1The premise of this paper is a self-evident statement: neither Greek nor Roman religion can be considered as a “religion of the Book.” In the common sense, they are not systems based on a sacred scripture encapsulating a revelation and giving birth to dogmas. But this premise must be immediately qualified because writings related to religion were not absent in these cultures. Orality coexisted with writing during centuries: sanctuaries were full of inscriptions and graffiti, while written ritual norms are attested in Greece and priestly books in Rome. We will respectively focus on the ritual prescriptions preserved on stone in Greek sanctuaries, and in the so-called “books of the priests” in Rome. The shared background of both parts of our study is the question of religious authority and “codification” with regard to ritual performance.

About Greek ritual norms (V. P.-D.)

2In ancient Greece, there is nothing comparable to the priestly books or commentarii of ancient Rome, which John Scheid addresses below, and such writings do not seem to have ever been the rule. Some “books” are mentioned in connection with the so-called “Orphic” rituals but the issue is much discussed and scarcely evidenced.¹ The religion of the Greeks was definitely not a religion of the book. Nevertheless, with the adoption of alphabetic writing in early archaic times (eighth century BCE), the Greek city-states progressively used this tool in order to endorse their decisions and publicly display them when necessary. Among the wide range of measures adopted by the cities, one finds decisions related to gods, sanctuaries, or rituals, made for the well-being of the community.

3Unfortunately, official Greek archives are almost completely lost but we know that some of them contained a section devoted to regulations concerning “sacred issues.”² What is now left of the measures taken in the area of religion is a very small part of the decisions once written down: those inscribed on durable materials, such as stone, more rarely metal, and found by chance in excavations. In addition, inscribed ritual norms were never considered as “practical guides” for concrete cult performance. Had we access to all the Greek civic archives related to rituals, we would still not grasp most details of ritual processes. I will return to this point below.

4Does this mean that “practical guides” for rituals did not exist in ancient Greece? Absolute certainty is impossible, but no evidence of this kind has been preserved. We do know that treatises about rituals were composed during the Hellenistic period. Some titles and the names of authors are known, but not a single work of this kind has been safeguarded. Only scattered pieces are available. They were gathered by Aloïs Tresp in 1914 in a book significantly entitled Die Fragmente der griechischen Kultschriftsteller. When reading these fragments, we can observe that the treatises to which they originally belonged were the products of learned inquiries rooted in scholarly erudition. As they were not prescriptive, they probably did not serve as “practical guides” to be used for a proper performance of rituals. They were descriptive and expository. We do not know their sources of information. Any Quellenforschung is impossible to carry out since almost everything is lost. We can always dream of a manuscript or papyrus still to be uncovered which would provide such an exegetikon, but this hope is unreasonable and unwise.

5Epigraphic prescriptions related to rituals give us a rare opportunity to collect pieces of information, which would otherwise have completely vanished. However, epigraphic documents referring to rituals can have many different objectives, depending on the kind of text we have to deal with. The material gathered in the study of Greek ritual norms is embedded in documents as varied as civic decrees regarding cult, sacrificial calendars, sales of priesthoods, familial foundations, funerary laws, oracles or even boundary stones and dedications. The authorities issuing these documents can be cities and their subgroups, families, associations, or even individuals, in some cases where a brief cult regulation is connected to the dedication of an altar.³ Accordingly, we must be conscious that our knowledge of Greek rituals, as they were effectively performed by local communities, is partial and severely limited –as is also the case for Rome, even though other types of evidence are available. We can therefore return to the question raised earlier: what kind of information is given in epigraphic ritual norms and why was it written down?

6Due consideration must be given to the oral and unwritten character of Greek ritual tradition. What is recorded in the epigraphic evidence is what is exceptional or at least noteworthy and worth writing down,⁴ even if we are often unable to reconstruct the background of such a need. When the evidence gives some keys to understanding what happened, political transformations impacting the religious life of groups can be identified (for example, the foundation of a new city by gathering former ones in a process of synecism), or financial pressures necessitating the cost-sharing of sacrifices. For instance, in the case of a civic decree issued by the city of Kos, on the eponym island, containing the codification of various rules of purity for priests and priestesses, the text itself states that it is inscribed, “so that the purifications and cleansings are performed according to the sacred and ancestral customs (or traditions).”⁵ These sacred norms to which the text refers were previously recorded in archives, but we do not know exactly why the inhabitants of Kos felt the necessity to publicise them on several stelae, erected in conspicuous places of worship. The extant stele recording such purity rules is heavily damaged, but what is readable shows that the rules it contained were more precise than many other documents referring to rituals, perhaps because purity became a sensitive issue on the island at some point during the Hellenistic period.⁶

7Purity rituals are well attested in our evidence, but the main ritual performed by Greek communities for over a millennium was undoubtedly animal sacrifice: thusia. As far as thusia is concerned, Greek ritual norms clearly confirm that a range of actions to be performed were not made explicit because, at a fundamental level, this kind of sacrifice involved a basic series of actions that were part of common knowledge and did not need to be spelled out. The possibility for private individuals to offer sacrifices in sanctuaries without need of a ritual expert or a priest is a strong confirmation of this basic level of knowledge.⁷ When new configurations emerged in a group, then the necessity of specifying at some level of detail what had to be done in ritual context arose: for example, gathering several settlements in one city required that the religious life of the new city be organised in such a way as to provide the same degree of information and knowledge to all its inhabitants. However, the basic elements presumed to be known by all the members of the previous communities were left implicit.⁸

8An interesting question, as far as Greek ritual norms are concerned, is the scale of the common knowledge presupposed by these documents: can we identify local, regional or even Pan-Hellenic levels of shared knowledge? In other words, is it possible to recognise something typically “Greek” in the sacrificial ritual performed by Greek groups, beyond the ritual specificities of these particular groups on a local or a regional level? A linguistic metaphor will help to clarify the issue even further: does a common Hellenic sacrificial language exist in parallel with its multiple dialectal expressions?⁹

9An inscription newly discovered in ancient Thessaly, in central Greece, gives a good opportunity to briefly address this issue. The text, dated from the third or second century BCE, belongs to the category of “epigraphic ritual norms” insofar as it prescribes a range of sacrifices to be offered to various gods. The authority issuing the document, if any was mentioned, is now lost in the damaged parts of the stele, but a vast majority of the prescribed sacrifices and some of the gods mentioned are so exotic that we can reasonably attribute this regulation to an association mixing together Syrian –or other Near Eastern people– and Greek members.¹⁰

10For the first time in epigraphic evidence written in Greek, we find the expression: “to sacrifice according to the Greek norm” (or tradition, or custom), depending on the translation chosen for the Greek expression: thuein… Hellenikōi nomōi.¹¹ The expression is unique in epigraphy and exceptional in other textual evidence. The procedure is described in detail: any sacrificial animal can be offered except a pig (an exception which is actually not Greek); some cakes are to be deposited on the table, as well as olive oil for a lamp, wine, the boiled breast and one raw leg of the animal. Some of the innards are cooked for the priestess (liver, lungs, diaphragm, left kidney and tongue), while the right kidney, right “extremities” (we do not know what is implied), heart, omentum, the front leg, and a part of the tail must go into the fire, on the altar.¹² The same ritual process seems to be followed, at least on some points, when sacrificing sheep, as well as bovines.

11In other inscriptions including elements related to sacrifice, issued by Greek groups such as cities, families, associations, etc., one of the main objectives of writing down prescriptions is to guarantee that each participant in the ritual –whether a divine recipient or human agent– is to receive his or her due. In the case of the “hybrid” Thessalian group, the same preoccupation probably underlies the regulation, but we also have to take into account the necessity of informing people who do not share the same level of ritual knowledge, which is at the heart of this investigation.

12One last point must be addressed before moving on to Rome. What is specifically “Greek” in the “Greek way of sacrificing” prescribed in this text? Some interpreters consider that it refers to sacrifices of sheep, bovines, and larger animals in general, in contrast with the sacrifice of birds or goats more common in the various Near Eastern offerings documented in the text.¹³ This is perhaps partly the case. But a close comparison between all the sacrifices mentioned on the stele shows a major difference in this case: cutting of portions from the animal carcass and putting some of them into the burning fire of the altar for the gods. The other sacrifices mentioned on the stele imply either the deposition of portions on a table without burning, or complete burning of the animal, what is called a holocaust. The “Greek way of sacrificing”, in this text, is a middle ground between these two types of sacrifice. Of course, we know that holocaustic sacrifices were also performed in Greek cult, but they were very rare¹⁴ and, in the present text, they are not considered as “a Greek way” of performing a sacrificial ritual.

13In any case, the necessity to label this ritual as typically “Greek” emerges from the mixed composition of the association, which probably lies behind the regulation. We would be in a better position to assess the situation if we had other sacrificial regulations issued by the city in whose territory the sanctuary of the association was built. Unfortunately, this is not the case and the comparison can only be made with sacrificial rules issued in other regions of the Greek world.¹⁵ The outstanding character of the Thessalian document lies in the level of detail achieved by the description of the divine part burnt onto the altar. Nowhere else in our evidence do we attain such a degree of precision, probably because the composition of the divine part to be given to a deity in its sanctuary normally belongs to the shared knowledge of the community offering sacrifices.

14Libri sacerdotum or commentarii of some sort would have been wonderful tools for scholars trying to understand the sacrifices offered in the various Greek cities, but, as mentioned above, this kind of book probably never existed in the Greek world. Rome is not Greece, even if, at some point, Greece became part of Rome, to which we now turn.

The books of the Roman priests (J. S.)

15As mentioned in the introduction of this paper, the traditional religion of the Romans was a religion of the written word, without being a “religion of the Book.” There were nevertheless books of some kind, which are known as the books of the priests (libri sacerdotum). But even if modern historians have for a long time dreamed, like their Roman colleagues, that the books of the priests would reveal rare facts about archaic Rome, or help us discover arguments for a representation of the religion of the Romans, which would not be very far from that of a religion of the Book. But the relation of Roman religion with writing has in itself attracted very few modern historians. Only a handful of relatively recent studies have been devoted to this question.¹⁶

16The problems posed by the relations between writing and religion are manifold, and there is no question of examining them all. By examining the Roman priestly books, I do not intend to question their existence or the few elements that are preserved, but rather to question their essence and their raison d’être. Current studies simply take up traditional positions, examine the problem globally, or, finally, explain the evolution of a single type of writing (for example, the so-called commentarii¹⁷ of the arvals, the annual reports kept by these twelve priests who celebrated a State cult in the honour of an agrarian goddess, Dea Dia). It is therefore recommended to face these formidable questions with the necessary caution. Priestly books are not the only written documents concerning religion, but as they were at the centre of the Roman religious and scholarly tradition, it seems reasonable to begin the investigation with them.

*

17The existence of these books, attested by many sources, is beyond doubt; what is less assured, however, are the reconstructions that we want to make of them. What was a Roman sacerdotal book? Two opinions confront each other since the beginning of the nineteenth century at least: some consider the books of the priests as normative books expressing among others the rules of worship;¹⁸ others regard them as accounts of the activities and decisions of the colleges.¹⁹ In fact, recent studies have proved that this kind of “Urbücher” or “Ritualtexte” actually have never existed, and are largely an invention of modern historians. Now that conjures up new questions: why do the Romans mention libri sacerdotum, “books of the priests,” and what is a Roman sacerdotal book? The most logical and reliable way is to start from evidence, including direct sources, which have generally been excluded from the survey, collections like the commentarii, “the reports,” of the arval brothers, and the two commentarii of the Games, that were miraculously preserved.²⁰

18These exceptional documents are commonly referred to as the commentarii (or acts) of the arvals or of the quindecemviri. Although the inscriptions of the quindecemvirs have retained the title of Commentarium ludorum saecularium, and commentarii are attested on the reports of the arvales, this conception, without being wrong, is too sketchy. Epigraphic documents are only transcripts and should not be considered a priori as complete versions of the priestly commentaries concerned.

19The two commentarii (“records”) of the Secular Games clearly show this. The college of the quindecemviri, the “Fifteen men for the consultations of the Sibylline oracle,” was responsible for managing and executing the oracles delivered by the Sibylline Books. This means that besides the consultation of the Books, the preparation of the oracles and the eventual celebration of the prescribed rituals, this college was also in charge of supervising the execution of the recommendations made by the oracles and accepted by the Roman senate. As a result, the daily tasks of the quindecemvirs far exceeded the celebration of the Secular Games, and it is evident that the documents of 17 BC and AD 204 are only excerpts from their commentarii: this is also what their title explicitly says –at least in the case of the 204 reference, since it speaks of the “commentarium on the Seventh Secular Games” and not “commentarium quindecemvirum.” Nevertheless, it is undeniable that the extract itself is complete. Besides the documents give the reason for this transcription on perennial media of the extract concerned: the reason invoked is the conservation of memory. In 17 BC, a special decision of the Roman senate decided that the “record of the ludi saeculares should be written on a marble and a bronze stele and ordered that [the two stelae be erected] on the place where the Games would be celebrated to keep in the future the memory of the event…].”²¹

20The records of the arvales brothers –established since around the years 21/27 AD– consist in a more or less complete transcription of the activities of the brotherhood; but before these years the arvals only transcribed relatively short summaries. Certain rubrics describing rituals that have probably been celebrated since the creation of this cult were included in the copy only a century later: thus, the expiatory sacrifices (piacula) concerning the introduction of iron tools in the grove of Dea Dia for the engraving of the inscriptions, and their exportation once the work accomplished, appear on the epigraphic reports only from the year 81 on. The reasons which determined the annual transcription of the commentary of the arvals on the walls of the grove were different from those concerning the Secular Games. An examination of the records of the first decades of the Empire shows that the decisions leading to the result we know were gradually taken. The first series of transcriptions of the commentarii remained in line with what was found in most of the public sanctuaries of the time, that is to say, fasti, a calendar. The festive calendar of the Julian year was displayed in most of the sanctuaries with the list of the consuls and the praetors, in order to give the date of the ritual duty to be celebrated in the sanctuary. Now, as the date of arval sacrifice to Dea Dia was mobile, it could not appear on a festive calendar; this is why the arvales decided to complete each year the calendar by an extract of their commentarii giving, year after year, the date of the sacrifice. The formula chosen then compelled them to include in the extract of the commentarii the report on the election of new members of the priesthood; and, of course, the name of the annual president of the college, thus saving the creation of one or two other complementary lists of priests elected and the sequence of the presidents. Once put in place, this custom developed. The transcript progressively comprised decrees of the brotherhood, the date of the public vows in January which, at the beginning of the Empire, was not inscribed on the regular calendar either. Finally, the transcription was so complete that from the end of the principate of Tiberius (AD 14–37), there was no reason to omit the other rituals celebrated by the arvals. From that date on, there were variations in the drafting –sometimes succinct, sometimes prolix– of these annual commentarii, depending on the mood of the secretary, or the period,²² but the general structure of the documents remained the same until the third century AD.

21The epigraphic transcripts of the records of the arvals nevertheless provide details regarding how the commentarii were kept. We know that the support of the annual commentary was a codex, held by the president or by the commentariensis, the secretary. These codices were obviously archived and deposited in the sanctuary of Dea Dia, or in a room located in the grove. Since AD 134 (and probably already once in 109) during an interruption of the circus races, the arvals questioned business concerning their liberti or public slaves. To find answers, they consulted these codices and even read them entirely (perlectis codicibus) on the spot.²³ It is obvious that the verifications were not made by consulting the wall inscriptions, as they could not be called codices. These passages thus show without any ambiguity that the arvales had in the sanctuary of Dia a tabularium, a library of codices, which constituted the only possible reference for the priests. In other words, the codices of this archive, “on which the decisions of the previous fratres arvales were recorded,” formed the commentarii of the arvals. The inscriptions probably did not give an exact representation of these codices, since they were each year formatted for their “publication,” whereas the waxed tablets (then bound in codices) were written day by day and in cursive form (hence including transcription errors of the scribe when making copies on marble), but we can assume that the commentarii are closely related to the text of the codices.

22It should be added that the arvals records show that the colleges also used prayer books, and perhaps a booklet with a list of members or a formula recording the election of the annual president of the brotherhood.²⁴ But it is not appropriate to equate these booklets with the priestly “Ritualtexte” as G. Rohde imagined. These prayer texts are in fact instruments of worship, allowing the recitation without error of the prayer or election formulas.

23There is no reason to consider that only the arvals had commentarii, and, of course, prayer books. The quindecemviri, as we have seen, also possessed them. This proves that it was not an exceptional practice linked to the Augustan “restoration” of the relatively obscure fratres arvales, but of a traditional procedure of all sacerdotal collegia. This deduction is confirmed by the existence of a commentariensis sacerdoti(i) septemuirorum, “a secretary of the priesthood of septemviri epulonum,” one of the major Roman priesthoods.²⁵

24Despite the differences in the functions of the two priesthoods concerned, the content of the commentarii of the arvals and the quindecemvirs is similar. Under a title including the consular dates and the indication of the annual presidencies are displayed descriptions of the rituals performed, attesting the execution of the duties on the day prescribed by custom, the sententiae, “decisions” –to use the term employed by the commentariensis– voted by the priests on all matters falling within their jurisdiction, or documents communicated to the college. Thus, the beginning of the commentarii of the Secular Games is formed by a succession of senatus-consulta relating to the organisation of the Games;²⁶ the inscriptions of the arvales record at least one decree of the pontifical college²⁷ as well as a communication of the praefectus annonae of the year AD 80 attributing to the arvals places in the Colosseum.

25But if one examines the known excerpts of the other priestly books,²⁸ one realises that they all could find a place in the commentarii of the arvals or the quindecemviri. Most may be excerpts from a decree or a responsum; some formulas may have been borrowed from prayer texts recorded in religious service records, or copied from prayer booklets. Most of these fragments were extracted from their context by the Roman antiquaries, who quoted only the rule or expression that interested them, ignoring the proper elements of the priestly decree. One can, of course, imagine that antiquaries exploited decrees that enacted sets of religious rules, in short, great rituals. But nothing proves it, and above all nothing allows us to establish that this kind of text would be older than the other documents. In Rome religious regulations that contain all the prescriptions for the cult of a god or goddess are very rare, if not non-existent. As such, the Books of the Roman priest were not a collection of norms.

26Let us take two examples. The first is provided by the list of prohibitions and taboos of the flamen of Jupiter, which one finds in Aulus Gellius:²⁹ Georg Rohde proved that, contrary to what one might believe, the ritual obligations (caerimoniae) imposed on the flamen were not extracted from a priestly book, but from an antiquary, most likely Masurius Sabinus, quoting himself the Augustan jurist Ateius Capito; the latter had collected a set of customs and prescriptions from his personal observations, or found in specific decrees, such as those that Augustus took in 11 BC about this priest. In any case, we have no evidence of the existence of a priestly document containing all the prescriptions that had to be observed by the flamen of Jupiter and his wife.

27A second example is the famous document on the Argei (objects that resembled a human form that were first exposed in a certain number of stations in the centre of Rome, then eventually thrown in the Tiber) which seems to give proof of the existence of great liturgical texts.³⁰ This document cannot be older than the third century BC, because of the buildings it mentions.³¹ As the rituals in question mobilised the Vestals, the flaminica Dialis, and “the most eminent priests” (ἱερέων οἱ διαφανέστατοι), which were all members of the pontifical college, modern scholars supposed that the document of the Argei came from the archives of the pontiffs. Though not in itself objectionable, this deduction gives no clue as to the precise nature of this document. Noting that Varro had not given the name of the document, but speaks either of “the sacrifices of the Argei” (Argeorum sacrificia, or sacra Argeorum), Rohde concluded that the text had no title; to his mind, all that could be said was that the pontifical text dealt with the sacra Argeorum, the “rituals of the Argei,” and even defined them as sacrificial. Therefore, he considered that the list of Argei came from a commentarius sacrorum Argeorum, a “record of the rituals of the Argei,” held by the pontiffs. If it is likely that the document goes back to the pontifical archives, there is no evidence that it was a particular commentarius about the Argei, which would ultimately be a recent form of the old “Ritualtext” sought by Rohde. It is much simpler to consider it as an excerpt from the annual commentarii of the pontiffs. This hypothesis can be supported –once again– thanks to the parallel provided by the comments of the quindecemviri, more precisely by the commentarius of the Secular Games of 204, which describes in the following terms the drawing by lots of the places where the priests were supposed to distribute the suffimenta, purification substances, to the citizens: “on 25 May, on the Palatine, the college (of the quindecemviri) convened in the temple of Apollo, in order to draw by lot on which places they should distribute from platforms to the people purifying substances: (list). After having inspected the lots and put them in an urn, they draw lots: on the Palatine, on the platform of our Augusti, which is in the square of the temple of Apollo (names of the priests drawn by lot)”, etc.³² Imagine that a scholar wanting to study the topography of Rome would have only this document to do so: he would quote this passage from the Commentarius of the Secular Games by deleting (or mentioning in the general presentation) all that concerns the drawing of lots and the names of the priests. In short, his text would specify that “purifying substances were distributed on the Palatine on the platform that is in the square in front of the temple of Apollo” and so on (suffimenta distributa sunt ‘in Palatio in tribunali, quod est in area aedis Apollinis, in tribunali, quod est ad Romam quadratam’). It should be noted that the excerpt does not affect the text of the commentarii, and that the use of the present tense, as in the Argei document, would well fit in the prescription of the quindecemviri. Our antiquary would, of course, have also been able to exploit the same lines II, 22 sq. of our commentarium, to deal with the distribution of suffimenta, “purification substances.”

28I therefore consider that for a given reason, perhaps marginal, the pontiffs were asked to define –or redefine– certain elements of the Sacra Argeorum, and recalled briefly all the stations of the procession on this occasion. Georg Wissowa and Georg Rohde imagined that Varro’s document concerned the very birth of the ritual. It is possible, but not necessary. The decrees of the pontiffs were often formulated only when new problems or circumstances had to be taken into account; in the absence of other evidence, the “late” date of the Varronian document does not prove that the custom itself is not anterior.

29In any case, the parallel provided by the Commentarium of the Secular Games of 204 seems to support the hypothesis that this list comes from priestly records rather than from a general commentarius sacrorum Argeorum –a title invented by modern scholars. Moreover, despite its title, the document of the Secular Games of AD 204 does not offer any proof that the priestly colleges possessed commentarii classified by cults. The Commentarius of the Secular Games is, in fact, a compilation of excerpts from the annual records of the quindecemviri, which were to include many reports and decrees other than those concerning these Games. As this was a very special holiday, the Senate prescribed the quindecemviri to compose this “commentarius” and to preserve it from destruction by transcribing it on bronze and marble. But this does not prove that the quindecemviri had isolated in their own commentarii the decrees and accounts related to the games. The testimony of the commentarii of the arvals even impel us to conclude in this sense: because the arvals apparently did not separate in their comments the commentarius of the sacrifice to Dea Dia, which represented their main obligation, from other decisions and descriptions, or legal documents received by the college. Finally, it is obvious that the Commentarium of the Secular Games is in any case far distant from the concept of priestly book as imagined by Ambrosch and his followers.

30Accordingly, the so-called commentarius sacrorum Argeorum is a decree of the pontiffs (or decemvirs if it was related to a Sibylline oracle) concerning the foundation or a modification of the tradition, which Varro or his source could find in the acts of the Senate as well as in the books of priests, if not in a scholarly book De sacris (or sacrificiis) Argeorum, that would have already exploited the above-mentioned sources.

31Therefore, the body of writings, which is sometimes called “archives of the priests,” consisted essentially of commentaries, that is to say, annual reports recording all the decisions taken by the college concerned in the context of their regular duties, or in response to a public or private interrogation, as well as reports on the rituals celebrated during the past year. Moreover, in the hands of the priests –but also in the hands of the magistrates– there were prayer books or some books containing oracles. We are therefore far from the pontifical or augural “Bibles” imagined by the historians of the nineteenth century. It emerges, however, from our limited survey of public priestly colleges that religious practice produced a very large number of “administrative” documents, preserved by the priests and, in some case also by the magistrates or the Senate. Very few are preserved, either more or less directly by copies on stone, or indirectly by the extracts made by Roman scholars. The disappearance of these “archives,” and, even of most of the essays that the Roman antiquaries had consecrated to them, represent an incommensurable loss. To illustrate the importance of this vanished treasure, not only for the knowledge of the public religious life, but for the whole Roman history, it is enough to convoke again the example of the commentarii of the arvales. Judging by the annual volume of these documents (about 4 000 characters), the magnitude of the commentarii of the four major priestly colleges can lead to us to dream or despair. Moreover, without the discovery, between 1865 and 1869, of nine well-preserved records of the arvals, the Mommsenian theory on the imperial power would have been different, and without the commentarii of the arvales as a whole, the knowledge of the political life under the Empire would often be thin, superficial and strongly marked by the topoi of ancient historiography.

32The loss is no less irreparable for the knowledge of the Roman religion. In fact, in this religion without a founding book, this vast group of commentarii and descriptions of rituals represented, with the priestly jurisdiction, the religious tradition: religion was oral tradition plus the annual commentarii of priests. Or rather, the religious tradition existed and was transmitted only through the annual archiving of practice and jurisdiction.

*

33In conclusion, types of documents from Greek and Roman culture relating to rituals cannot be superimposed, as can be seen from the comparison of the evidence discussed here above, be they the Greek ritual prescriptions or the Roman commentarii. Moreover, the difference is profound between the political background of Rome, notably the importance of the Urbs, the “city” par excellence, and the political background of Greece, where a thousand of independent city-states in the Aegean Basin regulated their own religious life during centuries. Beyond these “structural” differences, however, interesting similarities can be identified. On both sides, there were no “normative” ritual books for a cult, a temple or –even less conceivable– for the whole system. There were only archives, unfortunately almost entirely lost to us. Scattered pieces of evidence of this vast complex have been preserved, because it was recorded on stelae publicly displayed, which were uncovered by excavations, for the greatest benefit of modern scholars. Such records mainly attest the great flexibility of rituals in various places and circumstances, like a language able to express in various forms a message to the gods. A large range of possibilities, culturally determined by ritual performances that were particularly “Roman” or specifically “Greek,” were also attested within both cultures, from place to place, from time to time. Such flexibility is definitely one of the main characteristics of ancient polytheism.