New papyrological findings invite a question that scholars have long been asking: what is the background knowledge by the aid of which the documents in the Demosthenic corpus have been constructed? In other words, one must wonder how fabricated these documents are. It is sufficient for the purpose of this paper to mention that the majority of modem scholars regard the transmitted documents as utterly spurious, «bearing absolutely no relation to the documents that Demosthenes had the clerk read out to the court»¹. But none of these studies entirely covers the case for the documents in the speech De Corona. In a few studies conducted near the end of the nineteenth century, scholars defended – at least in part – the value of the documents inserted in this speech². The letter from Philip to the Peloponnesians (§ 157) showcases an example of a document, well represented in the manuscripts, that is «not so obviously a forgery as most of those [documents], which have preceded»³. The points of suspicion will be noticed as they occur in the passage of the letter preserved in the papyri and the byzantine manuscripts, as will also any point which merits attention as potentially genuine to the passage in question.

Similarly, it would not be necessary to conclude that the original documents were fabricated, because they were probably never included in the final copy of the speech⁴. In their great majority, the speeches of Demosthenes preserved in the papyri display no attempt to represent the documents cited⁵. But this is exactly what we might have expected, as ancient editors were much more at ease with the convenient distinction between text and subtitles, which might have referred the readers to the passages of the documents or commentaries for that matter⁶. Going back to the first question, one must also wonder: what degree of credit is due to the compiler who constructed them? For it does not follow that, because the compilers made the most astonishing mistakes, they were wholly without trustworthy materials and the intention to use them honestly. New Demosthenes papyri coming to light re-establish the view that «we must judge each document on its merits, and not condemn them all out of hand»⁷.

Two hitherto unpublished papyri from the Oxyrhynchus collection dating from the Roman period will serve as an example to set their irregularities quoted in the letter against any genuine elements preserved in them⁸. They complement each other in the sense of providing the lower and the upper parts of the letter transmitted in the medieval manuscripts and present readings which might go back to the earlier stages of the transmission of the Demosthenic documents, or even the genuine documents of the Demosthenic age. To avoid confusion, I shall divide the paper into two parts. In the first part, I shall evaluate each new reading separately and make the choice that, on my view, fits the contents of the transmitted letter best. In the second part, I shall attempt to argue that these variant readings might be indicative of sources now lost to us that Hellenistic scholars drew on to furnish their documentary passages.

I. Papyrus fragment A (P.Oxy. ined. Inv. C 229 22-7) preserves a fairly complete copy of the letter from Philip II to the Peloponnesians at § 157, and belongs to a copy from the speech to judge from the instruction to the court clerk to read out the «letter» (§ 156). The papyrus is of interest both as a testimony for the study of the forged documents in Roman Egypt and for its three new readings, one trivial but correct (16-17: μ[εθ] | [οπ]λ̣ω[ν] ε̣ρ̣χομ̣[ενοι instead of ἐρχόμενοι μεθ’ ὅπλων), another strikingly new and possibly original (7-8: τοι[c] cυνεδ̣[ροιc] | [τ]ω̣ν αρχειων instead of τoῖc cυvέδρoιc), the other puzzling but of equal textual importance (25: επιτιχιcμον corrected calamo currente to επιcτιcμον). The hand can best be viewed in its immediate palaeographical context as a representative of the general type commonly referred to as «sloping oval» capitals and, in the broader palaeographical environment, as a specimen of the «Formal Mixed» style⁹. The best evidence for the date is provided by P.Oxy. II 223, Homer, Iliad 5 (pl. I), written on the other side of a petition, II 237, dated to after AD 186.

The text of both fragments has been collated with (and missing portions of the text have been supplied exempli gratia from) the edition of Dilts (2002).

I. Papyrus fragment A

[ε]πιcτολην    (§ 156)

επιετọλ̣[η]   § 157

[β]α̣ιcλευc̣ Μακεδ[ονων]

[Φι]λιππ[οc] Πελοπ̣[ον]

5 [ν]ηcων̣ [τ]ων εν̣ [τηι]

[cυμμαχ[ιαι] τοιc δη̣[μιουρ]

[γo]ιc και τοι[c] cυνεδ̣[ροιc]

[τ]ω̣ν αρχειων κ̣α̣ι̣ [τοιc]

[α]λ̣λοιc cυμμαχοιc̣ [πα]

10 [cι] χαιρειν επ’ε̣ιδη [οι]

[λοκρ]ο̣ι̣ ọι̣ κ̣α̣λ̣ουμενοι Οζ̣[ο]

[λαι κα]τ̣οικουντεc εν̣ [Αμ]

[φιccη]ι̣ π̣λημμελου[cιν]

[ειc το] ι̣ερον του Απολλ̣[ω]

15 [νoc] τ̣ου εμ Δελφọι̣c̣ κ[αι]

[τη]ν̣ ιε̣ραν χω[ρ]αν μ[εθ]

[οπ]λ̣ω[ν] ε̣ρ̣χομ̣[ενοι λεη]

[λ]α̣τουcι· βọυλọμ̣αι τ[ωι]

[θε]ω̣ι μεθ υμων̣ βοη[θειν]

20 [κ]α̣ι̣ αμυναcθαι τ̣ουc [πα]

[ρα]βαινονταc τι τ̣ων̣ ε̣[ν]

[α]ν̣θρωποιc ευcεβων ω̣[c]

[τ]ε c̣υναντατε μ̣ετα τ[ων]

[οπ]λων ειc την Φωκ̣[ι]

25 [δα] ε̣χοντεc επι[|τ|]cιχ`τ´ιcμ̣[ον]

The wording in the prescript, namely Πελοππονηcίων τῶν ἔν τῇ cυμμαχίᾳ τοῖc δημιουργοῖc καὶ τοῖc cυνέδρoιc, has been considered a reason for establishing the letter as spurious¹⁰. I should like to set out what I have been able to collect as evidence for a hypothetical social function of the cύνεδροι and δημιουργοί in classical times to attempt to bring into relationship those terms with the first new reading of the papyrus, [τ]ω̣ν αρχειων. The cύνεδροι whom we find in inscriptions of the third and fourth centuries were almost always magistrates of the highest rank, as in, for example, IG XXII 686, 5 (III BC) and IG XXII 43, 44 (IV BC) (according to LSJ). Judging by all the literary passages where cύνεδροι occur, we could state that cύνεδροι were «select commissioners» or «delegates» sent or authorized by all member states «to the assembly of the second Athenian League» (LSJ s.v. cύνεδροι II 1). We meet cύνεδροι three times in Thucydides, 4, 22 (twice), for select commissioners from Athens to talk peace with Sparta and come to some agreement, and in 5, 85, for the Melian commissioners in the debate. Next, we find cύνεδροι in several passages from Isocrates, as in 8, 29, for delegates sent to the assembly of the second Athenian League.

Similarly, we shall try and reconstruct the creation and development of the term δημιουργοί, from one stage to another, however speculative and hypothetical the attempt may be. That the δημιουργοί played an independent part in the social life of the Peloponnesians is evident from the inscriptions unearthed in several areas of the Peloponnese. Thucydides 5, 47 mentions δημιουργοί in connection with the cities of Mantinea and Elis in the text of the hundred-year treaty between Athens, Argos, Mantinea and Elis; and Polybius (23, 5) in his account of the Achaean League. In an inscription, dated by Bingen to about the middle of the fourth century, we find δημιουργοί in Achaea in the Peloponnese, according to Bingen’s reading¹¹. As a designation of chief magistrate (LSJ s.v. δημιουργόc II), δημιουργόc very likely existed in other Peloponnesian constitutions also¹². However, we lack inscriptional evidence to show that δημιουργοί were found everywhere¹³.

The term τῶν ἀρχείων in the papyrus determines the meaning of the prescript. It would mean the «town-hall, residence, or office of chief magistrates» (LSJ, s.v. ἀρχεῖον I) or the «college or board of magistrates, magistracy» (LSJ, s.v. ἀρχεῖον II). cυνέδροιc might then be taken to mean βουλευταίc, in view of the new phrase (LSJ s.v. cύνεδροc II.2, with epigraphical evidence)¹⁴. Though certainly not the most natural way of expressing it, it is still conceivable that cυνέδροιc τῶν ἀρχείων as βουλευ-ταῖc τῶν ἀρχείων would mean the halls of various towns and the βουλευταί who attended those halls¹⁵. If this is the sense intended, the prescript of the letter must then be translated as «Philip, King of Macedonia, to the magistrates and councillors of the allied Pelopon-nesians [who attended] those halls and to all his other allies, greetings.» Additionally, the fact that ἀρχεῖον is placed in the genitive should be taken as implying that this was perhaps an original reading that did not simply infiltrate the text of the tradition, rather than a gloss that has intruded into the text.

The wording after the prescript is apparently identical with that of the letter from Philip in the medieval tradition. This is a fact of some interest for the textual history of the Demosthenic «documents». A striking and technical word, also unattested in the manuscripts, breaks the overall flow of the passage. The term ἐπιτειχιcμόc is used figuratively once in Demosthenes, 18, 87, to mean «plan of attack» (LSJ s.v. ἐπιτειχιcμόc = ἐπιτείχιcιc), as opposed to literally meaning «building a fort on the enemy’s frontier» or «fort or stronghold placed on the ennemy’s frontier» (LSJ s.v. ἐπιτείχιcμα 1), with several examples adduced from prose narrative¹⁶. The term ἐπιcιτιcμόc, on the other hand, is placed as an interlinear correction to denote in its regular usage the meaning of «store or stock of provisions» (LSJ s.v. ἐπιcιτιcμόc 2)¹⁷.

The two variant readings have clearly been designed for the same place in the speech. Only in this letter from Philip is it relevant to talk about planning an attack, placing a stronghold or stocking up on provisions. The variants may then be presumed to be alternatives. But which one is the most viable alternative? The word ἐπιτειχιcμόc might be defended as meaning figuratively what it means when used in Demosthenes, «having the plan to attack», or «having this place as the basis of operations», or «with a plan of attack», or «casting about a plan of attack against those who transgress in any way the sacred principles of religion». But ἡμερῶν τετταράκοντα, if it followed in the document – which would be right where the document breaks off – does not tie in well with ἐπιτειχιcμόν: «having the plan to attack for forty days» would place an intolerable strain on ἔχοντεc.

The corrected reading in the papyrus, ἐπιcιτιcμόc, seems the most appropriate noun to convey the sense intended for the passage, which is that of an army or state meeting in full force with forty days’ provisions. The original document must of course have specified a day by which the armies were to join Philip II, which is another reason why this document is not without objections. As Simcox plausibly argues, it would be absurd to require an army or state to be ready with forty days’ provisions, if the service might begin any day of the month¹⁸. There would also be much plausibility in Simcox’s theory that the Demosthenic documents appeared in the margin before they appeared in the text¹⁹; in this way we should be able to adopt the only satisfactory explanation for dates or the lack of them in the forged document: we may suppose that details of this kind disappeared in the process of transference of the documents from the margin into the main text, which is more than we can believe of a forger and the reasons for which he did not provide a date.

Papyrus fragment A now shows that such documents might exist in variant forms: the ancient and byzantine copies of the letter from Philip to the Peloponnesians differ only in a few individual but important details, as may be seen from the readings just presented. This is corroborative evidence for a long established view first expressed by Colin Roberts: «A compiler might add to whatever text of Demosthenes he chose his own selection of the documents; thus we could account for the various combinations of traditions in text and documents which are found in the mediaeval MSS., the papyri and the ancient citations.»²⁰ Furthermore, two papyri were discovered to contain documentary passages of § 221 that are found in none of the Byzantine manuscripts and yet differ from each other in the prescript²¹. But the prescript of the letter in P.Oxy. LXII 3009 is identical to the wording of the prescript in the letter preserved in the Byzantine manuscripts at § 157. The papyri provide further evidence that «the documents of this speech were fabricated by various people and circulated in different forms and contexts»²².

Like the variant readings preserved in the two papyri discussed above, the variant ἐπιτειχιcμόc might be due to a different version of the letter of Philip circulating in Hellenistic times, and there it might perhaps have made perfectly good sense. Hence, too, the wrong variant, ἐπιτειχιcμόc, were first introduced in the text of this document instead of the right variant, ἐπιcιτιcμόc. This accounts for the need to pertinently evaluate the variants in many cases (as in the case of τoῖc δὲ cυμβούλοιc, unanimously attested in the Byzantine manuscripts vis-à-vis the new reading in Papyrus fragment B), and for the historical difficulties met in the document. Similarly, cυνέδρoιc τῶν ἀρχείων must be nearer the original reading than the version of the prescript preserved in the Byzantine manuscripts; the phrase restores sense in an obscure prescript without being in a glossator’s language.

Papyrus fragment B (P.Oxy. ined. Inv. 34 4B.77/D (2-3) a, 34 4B.77/D (4-6) c, d) covers a wide area of the text of Demosthenes De Corona (§ 110-194), and inserts one fairly complete passage of a law at § 120 and a copy from the letter of Philip II to the Peloponnesians at § 157. The single hand at work is a small elegant specimen of the «Formal Mixed» style of the «sloping oval» type²³. It is closely similar to P.Oxy. VII 1016 Plato, Phaedrus 227a – 230e, on the front of which there is a list of landowners of the «past thirteenth year» of an unnamed emperor (P.Oxy. VII 1044). If we judge that the hand belongs to the late second century on palaeographical grounds, then we might assume as well that this thirteenth year belongs to the reign of Marcus Aurelius (173/174) or that of Septimius Severus (205/206); but such judgements are almost always bound to be fallible²⁴.

ΙΙ. Papyrus Fragment B

[ενεcτωτο]c μην̣[οc λωου] (§ 157)

[ωc η]μ̣ε̣ι̣c α̣γ[ομεν ωc δε]

[Αθηνα]ι̣̣οι βοη[δρομιω]

[voc ωc] δε̣ Κορ[ινθιοι Πα]

5 [νημο]υ̣ τοιc̣ δ̣ε̣ [μη cυ]

[ναντη]c̣αcι π[ανδημει]

[χρηcο]μ̣εθα̣ τ̣[οιc εν τηι]

[cυνθ]η̣κηι κ[ειμενοιc]

[επιζη]μ̣ιοι[c ευτυχει]

10 [τε ο]ρ̣α̣θ̣’ ο̣τ̣[ι φευγει]

[ταc ιδια]c̣ π̣[ροφαcειc]

In so far as one can judge from so short a stretch of surviving text, the letter of Philip has to a large extent the same wording as that of the Peloponnesian letter in the Byzantine manuscripts. But it also seems to offer a possible true reading, which has been anticipated by modern conjecture, partly preserved in the text and partly recovered from the lacuna, in a phrase suspected by nearly all editors of being corrupt.

The words attested by the medieval manuscripts, τoῖc δὲ μὴ cυναντήcαcι τoῖc δὲ cυμβούλοιc ἡμῖν κειμένοιc, have normally been considered to be corrupt by scholars as being unnecessary for the meaning intended in the passage, and grammatically defective due to the presence of a second connective particle within the same clause. Most editors have followed Schaefer, who proposes to leave out all that intervenes between χρηcόμεθα and ἐπιζημίοιc, i.e. τoῖc δὲ cυμβούλοιc ἡμῖν κειμένοιc. Those editors (Butcher et al.) and LSJ (χρηcόμεθα ἐπιζημίοιc = ἐπιζημιώcομεν, I 2) treat ἐπιζημίοιc as an adjective qualifying the substantively used participle τoῖc μὴ cυναντήcαcι, that is the object of the governing χρηcόμεθα. This would leave a kind of sense to the passage: «those who refuse to attend the meeting with all their available forces we shall treat as liable to the penalties of war».

Several other scholars, who have variously attempted to restore sense by emending a presumed corrupted reading, rightly reject deletion of the passage. Westermann’s universally accepted conjecture τοῖc ἐν τοῖc cυμβόλοιc ἡμῖν κειμένοιc (according to Dilts) is a distinct improvement to the sense and could even be regarded as the true reading that lies behind the meaningless cυμβούλοιc. All other editors who have emended the passage (Lipsius et al.) and some later ones generally treat ἐπιζημίοιc as a substantive and τοῖc μὴ cυναντήcαcι as a relative dative participle. The sense would then be: «as for those who refuse to attend the meeting with all their available forces we shall use the penalties provided [in our agreement?]».

In papyrus fragment B, we are most likely to have unearthed a possible new reading which Simcox believes must have fallen out of the original document. The corrupted reading, τοῖc δὲ cυμβούλοιc ἡμῖν κειμένοιc, may have arisen from mistaking two clauses, which Simcox reads in the document: τοῖc δὲ cυναντήcαcι πανδημεί cυμβούλοιc χρηcόμεθα· τοῖc δὲ μὴ [κατὰ τὰc cυνθήκαc] ἡμῖν κειμέναιc, ἐπιζημίοιc²⁵. This theory would explain how the term arose in the papyrus: if rightly supplemented, [cυνθ]ή̣κηι would retain the mood of what Philip supposedly said to the δημιουργοί and cύνεδροι of the Peloponnesians or thought in the Peloponnesian correspondence, while there are enough parallels in the forged decrees of De Corona to substantiate the word (29, 164 twice, 165-166 and 181).

In all the instances stated above, compilers talk in the plural about the «articles of agreement, and hence, covenant, treaty between individuals or states» (LSJ s.v. cυνθήκη II 2). A remarkably close parallel comes from the first letter of Philip to the Athenians at 18, 39, καὶ ταῦτα οὐδὲ cυμπεριειλημμένων τῶν Φωκέων ἐν ταῖc κoιvαῖc ἡμῶν cυνθήκαιc. Here, cυνθήκαιc does not mean «treaties» in general but «terms of treaty», which might otherwise have turned into «penalties» (§ 157: ἐπιζημίοιc) and must include those having an agreement. But Philip claims that the Athenian policy to take the field against him strikes him as unreasonable, for he subdued the Phocians who were not included in the terms of treaty upon which they had agreed.

Another document which, as far as I know, has not been utilized for the elucidation of Philip’s letter to the Peloponnesians, although it provides the closest parallel to the τοῖc ἐν τοῖc cυμβόλοιc ἡμῖν κειμένοιc clause (and now to the τ̣[οιc εν τηι] | [cυνθ]η̣κηι κ[ειμενοιc]|[επιζη]μ̣ιοι[c clause), is the second letter of Philip to the Athenians (18, 77). The relevant portion reads οὐ cυμπεριειλημμένοιc δὲ ἐν ταῖc τῆc φιλίαc κοινῇ κειμέναιc ἡμῖν cυνθήκαιc. In this letter, Philip states that the Athenians concealed the true reasons of their enterprise, namely their design to send vessels to help the Selymbrians who were not included in the articles of friendship mutually agreed upon between the Athenians and Philip. It seems to be taken for granted that each letter of Philip is in the regular style of all other letters, which have come down to us under the name of Philip, whereas we could of course maintain that they all might have been taken from somewhere. This argument will be dealt with in the last part of this paper.

Westermann’s alternative has similar qualities in the meaning, without the advantage of the documentary parallels of cυνθήκη: the term cυμβόλοιc has a definite point as the normal term used in documents for a «treaty between two states providing for the security of one another’s citizens and states for the settlement of commercial and other disputes (usually in the law courts of the defendant’s city)». It also becomes clear from the examples cited in LSJ (s.v. cύμβολον II 3) that the term was quite general in scope, not merely commercial as was once believed by scholars²⁶. But cύμβολα is not the term compilers use for treaties, nor do they anywhere else say that procedures were regulated in cύμβολα in cases that involved citizens of both states.

This discussion leaves open the question whether the restored cυνθήκῃ is the original reading in the document. If cυνθήκῃ is the original reading, it is conceivable that it would itself have been glossed by a reference to cύμβολα though utrum in alterum may not be in its favour: cυνθήκῃ gives the same sense more explicitly, which means that it may have itself originated as a scholiast’s gloss intended to clarify the meaning of an original cύμβολα, but not the other way around. But I would have thought that if cυνθήκῃ were the only reading attested in the manuscript tradition, no one would have objected to it and everyone would have defended it as perfectly fitting for the sense, pointing irrelevantly to all the other examples of the word cυνθήκῃ in the forged documents. But who could have guessed the true reading in a forged document? How many other variant readings may have been current in ancient copies of the letter from Philip to other member states? We have no ground for supposing that Philip and the Peloponnesians had adopted any federal constitution. Yet this historical difficulty may be accounted for by supposing inaccuracies in the original draft of the document, on account of the documents referring by nature to historical matters, which their compilers did not fully understand.

II. In the light of the striking new readings, the survival of the two partly preserved copies from the same letter of Philip suggests that here we have remnants descending from the same manuscript. Such a hypothesis, undeniably economical, encounters one main objection which scholars have made in the discussion on the distribution of the forged documents in the tradition: instead of the linear transmission of the documents, one has to envisage multiple recensions of forgeries that circulated contemporaneously²⁷. Yet it remains true that both papyri must have been copied from somewhere, and it is prima facie likely that papyri of at least well-known pieces such as the diplomatic correspondence of Philip were copied locally, whether it be in particular scriptoria, by master-scribes or scholars. This is perhaps not a conclusion to be dismissed: the hypothesis should remain open.

But even assuming that such scriptoria or master-scribes existed, as is required by the hypothesis, the official letters themselves are of a kind to strengthen the presumption created by the presence of all other documents, which from their nature cannot have been the letters which Demosthenes had the clerk recite in court. However, it is not necessary, in order to make out their genuineness, to suppose that Demosthenes had the documents inserted in the final copy of the speech, but «only that they were actually taken from the public archives, on the supposition that they were the identical records referred to in the text, and not fabricated». If the original documents were initially preserved separately from the speech, it would not have been difficult to imagine how they would have come to be incomplete. Thus, «it is very conceivable to imagine that the records themselves, from a certain point, may have been lost or mislaid, or from some other cause rendered defective»²⁸.

It may be fairly doubted whether it is worthwhile to put forward any theory concerning the Attic archives, when we clearly lack any evidence for them. The documents that could have been taken from the Attic archives are fabricated in their temporal frame: names and dates are almost uniformly wrong. If we suppose with Boeckh (as cited by Simcox and Simcox) that the documents were put up separately, containing the name of the archon, the day and the month of the archonship, it is still inconceivable that anybody could have mistaken the name of the archon, or the day and the month for that matter²⁹. However, we somehow feel forced again to the argument that true readings in the documents, such as these we have encountered in the papyri, can wander only if their compilers had access to a central repository of documents with authentic material preserved.

Boeckh has proposed a tempting hypothesis to meet the difficulties in making out the correspondence of Philip in the temporal frame placed, which is generally assumed as the starting point of any attempt made to maintain that the documents are in some respects genuine³⁰. He supposes that the documents were taken from a collection of decrees etc., which followed the arrangement, as Boeckh conceived it, of the Attic archives³¹. When, now, these documents were taken from the archives for collection, the name of the archon may have been lost or overlooked, the date of the archonship of such and such a one mistaken for such and such day and month³². The compiling of collections of such documents would in a broad sense be part of the same intellectual trend, oriented towards the acquisition, exhibition and ordering of solid background knowledge for educational purposes, that produced such specialized works in Hellenistic times as the books of Callimachus and other literary works.

Encyclopedic collections of various types of documents would be meant to serve as storehouses from which contemporary scholars could glean many an erudite fact or turn of a word, phrase or passage with which to season or compose their documents. Depending on their educational level, they might make virtually no mistakes or many and incomprehensible mistakes in using the collection, especially if they tried to produce summaries instead of copying at length; one can only imagine how an inept scholar, set to make epitomes from such a collection, might have produced an objectionable document like this letter of Philip. The actual process of collecting documents into catalogues or prose hand-books is not documented for the Hellenistic period proper. However, considerable interest in collecting decrees is known to have been current from the writings of Alexandrian grammarians of the time³³.

All in all, these (τoι[c] cυvεδ̣[ρoιc] | [τ]ω̣ν αρχειων, επιcιτιcμον, [cυνθ]η̣κηι) and other readings in the transmitted documents of the speech De Corona constitute cases where we could take editors to task for the implications of deleting or omitting the spurious documents from their editions as extraneous to the Demosthenic corpus. The spurious documents are neither outright forgeries nor school exercises; they are the product of the work of different writers with differing degrees of skill, making use of authentic elements and creating their own documents. The documents transmitted in the manuscripts do merit a place in the text of Demosthenes’ speech.