ‘real' language of Roman law? Besides, Wacke noted, concerning the stipulatio, that it could be con­ducted in Greek and, since Masserius Sabinus, also in other languages. The passage quoted explicitly mentions Syriac, which should be equated with Aramaic.^[200] If one accepts that a stipulatio in Aramaic would have been valid under Roman law, it does not seem likely that a legal act writ­ten in Aramaic would be invalid in a Roman court of law solely on the basis of its language.

Consequently, a closer investigation is wanted, both of the possibil­ity of indigenous courts and their status vis-à-vis the Roman court, and of the use of Aramaic in legal acts after the conquest, to determine the status of Aramaic as a legal language in a Roman context, as it appears from our documents.

Roman vs. indigenous courts

The problem with a division between Roman and indigenous courts is that there is no indication in any of the documents that there actually were local courts. The information from the archives, from the sum­monses and other lawsuit related documents we find there, concerns litigation before Roman courts, de facto the court of the Roman gover­nor.

Isaac, already quoted above, remarked on this:

The absence of any reference to Jewish courts or local officials who might have settled financial disputes between Jews is striking. Indeed, Jewish institutions are not mentioned anywhere in the Babatha archive.^[201]

Cotton, who wrote about the possible existence of indigenous courts on several occasions, initially observed:

It is a remarkable fact though that no court, Jewish or non-Jewish—apart from that of the Roman governor of Arabia—is mentioned in any of the documents from the Judaean Desert—a great many of which are legal documents.^[202]

Her conclusion for Arabia was then:

After 70 conditions prevailing in Judaea became similar to what condi­tions in Arabia had always been: there was no Jewish court which had the authority to enforce its decisions. In Arabia there had never been Jewish courts of law as the exclusive use of Nabataean in the regal period dem­onstrates.^[203]

Here again the use of language prompts the conclusion: because in the regal period only Nabataean is used, there will not have been Jewish courts. However, even if one accepts that this is true, that there were indeed no Jewish courts, this does not say anything about the law applicable to the documents in Nabataean Aramaic. As I will show in my discussion of P.Yadin 6, it is definitely not true that documents in Nabataean Aramaic have to draw on Nabataean law.^[204] Internal evidence, references to law, should be conclusive in this respect and not the doc­ument’s language.

Obviously the use of Greek was not exclusive after the Roman conquest. There are still documents written in Aramaic and in Greek documents Aramaic is used for subscriptions and signatures. This means that Ara­maic continued to play a part as a legal language. This can lead us in two directions: either we assume, as I am inclined to do, that Aramaic could play a part as a legal language in a Roman court context, or, following Cotton's argument about the situation in Nabataea, one has to assume that the continued use of Aramaic indicates that there were indigenous courts. Cotton took this latter line of argument and investigated in detail whether there is any evidence, necessarily from outside the archives, for the existence of such local courts, possibly courts of arbitration:

Perhaps one should think in terms of courts of arbitration acceptable to both parties to the contract of litigation. In the course of time, the Romans, even if not officially recognizing these forms of Jewish jurisdiction, none­theless came to tolerate them.^[205]

Here Cotton bases herself on evidence from the rabbinic sources that mention ‘courts of different sizes in towns and villages' as well as discus­sions about ‘the validity of contracts made in Greek, about the use of gentile witnesses, courts and archives.'^[206]

In a later article Cotton investigated the possibility of the existence of local courts (perhaps of arbitration) in more detail.^[207] Although this possibility cannot be excluded, it seems that the suggestion is sooner prompted by speculation (about the Roman courts' workload if all cases went there and so on) than by facts. Cotton's suggestion of thinking of other possibilities for conflict settlement than ‘formal iuris dictio of a court of law' is attractive: indeed, it makes sense to assume that people looked for ‘other solutions, less cumbersome, less expensive and less time-consuming.'^[208] Yet Cotton's argument that evidence to such a prac­tice can be found in P.Hever 63 is not very compelling.

Cotton assumes that the renunciation of claims as described in this act may well be the result of a dispute settlement in front of a arbiter. This could be, but even if it was, this says nothing about the status of this type of jurisdiction, indeed whether it should be called jurisdiction at all. No matter how the dispute was settled, the legal validity of the renunciation lies in the written act, not in the possible arbitration that preceded it. In this context I note we have no documents pertaining to this case (or indeed any other) of arbitration. What is more, as Cotton acknowledges herself, the act is written in Greek and contains a stipulatio clause, which makes it comparable to other acts from the archives that were clearly intended for a Roman court context. Cotton then states that the parties might have wanted to keep recourse to the Roman court open, imply­ing their initial legal obligation would have arisen from the assumed arbitration. However, as far as I can see, the obligation arises from the recording in the legal act of the dispute settlement reached. And there is nothing in the legal act to suggest it was envisaged as subjected to any other kind of jurisdiction but that of the Roman governor. Therefore, P.Hever 63 does not contain conclusive evidence as to the existence of local jurisdiction (in the form of arbitration), and it certainly does not position this jurisdiction as an alternative to Roman jurisdiction. On the contrary, it appears that a dispute settlement (perhaps through arbitra­tion) gained validity through a legal act that was subjected to Roman jurisdiction.^[209]

What is more, regardless of whether one finds the evidence for the existence of such local courts compelling or not, there is a technical- juridical problem with assuming that these local courts existed and had jurisdiction to judge certain cases. Both Cotton, and Wasserstein, in his assessment of P.Yadin 18, quoted above, assume that local courts had little power to actually have their decisions enforced which would mean that someone turning to those courts would have to depend on the willingness of the other party to subject itself to the court's ruling.^[210] This implies that drawing up a contract in Aramaic meant for the par­ties involved that they subjected themselves to the jurisdiction of a local court.^[211] This is clearly the conclusion by Safrai:

If a Jew wanted a document between him and his fellow to have legal validity, he was forced to write the document in Greek, and in a manner that would meet the requirements of the court in Petra or in Rabbah. One who wrote his document in Aramaic thereby decided that he would not need the official courts. He did so either out of naivete and good will, or because he relied upon another, unofficial, court, probably a Jewish one. The Shelamzion who received an Aramaic marriage contract knew that she would not have any legal recourse to the Roman court and apparently relied upon another court, or possibly upon a second marriage contract written in Greek.^[212]

Theoretically, this would be possible, but I believe that in reality the pos­sibility is excluded by the evidence provided by the acts in the archive.

In P.Yadin 10 Babatha has her marriage contract drawn up in Ara­maic. However, when an argument arises with her deceased husband's first wife Miryam, Babatha turns to the court of the Roman governor and not to some kind of local court (P.Yadin 26). It is likely that her claim in the case is based on her (Aramaic) marriage contract, which means that Babatha approaches the Roman governor in a case founded on rights established in an Aramaic deed. In any case it is clear that Babatha bases her rights recorded in P.Yadin 21-22 on her Aramaic marriage contract. There Babatha bases her right to sell dates from orchards that belonged to her deceased husband on her dowry, thus basing a deal laid down in a Greek contract on a right acquired through an Aramaic contract. If jurisdiction was determined by the language of the contract, this would be impossible. After all, the rights of the marriage contract in Aramaic would be subjected to local jurisdiction, while the contract based on those rights would be subjected to Roman jurisdiction. What the actual evidence from the archives indicates, is that an Aramaic contract could be produced as evidence in a Roman court context, or in any case that rights derived from such a contract could be subject to Roman jurisdic­tion. Consequently, one has to accept that parties to a deed in Aramaic had not subjected themselves (exclusively) to local jurisdiction.

Following this line of argument, a determinative relationship between language and court/jurisdiction is untenable. The existence of legal acts in Aramaic is in itself not enough to prove that there was local jurisdic­tion. On the contrary, it seems more likely that the Aramaic deeds did function in a Roman court context. Therefore, it is safer to concentrate on the Roman court context and try to understand what the use of sev­eral languages meant in that context.

Greek vs. Aramaic

It seems that the sole reason for assuming that the Aramaic docu­ments were meant for local courts is that they are in Aramaic and not in Greek.^[213] If it is argued that a document has to be in Greek to be valid in a Roman court, the logical conclusion is that documents that are not in Greek cannot be used in those courts.^[214] However, the very fact that no evidence for the existence of local courts is found suggests that the Aramaic documents were actually used in Roman courts, or at least people believed they could use the documents for that purpose.

In a review of DJD XXVII (which incorporates the Salome Komaise archive) Bagnall observed:

The use of Greek in these legal texts, Cotton argues, reflects a desire to have legal acts easily recognized in Greek-language courts, i.e., Roman courts; the only court actually mentioned is that of the Roman governor of the province of Arabia. This may be true, but it is all the more striking that under Roman rule legal documents were also written in the local lan­guages; see further on this below.^[215]

and below:

The use of languages here (and in the still-unpublished part of the Babatha archive [i.e., the Aramaic part of the Babatha archive, JGO]) certainly shows that Aramaic remained usable in legal documents under Roman rule and that competent scribes in both languages were available in at least the more important villages; we still do not know accurately why one lan­guage was chosen for one document, the other for another. There is no evidence that the Romans discouraged the use of Aramaic.^[216]

An important point to take into account in this respect is that even in the Greek documents subscriptions are still in Aramaic. Cotton takes this as a sign that the picture painted by the languages used in the documents need not have been true to the real linguistic situation. Consequently, it is in her opinion wrong to view the Jews involved in the papyri as hel­lenized or even semi-hellenized.^[217] I agree with that completely, since the documents are legal documents. The language there serves a completely different function than it does in everyday life.

It is precisely this special character of the legal document that poses the problem with the use of Aramaic in the subscriptions. Considering that a legal document has the function to establish, or at least record, rights and obligations, such a document establishes legal ties. A sub­scription is the part of the document where the party testifies to his obligation to the arrangement. Consequently, a subscription can be con­sidered to be an essential part of a legal document. If Greek was used to make the documents ‘valid' in a Roman court of law, why have the parties' subscriptions in Aramaic? Why not always include them in a Greek translation as is done in P.Yadin 11 and 16, or at least add a Greek translation to the Aramaic subscription, as is done in P.Yadin 27?^[218] A comparison with the situation in Egypt is illuminating.

In an article about autograph confirmation in Demotic contracts DePauw touches upon a drastic change in languages used in legal docu­ments in Egypt at the moment of the Roman conquest.^[219] Before that moment typically both the main text of a document and its subscrip­tions, if any, were written in Demotic. After the conquest subscriptions were written in Greek, regardless of the language of the main document. To put it differently, contracts continued to be written in the indigenous language, while for subscriptions the language of the new dominant power was used. Documents that are completely in Greek are most likely translations of originally Demotic contracts with Greek subscriptions.^[220] There are a few contracts that have a subscription in Demotic but this is never the first subscription. Furthermore, subscriptions become more elaborate and both parties subscribe, something that was unusual in earlier times. Subscriptions also occur far more often, where previously even a mere signature was but ‘an optional feature.'^[221] This shows that a greater importance was attached to subscriptions. DePauw argues that all of these changes have to be related directly to the Roman conquest and more precisely to Roman requirements:

the evidence suggests that the first party had to sign the contract in Greek before a document could be registered in the grapheion. Demotic could still be used in the subscriptions, but only in addition to the Greek, or for the declaration by the beneficient [i.e. the second subscription, JGO]. Neither of these had the same importance for the validity of the contract as the subscription of the first party.^[222]

This marked difference between the documents drawn up prior to and after the conquest shows that in Egypt under Roman rule subscriptions began to play a more important role in documents and that the use of language was related to this. The mandatory use of subscriptions, where first a signature had not even been obligatory, and the elaborate char­acter of the subscriptions suggest that those were deemed to constitute important, if not essential, parts of the contract. It seems inevitable to relate the use of Greek for party subscriptions to this alleged impor­tance: just because the subscriptions where the parties agreed to their obligations were considered to be the essential, constitutive parts of the document they had to be written in the language of the dominant power in the area, or more general, the lingua franca.

The situation in the Judaean Desert documents should obviously be assessed differently. Instead of maintenance of the original language for the main text of the document and change to Greek in the subscriptions we find a change to Greek in the main text of the document and mainte­nance of the indigenous languages in subscriptions. This is remarkable because the evidence from Egypt suggests that the use of language was related to the significance of the party declarations: they constituted the binding parts of the contract.

A single instance of a Greek contract from Egypt where the first declar­ant's subscription is in Demotic prompted the suggestion that subscrip­tion in Demotic was allowed when the main text of the document was written in Greek.^[223] This would suggest that in Egypt two types of docu­ments would have been valid: Demotic documents with Greek subscrip­tions or Greek documents with Demotic subscriptions. The latter type would obviously resemble the Judaean Desert documents: Greek docu­ments with Aramaic subscriptions. It would then be deemed sufficient to have at least one part of the document in which the parties' obligations were expressed in the language of the dominant power, i.e. Greek. I note, however, that in the Egyptian material this occurs but once: in a com­parable instance of a Greek document with a Demotic subscription the Demotic subscription is accompanied by a Greek one.^[224] In the Judaean Desert documents the subscriptions were written in the indigenous lan­guages usually without any form of translation into Greek.^[225] Only in copies were the subscriptions incorporated in a translation rendering a completely Greek document.^[226] This means that it seems that the Roman conquest led to different developments of the legal languages used in the Judaean Desert area than it did in Egypt. The continuing use of Aramaic for the party subscriptions seems to suggest that Aramaic could play some part in a Roman judicial or administrative context. The difference with the situation in Egypt is all the more relevant for our documents as DePauw concluded for that Egyptian situation that ‘the legal value of the Demotic part [of acts] was thus reduced to next to nothing' and that ‘Demotic was thus degraded to a second rank language.' The role of Aramaic in the subscriptions to the Judaean Desert documents suggests that there the indigenous languages did not suffer this same fate. There­fore, far from assuming that Aramaic became replaced by Greek as the language for valid legal acts, we should accept that Aramaic continued to play a part as a legal language.^[227]

The case of the subscriptions at any rate proves that there had to be some kind of knowledge of Aramaic in a Roman court. Otherwise, the judge would not be able to understand what the parties had promised to one another, as he would not be able to compare the subscriptions with the main text, or in any case read and interpret the entire docu­ment. There may be two documents in the archive that are completely in Greek, giving a translation of the originally Aramaic subscriptions, but these are exceptions.^[228] In the majority of the cases the documents include Aramaic subscriptions while they are clearly meant for use in a Roman court context, for example the summonses of P.Yadin 14 and 25, and the proposed settlement of P.Yadin 15. It is important to note that these documents were not, like legal acts in general, drawn up with the idea that they might at some future time in a dispute situation have to be produced in a court context, but they were drawn up when the dispute had already arisen, i.e. when it was already clear to what court the par­ties wanted to turn. Consequently, it cannot be maintained that docu­ments intended for Roman courts had to be written in Greek.^[229]

In view of the Egyptian situation the cases in the Babatha archive where the document is entirely in Greek seem to further complicate our understanding of the language issue. For these two cases, P.Yadin 11 and 16, it has been argued that both represent copies of official acts, assum­ing that the original acts may have included Aramaic subscriptions.^[230] At first sight this seems logical: an act was drawn up with the parties' subscriptions in their indigenous language, while the copy was made up entirely in Greek for the sake of uniformity and accessibility. But why was this copy apparently given to the parties and not kept by the authorities? It is found in the archive of the parties involved. Alterna­tively, taking the reverse argument, what happened to the original act containing the party subscriptions in Aramaic? The obvious answer should be: this original was kept by the authorities, or in the case of the loan in P.Yadin 11, by the creditor, the Roman centurion. This presents us with a seemingly contradictory situation: the authorities—who are presumed to have little knowledge of Aramaic—keep an original docu­ment that contains Aramaic subscriptions, while the parties, who did not know Greek, are supplied with copies that are completely in Greek! Of course, this latter practice can be explained from the viewpoint of the Roman authorities: if they had completely Greek copies of acts drawn up, for uniformity's sake, we can hardly expect them to have taken the fact that the parties could not understand Greek into account. In this light I emphasize that Babatha could not even write (and presumably read) Aramaic.^[231] This already indicates that the use of documents is cer­tainly not directly related with literacy of the parties in the languages concerned. It is therefore not so much the fact that the copies are in Greek that is extraordinary, as the related conclusion that the original acts kept by the authorities contained subscriptions in the original lan­guages. Since such acts, even in a thoroughly Roman administrative matter like the census, were drawn up in several languages, it cannot be maintained that legal acts by private parties had to be drawn up in Greek to make them acceptable in a Roman context. Furthermore, the context of P.Yadin 16 shows that the use of Greek is also not directly related to the need to register the document in a public archive: despite the obvi­ous need for this in a census context the language used for subscriptions in the original document was not Greek.^[232] This supposes an acceptance of Aramaic as a language for legal acts, although it was clearly not used to represent the document's main text.

The observations made above, when put together, present the following picture: after Roman rule a gradual process led to the practice of having legal documents drawn up in Greek, maintaining party subscriptions in Aramaic. Contrary to Cotton's repeated conclusions, the apparent desire to have the main text written in Greek cannot be explained exclusively by referring to a Roman court context, as the indigenous language was still used for parts of the documents. Consequently, it cannot be maintained that Aramaic was not valid as a legal language either. Documents that are completely in Greek are copies of acts that were originally drawn up with Aramaic (and in some cases Latin) subscriptions. Those originals were apparently kept by the authorities, denoting that the indigenous language kept on playing a part not only in the context of jurisdiction but also of administration.

Choice of language is choice of law?

Above it was shown that there was no direct determinative relationship between language and court, that is, that the languages used do not indi­cate to what jurisdiction the parties involved subjected themselves. A related, but not identical, question is whether there is a direct and deter­minative relationship between language and law. Does the language cho­sen for the contract determine what law is applicable to the contract?

The question is related to the question of a relationship between lan­guage and court, because one could assume that language indicates the court one wishes to turn to and therefore also the law one wishes to apply to the contract. For example, it would be logical to assume that if the choice for Aramaic implied subjection to local jurisdiction, this would include application of local (perhaps Jewish) law.

But if a choice for Greek was related to subjection to Roman jurisdic­tion, does this also imply subjection to Roman law on a deeper level?

A direct determinative relationship between language and law suggests itself from a comparison with bilingual family archives from Hellenistic Egypt containing Greek and Demotic documents.^[233] In discussing the Tatehathyris archive Pestman maintained for this Egyptian material that the choice of language can directly be related to a choice of law: Greek documents seek to connect with Greek Hellenistic law, Demotic documents with Egyptian law. This can for instance be seen in the use of guardians: Greek documents present women acting with guardians, in accordance with the requirements of Hellenistic law. Demotic docu­ments, on the contrary, do not mention guardians, which is consistent with Egyptian practice.^[234] According to Pestman, those documents in the archives concerning marriage and divorce seem to be drawn up in Demotic, because of the more favourable position of women under Egyptian law.^[235] This implies that language was chosen to put the docu­ment in a certain framework, or argued the other way around, that the language was determined by the law one wished to apply to the act.

The same idea is phrased in Isaac's observation on the Greek marriage contract of P.Yadin 18:

Any social changes in this period must be seen against the background of the imposition of Roman provincial administration. If Jews in Arabia preferred Greco-Roman marriage contracts to traditional ketubbot, the most likely explanation is that the former offered advantages that had been unavailable under Nabataean rule. The position of the woman, for instance, is more favorable in ‘Greek law' (document 18, ll. 16, 51).^[236]

The question that should be raised here is: does P.Yadin 18 connect with Greek law, because the document is written in Greek, or because the document clearly refers to Greek Hellenistic law as the applicable law to (part of) the arrangements? To put it differently, is the relationship between language and law determinative or should internal evidence decide what law was thought to be applicable to these acts?

In this light it is worthwhile to repeat Wassersteins remark on P.Yadin 18 already cited above, indicating that it may have been

no more than a further safeguard (additional to the kethubbah) for the pecuniary interests of the bride, enforceable in a non-Jewish secular court.

This interpretation does not regard the Graeco-Roman marriage con­tract as a replacement of the ketubba, but as an addition to it. The link between the additional character of the document and its supposed enforceability in a non-Jewish court suggests that a ketubba would not have been enforceable in such a court. But does this mean that the Greek document is merely a Greek version of an Aramaic document, the lan­guage required to have the document enforced in the Roman court con­text, or does the change in language imply more?

The first option, a Greek version of an Aramaic document, seems to apply to a document like P.Hever 64, which has been plausibly argued as being a Greek translation of an Aramaic ‘Urtext.’^[237] Here the contents of the document do not seem to be affected as the gift is made in the same terminology as the one of (Jewish Aramaic) P.Yadin 7.^[238] This could denote that only the language of the documents changed but not the law and legal context to which they referred.

However, if a Greek document explicitly determines that Greek (Hel­lenistic) law is applicable to (part of) its arrangements, does this mean that the legal context changed with the language? If so, is this not due to the reference to law, rather than to the language used? To put it dif­ferently, would Greek (Hellenistic) law have applied to the contract, if it had been written in Greek but had not incorporated a reference to Greek law? The comparison with the case of P.Hever 64 seems to imply that internal evidence should determine what the applicable law prob­ably was, rather than the language used.

Concerning the use of Greek by Jews in general Wasserstein observed that using Greek does not always ‘point to direct acquaintance with the Greek sources.'^[239] To put it differently, the scribe may use a Greek expres­sion without being aware that it stems from a certain source. It is obvi­ous that this calls for caution when drawing conclusions from the fact that the documents use Greek phrases: we do not know whether this meant that there is a direct relationship with other Greek documents or something like a Greek (Hellenistic) tradition. Wasserstein continued to call for caution when he pointed at the occurrence of Greek loanwords in both Jewish Aramaic and Syriac. Even though he was there speaking about a later era, his observations are obviously relevant here:

This suggests, not that the Rabbis had borrowed these words directly from Greek, but rather that they found them ready-made, readily available, in the Aramaic koine, which they shared with their non-Jewish, non-helle- nized, non-Greek-speaking neighbours, not only in Palestine but in the whole region both before and after the Christian period... It was the com­mon Aramaic inheritance, the common Aramaic language, that served as the principal conduit for hellenistic influences on non-hellenized Jewry. Thus, hellenistic elements in non-hellenized Palestinian Judaism can, paradoxically, be seen not as deliberate and conscious adoption of for­eign, Greek, ways, but, on the contrary as a sign of belonging to the home­grown culture of the Aramaic East within the Empire as well as outside its borders.^[240]

Consequently, in a quest to say something about law behind the docu­ments it is essential to question in all cases what legal context the par­ties and foremost the scribe wanted the document to conform to, and in what way he expressed this in the documents' use of terminology. Instead of focusing on language in the discussion, internal evidence should be considered to see whether this indicates what law was thought applicable to the document.