Greek, Roman, and Byzantine Studies 46 (2006) 339–374 
© 2006 GRBS 

 
 
 
 

The Athenian Law of  Agreement 
Edwin Carawan 

ka‹ tåw m¢n fid¤aw ımolog¤aw dhmos¤& kur¤aw énagkãzet' 
e‰nai, tåw d¢ t∞w pÒlevw sunyÆkaw fid¤& tÚn boulÒmenon 
lÊein §ãsete (Isoc. 18.24) 

OON AFTER THE RECONCILIATION of 403 Isocrates posed 
this paradox to the jury: if you side with the plaintiff 
Callimachus, “(though) you make private agreements 

binding with public authority, you will allow anyone who 
wishes to abrogate the city’s covenants for private interest.” He 
appears to be citing a “law of agreement” newly enacted by the 
restored democracy,1 a statute making homologiai idiai enforce-
able in court. This rule was restated in subsequent legislation 
and in the course of the fourth century we find it invoked in a 
wide range of cases. But the rationale behind the law is often 
obscure: What is it that the Athenians regarded as particularly 
binding about “agreements”? The aim of this essay is to define 
that sense of obligation, common to the city’s covenants and 
private contracts, that the law of agreement enforced. 

But there is a broader dispute that complicates our inquiry.  
On one side we read that the Greeks knew nothing of consen-
sual contract: 2 as in other cultures east of Rome, an exchange 

 
1 Thus L. Gernet, “Le droit de la vente et la notion du contract en Grèce 

(d’apres M. Pringsheim),” RD 29 (1951) 560–584, at 579–580 (= Droit et 
société dans la Grèce ancienne [Paris 1955] 220–221); M. Alliot, “L’ımolog¤a 
dans la vente en Grèce,” RD 32 (1952) 462–463. But cf. H. J. Wolff, “Die 
Grundlage des griechischen Vertragsrechts,” ZRG 74 (1957) 26–72, at 61 
n.81, finding the connection “quite uncertain.” 

2 R. Gneist, Die formellen Verträge des neueren römischen Obligationenrechts in 
Vergleich mit den Geschäftsformen des griechischen Rechts (Berlin 1845), discounted 
the consensual foundation, emphasizing the formal element of written 
instrument. F. Pringsheim, The Greek Law of Sale (Weimar 1950), esp. 17–85, 
emphasized formality of witnesses, asserting that sale required full exchange 
 

S 



340 THE ATHENIAN LAW OF AGREEMENT 
 

 

of promises created no enforceable obligation. For an agree-
ment to be binding there must be some basis in a “real” asset 
that one party took from the other; the exchange is marked by 
a particular formality which might seem binding in itself, but 
the lawsuit is a claim for loss or damage to that property, not 
for breach of promise. Let us call this school of thought (despite 
their differences), “the realists.” On the other side stand those 
we might call “the consensualists,”3 for whom the very mean-
ing of the word “agreement,” homologia, suggests something like 
the concurrence of wills, the consensus in idem that is the essence 
of Roman consensual contract. There are certainly arguments 
in the speeches that point to this principle. After all, even in 
modern business, most contracts only become enforceable with 
some expenditure (in payment or “reliance”), but we readily 
regard the meeting of minds as the essence of the obligation.4 

___ 
(credit sale was unenforceable). Emphasizing the real exchange: J. Partsch, 
Griechisches Bürgschaftsrecht (Leipzig 1909), esp. 76–86, treating surety as the 
foundation of Greek contract; H. J. Wolff, “Consensual Contracts in the 
Papyri?” JJP 1 (1946) 55–79, against consensual contracts in the Ptolemaic 
period, emphasizing the real transaction as in archaic antecedents (esp. 76–
77); Wolff, ZRG 74 (1957) 26–72, dealt more directly with the classical texts. 
For sound objections against Pringsheim and Wolff, but skeptical of any 
“doctrine of contract,” cf. S. C. Todd, The Shape of Athenian Law (Oxford 
1993) 255–257 and 264–268. 

3 Gernet, Droit 218–222, suggested that formalistic features in the fourth-
century law might be a regression from earlier consensual obligations. Also 
defending the consensual base: L. Beauchet, Histoire du droit privé de la ré-
publique athénienne III (Paris 1897) 14–31; A. Biscardi, Diritto greco antico (Rome 
1982) 133–161; P. Millett, “Sale, Credit and Exchange in Athenian Law 
and Society,” in P. Cartledge et al. (eds.), Nomos: Essays in Athenian Law, 
Politics, and Society (Cambridge 1990) 167–194, emphasizing “positive reci-
procity” as a force of trust in transactions; J. Velissaropoulos-Karakostas, 
“Altgriechische pistis und Vertrauenshaftung,” Symposion 1993 (1994) 187–
189, and “Merchants, Prostitutes and the ‘New Poor’: Forms of Contract 
and Social Status,” in P. Cartledge et al. (eds.), Money, Labour and Land (Cam-
bridge 2002) 131–135; and esp. E. Cohen, “A Legal Fiction: ‘The Athenian 
Law of Sale’,” Symposion 2003 (forthcoming). 

4 As Pringsheim conceded (in response to Gernet), “L’origine des contrats 
consensuels,” in Gesammelte Abhandlungen II (Heidelberg 1961) 175–193, esp. 
186; a sort of consensual contract emerges with the transition to written 
documents. For the state of the debate see now E. Cohen’s appraisal in M. 
 



 EDWIN CARAWAN 341 
 

 

So, too, in Athenian lawsuits, where this law of agreement is 
specifically invoked what seems to be at issue is precisely the 
consensual element: even in cases where the law insists that 
claims be based on a real transaction, the plaintiff is demand-
ing something more than his actual losses, something to which 
he is entitled only because of the agreement.  

As this profile suggests, the two sides may be arguing at cross-
purposes. By “consensual contract” the realists seem to mean 
one thing and the consensualists something else. To be clear 
about what we mean in this study, let us consider the instruc-
tive example of a modern casebook.5 

A few days before Christmas in 1952, at a roadside restaurant in 
rural Virginia, one neighbor offered to another $50,000 as the 
sale price for a farm. Both men had been drinking but were not 
noticeably impaired by it. The owner of the farm agreed to the 
sale and even wrote out a simple contract on the back of a 
restaurant bill. The buyer happily proceeded to raise the money 
and research the title. But the seller then dismissed the deal as a 
joke. The court found to the contrary: although nothing of any 
value had been exchanged, and what the buyer had spent in 
expectation of the sale (his “reliance interest”) was negligible, 
nonetheless, the fact that the two parties undertook the formality 
of a written contract, evidently intending to be bound by their 
agreement, was sufficient to make the deal binding. It does not 
matter that the seller was mocking his buyer (convinced he could 
never raise the money): any reasonable person would have con-
cluded from his outward actions that he intended to make the 
sale—“the undisclosed understanding … is immaterial.”  

Two features of this case make it especially useful for defining 
the issue before us: It is a straightforward example of a consen-
sual contract, uncluttered by part-payment or prior obligation. 
And it demonstrates how forceful a thing is formality. 

The first feature illustrates a basic distinction: in a strictly con-
sensual contract the exchange of promises is binding in itself.6 

___ 
Gagarin and D. Cohen (eds.), The Cambridge Companion to Ancient Greek Law 
(Cambridge 2005) 290–302. 

5 Adapted from B. Frier and J. White, The Modern Law of Contracts (St. Paul 
2005) 1–14. 

6 That is, the essence of the obligation is that promises are inherently 
 



342 THE ATHENIAN LAW OF AGREEMENT 
 

 

Among the known cases the Athenian law of agreement has 
nothing to do with deals of that sort, where a new obligation is 
entirely constructed on the exchange of promises. That is why 
the realists roundly reject any notion of consensual contract at 
Athens. But even at Athens, in dealings where the exchange of 
some asset seals the deal, the “agreement” creates obligations 
beyond the quid pro quo, and these seem to be what the law of 
agreement specifically enforces; this is what the consensualist 
emphasizes. After all, if the plaintiff merely asks for compen-
sation, to be paid for the money or goods he lost, the lawsuit 
aims at enforcing proprietary rights, not the “agreement” itself. 
The law of agreement particularly applies to what we 
might distinguish simply as “the further obligations,” beyond 
merely repaying the debt: without the law of agreement the 
plaintiff would still have a claim for what he paid or loaned; 
with the law-bound agreement he has a right to something 
more. 

By contrast the second feature—the force of formality—is 
important because it agrees so well with the Athenian cases: 
here, too, the law looks to the outward representation, however 
much the litigants plead for a deeper understanding. The 
realists may discount the force of binding consensus because it 
is cloaked in form and fiction; but the consensualist may argue 
that it is precisely that demonstrative expression of one’s will 
that marks it as a binding commitment.  

But what sort of commitment is it that this law particularly 
enforces? When we speak of contract, we often rely on Roman 
terms and modern thinking: consensus is essentially an ex-
change of promises, creating new obligations for the transfer of 
assets in future (as in Roman stipulatio).7 But in order to under-

___ 
binding: see C. Fried, Contract as Promise. A Theory of Contractual Obligation 
(Cambridge [Mass.] 1981); cf. D. Kimel, From Promise to Contract. Toward a 
Liberal Theory of Contract (Oxford 2003). One could argue that strictly con-
sensual contracts were a viable (even essential) part of Athenian commerce, 
but that question lies beyond our scope (below, n.7).  

7 On the Roman development cf. R. Zimmerman The Law of Obligations 
(Cape Town 1990) 546–577. At Athens unsecured contracts often relied on 
reciprocity within the group (as Millett has emphasized: in Cartledge, Nomos 
167–194), e.g., the leases of demes and orgeones in IG II2 2492–2501; cf. E. 
 



 EDWIN CARAWAN 343 
 

 

stand how the Athenians construed their obligations under this 
law of agreement, we need to focus on the cases where that law 
is specifically invoked and parse those arguments on their own 
terms. We cannot simply assume that all commercial dealings 
are governed by this rule; it appears to be only those concluded 
with a particular formality. And of the known cases, where that 
formality is invoked, more than half have nothing to do with 
commercial transactions. How then, in this odd set of cases, 
does the law of agreement apply—what does it have to do with 
the matter for the court to decide? Or, in sum, what exactly is 
enforceable by this rule, homologiai kyriai?  

It is best to begin near the end of the fourth century, as the 
later texts address these questions more directly in regard to 
commercial contracts. So in section 1 we focus on Hypereides’ 
Against Athenogenes (of the 320s), where the speaker insists upon 
the meeting of minds as an issue for the jury to consider, 
though he cannot find that criterion in the law of agreement 
itself. In section 2 we turn to the loan contracts governed by 
dikai emporikai. From well-attested wording of the law for this 
special jurisdiction, it is clear that “agreement” alone did not 
make a contract that the court would enforce: there must be a 
symbolaion, an obligation to complete the transaction once 
money or goods change hands. But it is also quite clear that the 
law of agreement specifically applies to the further obligations, 
especially the duty to pay an additional penalty for any devi-
ation, however understandable. In section 3 we turn to the two 
cases that treat procedural arrangements as guaranteed by the 
law of agreement, without any real transaction behind them. 
But in this singular application (it turns out), the obligation to 
honor a new deadline (and deliver then as promised) is un-
enforceable. Finally, in section 4, we return to the testimony of 
Isocrates 18 in the context of other cases of the same type: “set-
tlement contracts” (diallagai) that dispose of some prior liability. 
I argue that the law to which Isocrates refers was a broad 

___ 
Carawan, “Oath and Contract,” in A. Sommerstein and J. Fletcher (eds.), 
Horkos (Bristol 2006) [forthcoming]. These do not appear to be governed by 
the law on homologiai idiai (perhaps by the Solonian law on synthekai, F76a 
Ruschenbusch). 



344 THE ATHENIAN LAW OF AGREEMENT 
 

 

measure affecting private obligations (symbolaia idia), delictual 
liability as well as contract. The original rationale was not that 
“agreements” are binding because promises are binding but 
that decisions about the past must be closed with a certain 
finality. And that principle, crucial to the Reconciliation, helps 
to explain the peculiar reach of this rule that has provoked so 
much scholarly dispute.  
1. A contract for sale: 

˜sa ín ßterow •t°rƒ ımologÆs˙, kÊria e‰nai (Hyp. In Ath. 13) 
Hypereides’ speech Against Athenogenes is the only text that spe-

cifically applies the rule homologia kyria to a contract for sale. It is 
cited in such a way as to suggest that it is part of a statute that 
directly applies to sales such as this, but it may simply be part 
of a more general law on transactions. That ambiguity may be 
key to the dispute. The plaintiff Epicrates treats the transaction 
as a cash sale, yet it was concluded with a detailed contract set-
ting forth further obligations. There seems to be little question 
but that the plaintiff is legally bound by the letter of that agree-
ment. Yet he poses the issue for the jury to decide as essentially 
a question of consensus: should the buyer be bound by liabil-
ities that he never recognized?  

The essential background is as follows. Epicrates was in-
fatuated with a slave boy belonging to Athenogenes; the two 
men had quarreled and then reconciled, with the understand-
ing that Athenogenes would free the boy along with his father 
and brother, for forty minae (5). Following that settlement—
indeed, as though honoring his pledge of “friendship”—
Athenogenes offered, for the same sum, to sell the slaves out-
right together with the perfume business that the father, Midas, 
operated. When Epicrates proved receptive, Athenogenes pro-
duced a written contract that he had already drawn up, read it 
out, and evidently explained certain clauses. Epicrates would 
be assuming responsibility for the debts of the perfume business 
(6): Athenogenes assured him that the stock-in-trade would 
cover any such debts; yet a certain Nicon is named as surety for 
Epicrates. The parties met at the perfumery and there the sale 
was concluded (9): the money was paid, Epicrates took posses-
sion, and the contract was given to a trustee named Lysicrates. 
Soon after the deal was done, Epicrates was hounded by credi-



 EDWIN CARAWAN 345 
 

 

tors: he discovered that the debts totaled five talents. It was 
only then that Epicrates read the contract for himself (or with 
help from his friends) and realized the meaning of the crucial 
clause (10): that he would be liable for whatever Midas owed 
(ka‹ e‡ tƒ êllƒ Ùfe¤lei ti M¤daw). He then confronted Athe-
nogenes, apparently demanding that he cover some of the 
debts; but the seller simply answered that he had no idea what 
debts Epicrates was referring to and, in any event, it was no 
concern to him as he had a document disposing of their trans-
action (12): …w oÎte tå xr°a gign≈skoi ì l°gomen, oÎte pros°xoi 
≤m[›n] tÚn noËn, grammate›Òn t' e‡h aÈt`“` ke¤[m]enon prÚw §m¢ 
per‹ toÊ[t]vn.  

Athenogenes appears to have a strong defense, that the plain 
wording of the written document trumps any deeper under-
standing (13): “straightaway [he] will tell you that the law says 
‘whatever terms one man agrees to with another are binding’.” 
What is striking to us is that the plaintiff Epicrates never argues 
to the contrary—that, surely, “to agree” (homologein) naturally 
requires that both parties clearly understand and willingly con-
sent to what the deal entails. Instead he must argue that the law 
honors only dealings that are dikaia, rightful or equitable. By 
that objection he seems to mean that the transaction should 
convey equivalent value, not that true intentions are somehow 
paramount. But let us carefully consider Epicrates’ rationale, 
first weighing the points of law and then the argument on con-
sensus.  

It was once supposed that this suit relies upon a rule anal-
ogous to the Roman exceptio doli: if a contract is made under 
false pretenses, the defendant may enter this plea to cancel the 
obligation. In our case the whole thrust of Epicrates’ complaint 
is that Athenogenes plotted to deceive him by means of the 
contract. Indeed Epicrates repeatedly describes the seller’s 
scheme as a plot against him (epiboulê) and an “ambush.”8 But, 

 
8 §ntaËya ≤ §p̀ìboulØ ka‹ tÚ pl̀ã̀sma t[Ú] m°ga (8); §n]edreÊsaw ≤mçw ta›w 

sunyÆkaiw, and from the reading of the contract the jurors will realize the 
epiboulê (12); §nedreÊsant°w me, with fragmentary reference to “plotting,” bou-
leuseôs (18); the law is not made by (or for) those who plot against others, 
epibouleuontes (21). It was once supposed that the charge is bouleusis, in the 
 



346 THE ATHENIAN LAW OF AGREEMENT 
 

 

as Meyer-Laurin showed,9 the nature of the case is quite other-
wise: Epicrates cannot point to any specific law against fraud 
that applies to this case. And he seems to have no intention of 
canceling the contract. Rather, by drawing parallels from laws of 
dubious relevance, Epicrates builds his case on a more basic 
idea, that Athenogenes has done him damage by means of the 
contract just as if he had used any other instrument.  

The most nearly relevant of his counter-examples is a statute 
on sale of slaves (15). The text is unfortunately marred by a 
lacuna, but the sense seems reasonably secure: “when the 
parties agree on the terms of their transaction” the seller is to 
disclose any illness or infirmity; if not, the buyer has the right to 
return the slave.10 The restoration [ımologoËn]tew, “when they 
agree,” is not certain but most probable from the grammar and 
the context.11  Leading up to this passage, Epicrates has ac-
cepted responsibility for what he knowingly decided: the debts 
“I learned of in the contract, I do not dispute but admit that I 
owe” (˜souw [§puyÒmhn, oÈd¢n éntil°]gv soi, éll' ım̀olog[« 
Ùfe¤lein). The law on slave-sale is cited as a further instance of 
this principle.  

If the wording of this law included the crucial term, homo-
logein, it might seem surprising that Epicrates makes so little of 
it. His situation and that addressed in the law would seem 
perfectly parallel: he has unwittingly taken on losses, just as if 
he had bought a slave with a hidden disability.12 The chief 

___ 
sense of concerted fraud, but see D. Whitehead’s sensible comment on 18: 
Hypereides (Oxford 2000) 315–317.  

9 H. Meyer-Laurin, Gesetz und Billigkeit im attischen Prozess (Weimar 1965) 
16–19.  

10 ßtero]w nÒmo[w §st‹ per‹ œn ımologoËn]tew éllÆloiw sumbãllousin, ˜tan 
tiw pvlª éndrãpodon, prol°gein, §ãn ti ¶x˙ érr≈sthma: efi d[¢ m]Æ, énagvgØ 
toÊtou §st¤n. 

11 The dative éllÆloiw (“[agreeing] with one another”) makes this 
restoration preferable to the alternative (pvloËn]tew). The participle supple-
ments the main verb sumbãllousin: on the legal sense, see esp. P. Kuss-
maul, “Zur Bedeutung von sumbÒlaion bei den attischen Rednern,” in C. 
Schäublin (ed.), Catalepton: Festschrift für Bernhard Wyss (Basel 1985) 31–44. 

12 Presumably the buyer recovers at least part of what he paid. Suda s.v. 
enagôgê oiketou seems to suggest that the buyer has the advantage: the law 
 



 EDWIN CARAWAN 347 
 

 

reason that he does not pursue the analogy is simply that it 
does not lend much support to his case. He is bound by a 
written document setting forth further obligations and guar-
antees. If there is to be anagôgê, it should be spelled out in the 
covenants: either the guarantee is not there or it is useless. The 
terms of the contract are final and trump any claim to the con-
trary, even rights that are otherwise secured by statute. That 
rule, that contract is more binding than law (below, n.30), was 
often expressly asserted in the document and was probably 
included in this one. The fragmentary passage 28–31 dealing 
with Athenogenes’ violations of other laws seems to answer this 
trump clause: Athenogenes “insists upon his individual rights ... 
while transgressing the common covenants of the city,” koinåw 
t∞w pÒlevw sunyÆkaw parabåw ta›w fid¤aiw prÚw §m¢ fisxur¤zetai 
(30–31).  

So the argument on consensus appears to be just that: argu-
mentation. Epicrates can find no statute that expressly penal-
ized or canceled contracts framed by deception. The statute 
closest to a rule against fraud is the law against telling lies in the 
marketplace which may have nothing to do with this case.13 
Athenogenes has the law on his side: it makes an agreement 
binding regardless of misunderstanding; the formal act of 
affirming the agreement is what counts.14 In this sense we can 
speak of homologia as a sort of consensus. Indeed, the Romans 

___ 
urges him to resolve the matter with the seller (§f¤hsin ı nÒmow t“ »nhsam°nƒ 
diakr¤nesyai prÚw tÚn peprakÒta); cf. Whitehead, Hypereides 315–317. 

13 Meyer-Laurin, Gesetz 18, argued that this law is an instruction to the 
officials charged with supervising the agora and largely irrelevant as the deal 
was not done in the market.  

14 It is the savvy “speechwriter” Athenogenes who will (supposedly) insist 
upon what was “promised” or “undertaken.” Thus, when he initially pro-
posed the deal (6), Athenogenes acknowledged the small sums owed to 
Pancalus and Procles and added, “‘if any of the regular suppliers has any 
other stores at the perfumery, of whatever sort, this’, he said, ‘you will 
undertake’” (sÁ énad°j˙). Then (7), Epicrates explains the consequences of 
buying the business: “If I bought ... outright, agreeing with [Athenogenes] to 
undertake the debts (ımologÆsaw aÈt“ tå xr°a énad°jasyai)—assuming they 
were negligible because I had no advance knowledge (d[iå] tÚ mØ p[ro]eid°nai)—
he intended to put his creditors onto me ... entrapping me in an agreement” 
(§n ımolog¤& lab≈n).   



348 THE ATHENIAN LAW OF AGREEMENT 
 

 

began with a similar assumption: the promises are binding be-
cause they adhere to a particular form of words.15 To be sure, 
Epicrates argues that he was deceived, that there was no 
meeting of minds, but it is clear from the way he argues against 
the statute that this meeting of minds was not what the law 
meant by homologein. Athenogenes will insist that the agreement 
is final; Epicrates can only contend that binding agreements 
must be “just” or equitable and he agreed to the deal on that 
basis. That principle is not contained in the law of agreement 
but is for the jurors to weigh by their “most just opinion.”16 

In this argument, knowledge weighs more than intentions. 
Epicrates insists (as we saw) that he accepts the debts that he 
had “learned of” (14); his complaint is over the debts of which 
he was given “no advance knowledge” (8). This emphasis, that 
a man is responsible for what he knowingly takes on, leads to 
an effective dilemma (19–20): If Athenogenes claims he did not 
know of the debts, that is not a defence (apologêma) but a con-
fession (homologêma) that Epicrates is under no obligation to pay 
them—for how could he be expected to know;17 but if Athe-
nogenes knew of the debts and concealed that crucial feature of 
the transaction, the deal is wrongful (adikon) and must be rem-
edied. Of course, Athenogenes must have known: he insisted 
on a surety because he was expecting Epicrates to find himself 
in some difficulty (20). And in the midst of their negotiations 
Athenogenes queried Epicrates’ friends about what he was up 
to: Why buy the whole business, when he could just buy the 
boy? In this gambit, Epicrates claims, Athenogenes was already 
providing cover for his scheme, pretending he had fully in-
formed the unwary buyer. 

 
15 On the origins of stipulatio see H. F. Jolowicz, Historical Introduction to the 

Study of Roman Law (Cambridge 1967) 293–295; further, below 361 nn.35–
36. 

16 Meyer-Laurin, Gesetz 16–19, in disproving the old theory of a legal 
mechanism against fraud, leads us in this direction: the arguments on 
“equity” in the law are just that, artful arguments, entechnoi pisteis, using the 
law-texts to construct a circumstantial model. 

17 Perhaps Athenogenes would argue that Epicrates, from his intimacy 
with the slave family, knew the nature of the business better than the owner 
did. 



 EDWIN CARAWAN 349 
 

 

This “argument from consensus” is the crux of Epicrates’ 
case: he appeals to the court for fairness claiming that he did 
not know or understand what he was agreeing to—that he was 
entrapped by an agreement. Such is the thrust of the whole 
scenario portraying Epicrates as love-struck, easily duped by 
the conniving Athenogenes. So he introduces the Solonian law 
on wills (17), the one statute that directly addresses state of 
mind (but not contract): a man is at liberty to bequeath his 
property as he sees fit, “unless he is impaired by age or illness 
or madness or a woman’s inducements, bound or under dire 
constraint.” Epicrates contends that his own case is parallel: he 
was ensnared by the wiles of a woman (the procuress Antig-
one), indeed, under a sort of erotic derangement.    

Of course, we can easily imagine as plausible an argument 
on the other side. Athenogenes’ naming a surety should have 
put Epicrates on notice that he was taking on risky obligations 
that could cost him his property. Just as a modern lawyer might 
argue, the outward act marks the decision: “the undisclosed 
understanding … is immaterial.” 

But Epicrates contends, to the contrary, that contracts should 
require a sort of consensus, an honest understanding of the 
consequences. And, I think, we have to suppose that such suits 
were admissible and it was a proper issue for the court to 
decide, whether a man should be bound only by what he 
knowingly “agrees” to. But let us be very clear about how the 
issue is framed: the crux of Epicrates’ case is that he thought he 
was getting fair value and that understanding should count for 
something, even though this principle is not contained in the law 
that “agreements be binding.” “Agreement” in the law may be 
no more than consenting out loud, not requiring an honest 
understanding of the commitment.  

That at least is the model that emerges from this admittedly 
singular case. Homologia involves a formal consensus: the plain-
tiff argues over the meeting of minds but never suggests that 
the key term in law, homologein, should properly convey a true 
consensus. The focus of the dispute is there at the margin, be-
tween what the law recognizes in a man’s formal agreement 
and what basic fairness would require in the jury’s best judg-
ment.  

That formalism is reinforced by prevailing usage. Typically 



350 THE ATHENIAN LAW OF AGREEMENT 
 

 

when a man “agrees,” homologei, he acknowledges some fact at 
hand, something he knows from the outward act. It may have 
consequences for the future and he acknowledges these in the 
same breath, but there is no suggestion that promises are bind-
ing in themselves. In the well-worn usage of legal proceedings, 
when a man “agrees” he affirms as one integral decision what 
he has taken and what he must repay in the balance.18   

Thus in a series of inscriptions from the same period, re-
garding trierarchs’ obligations for state property (IG II2 1623, 
1628, 1629, 1631: 334/3 to 323/2), we find the basic formula, 
taÊthn tØn naËn …molÒghsen pareilhf°nai ... ka‹ épod≈sein 
(with variations). The trierarch affirms that he has taken posses-
sion of the vessel and will repay: his promise is tied to his accep-
tance. He may acknowledge that he has received a particular 
vessel (identified by its name and shipwright) and promise to 
return that particular vessel in shipshape (1629.480, 496, 506). 
Or he acknowledges that he has received an old vessel and will 
repay with one “new” (1623.6–8, 26–32, 129–132). There may 
be considerable duties beyond simply restoring what was re-
ceived: the trierarch may specify that he received a vessel with-
out rigging (1629.302–305) or “half the trireme” (573–575) but 
nonetheless acknowledges the duty to restore it; if he has fallen 
into arrears, he affirms that he will repay two-fold (1623.109–
112). Of course these texts record a different kind of obligation 
from the freely-contracted dealings of the marketplace (the 
trierarch is obligated to the polis); but the pattern of usage is 
nonetheless indicative: in homologia as a binding legal decision, 
there is no taking without the promise to pay, and no enforce-
able promise without the receipt. 

Epicrates and the trierarchs of the 320s are bound by their 
agreements in just this way: their homologia is a formal agree-
ment and binding only if it belongs to a real transaction, sealed 
by some transfer of assets at hand. But that real exchange is not 
so simple as it might seem: there are standard arrangements 

 
18 As a rule of evidence, homologein means “confess” or acknowledge; that 

such statements are binding means essentially that having once made such a 
statement, one cannot go back on it: G. Thür, Beweisführung vor den Schwur-
gerichtshöfen Athens: Die Proklesis zur Basanos (SBWien 317 [1977]) 152–158. 



 EDWIN CARAWAN 351 
 

 

that often extend the reach of the transaction far beyond the 
quid pro quo, as the next set of cases will illustrate.  
2. Maritime loans, ca. 340–323: 

˜sa ên tiw •k∆n ßterow •t°rƒ ımologÆs˙, kÊria e‰nai ([Dem.] 
   56.2) 

 In the special jurisdiction for dikai emporikai we find the law of 
agreement invoked as specifically enforcing penalties that were 
agreed upon in the contract. The debt is essentially doubled if 
the borrower violates any terms of the agreement. This further 
obligation under the law of agreement forms the main issue in 
three cases: Dem. 56, 34, and 35.  

In these penalty clauses the obligation seems to us entirely 
devised from a meeting of minds. But the Athenians seem to 
regard it as a natural extension of the real transaction: the loan 
is secured by a pledge of ship and cargo, security that the 
lender may seize at any point if the borrower violates their 
agreement. Of course the borrower initially makes this pledge 
before he has actually acquired the cargo; thus much of the 
collateral is fictitious. So long as the security is actually avail-
able, the lender or his agent may seize it if the agreement is 
violated. But if the borrower cheats his creditor of this asset, 
then the lender’s claim may amount to (roughly) double in-
demnity, for the court to enforce. Indeed, in all three cases, 
defrauding the lender of his security (not defaulting on the 
loan) is the main complaint. 

So in the case against Dionysodorus, [Dem.] 56 (ca. 323/2), 
the plaintiff, Darius, demands a judgment twice the value of his 
loan, and he persistently links that penalty with the clause re-
quiring security of a certain value. He protests that the bor-
rowers, who never brought ship or cargo back to Athens, have 
cheated him not only of the loan but also of the security (4); 
therefore they owe the penalty. The defendants have offered to 
pay back the principal and to add the interest that accumulated 
to the end of the voyage, but Darius rejects that compromise: 
“What does the contract say? … If you do not … provide the 
security in plain view and unencumbered, or do anything else 
contrary to the contract, it orders you to pay twice the value” 



352 THE ATHENIAN LAW OF AGREEMENT 
 

 

(38).19 He then has the contract read out and proceeds to em-
phasize the chief violation: “Is there any place where you have 
provided the ship on which (security) you received the money 
from us?” (39). 

Such penalties are expressly recognized in statute (10), and 
Darius protests that the defendants, by off-loading their cargo 
of grain in Rhodes and selling it there, “have shown contempt 
for the agreement and for the penalties which they wrote into 
the contract to bind themselves, if they should commit any 
violation—contempt also for your laws which demand that 
captains and onboard agents sail to whatever port they have 
agreed upon; and, if not, that they be liable to the highest pen-
alties” (10).20 

Thus, however rooted in the real transaction, the issue before 
the court is the further obligation, beyond the quid pro quo: 
Don’t let them off with paying simply what they owe (opheilo-
mena). “Fine them with the penalties (they agreed to) in the con-
tract. It would be awful for them to have prescribed a two-fold 
penalty … but you prove more generous” than their judgment 
against themselves (44).    

There are at least two other cases where such issues arise. 
These are represented by speeches where the plaintiff argues 
against a paragraphê:  [Dem.] 34 Against Phormio and 35 Against 
Lacritus. In these cases the defendant finds a basis for challeng-
ing the lawsuit in the general statute for dikai emporikai. The law 
provides for expedited proceedings in disputes regarding con-
tracts at Athens, for trade to or from Athens, in cases where 
there is a transaction or real obligation (symbolaion) and a 

 
19 ≤ d¢ suggrafØ t¤ l°gei; ... §ån mØ épod“w tÚ dãneion ka‹ toÁw tÒkouw µ mØ 

parãsx˙w tå Ípoke¤mena §mfan∞ ka‹ én°pafa, µ êllo ti parå tØn suggrafØn 
poiÆs˙w, épot¤nein keleÊei se diplãsia tå xrÆmata. The same point, in much 
the same language, at 45. 

20 This provision is distinct though reinforced by the law prohibiting any 
resident in Athens from contracting to transport grain to any other port 
than Athens (e.g. Dem. 35.37, 50–51). In the law providing for penalty 
clauses we see that the rule applies not only to the end destination but to 
“whatever port they have agreed upon.” And the penalties are established 
by contractual commitment: these are penalties “they wrote into the con-
tract to bind themselves,” not penalties prescribed by law. 



 EDWIN CARAWAN 353 
 

 

written contract (syngraphê); if not, the defendant has recourse to 
the paragraphê.21  

In the one case ([Dem.] 34) Phormio contends that there is 
no real obligation because he has fully complied with the con-
tract and therefore the transaction is at an end.22 The law may 
indeed suggest that completion of the contract should serve as a 
bar to litigation, but only if the contract is properly terminated, 
by payment and cancellation of the written instrument. Our 
plaintiff suggests as much in a comment that is incidental to 
another point of evidence and thus all the more revealing: it is 
absurd that Phormio called no one to witness the payment he 
claims to have made to the ship’s captain, Lampis; that for-
mality is essential in payment to an agent; only if he had made 
the payment directly to the lender could he have dispensed 
with witnesses. For in that case they would have canceled the 
written contract, and “you would be quit of the obligation,” 
tØn går suggrafØn énelÒmenow épÆllajo ín toË sumbola¤ou 
(31).23 In any event, the plaintiff argues that the defendant has 
not completed but violated the contract, and the penalty must 
therefore apply. This contingency is among the terms of the 
contract: if Phormio failed to comply he would be liable to 
(essentially) double the sum of principal and interest.24 That 
 

21 Cf. [Dem.] 32.1–2, 33.2–3, 34.4–5. Kussmaul, in Schäublin, Catalepton 
33–35, here takes the unlikely view that the relative clause might actually 
recognize contracts other than those secured by transactions at Athens; but 
his main point seems secure: symbolaia and syngraphai in this context are not 
synonymous; both are essential elements of the paragraphê. That the written 
contract is required, not an alternative, is indicated by the argument at 
[Dem.] 34.31 (discussed below).  

22 otoi oÈ tÚ parãpan sumbÒlaion §jarnoËntai mØ gen°syai §n t“ §mpor¤ƒ 
t“ Ímet°rƒ, éll' oÈk°ti e‰na¤ fasi prÚw aÍtoÁw oÈd¢n sumbÒlaion: … ofl m¢n 
oÔn nÒmoi … oÈx oÏtvw l°gousin, éll' Íp¢r m¢n t«n mØ genom°nvn ˜lvw sum-
bola¤vn ÉAyÆnhsi mhd' efiw tÚ ÉAyhna¤vn §mpÒrion paragrãfesyai ded≈kasin, 
§ån d° tiw gen°syai m¢n ımologª, émfisbhtª d¢ …w pãnta pepo¤hken tå sug-
ke¤mena, épologe›syai keleÊousin eÈyudik¤an efisiÒnta, oÈ kathgore›n toË 
di≈kontow (3–4). Phormio claimed to have paid principal and interest to the 
ship captain as the contract allowed him to do if he could not complete the 
return voyage (32); and, evidently, he took refuge in the usual clause that 
canceled the debt if the vessel were lost or disabled (33). 

23 Cf. Kussmaul, in Schäublin, Catalepton 37–39. 
24 From 25–26, R. Bogaert, “Notes critiques, juridiques et économiques 

 



354 THE ATHENIAN LAW OF AGREEMENT 
 

 

double indemnity is what the plaintiff Chrysippus is now de-
manding, on much the same rationale as we saw in the case 
against Dionysodorus.  

Here again we find the penalty clause linked to the security 
(6): “I loaned to this man Phormio,” Chrysippus acknowledges, 
“twenty minae for the round-trip (émfoterÒploun) to Pontus, 
on security of either cargo” (§p‹ •t°r& ÍpoyÆk˙). The last 
phrase is elliptical (the assumptions would be familiar to the 
special court): its full meaning should be that Chrysippus had 
the right to seize and sell the cargo on either leg of the 
voyage.25 Thus he charges that “[our] contract ordered a cargo 
worth 4000 drachmas to be loaded”—that is, a cargo double 
the value of his loan. This does not mean that Phormio was 
somehow supposed to take the twenty minae and buy in 
Athens a cargo worth twice as much. The sum is calculated 
from the ultimate security on the return voyage: Phormio 
should have profitably sold his cargo at Pontus and there pur-
chased a cargo for sale at Athens; thus the security doubles in 
value.26 But it was all a scam: “for he put no assets at all in the 
ship, though the contract required it as an absolute necessity.” 
Chrysippus repeatedly emphasizes that linkage, between the 
cargo as security and completion of the contract (22): “From 
what asset could he expect to pay the money; for he set sail 
from here without loading marketable goods in the ship and 
(thus) with no security.” Again, at 33 the penalty is specifically 
tied to this particular breach of the contract—not that Phormio 
would not pay, but that his security was a fraud: “It orders you 
to load marketable goods in the ship; and if not, to pay 5000 
___ 
sur le discours contre Phormion,” in Studi in onore di Edoardo Volterra III 
(Milan 1971) 123–134, at 132–133, suggested that there was a separate pen-
alty of one and a half times principal and interest (= 39 minae) prescribed 
for the first leg of the journey. But the whole calculation is argumentative 
and may be referring to a standard option. 

25 Thus in 8 Chrysippus protests that he had sent a letter to his agent in 
Bosporus, alerting him to protect his security, presumably to seize the cargo 
or denounce it if there were any fraud. 

26 Phormio proceeded to make other loans on the same security, loans 
that would ultimately have made him liable for more than a 100 minae. 
Bogaert, in Studi 123–134, argued that the sum should be 95 minae, not 115 
as the text has it or 150 as editors have supposed.  



 EDWIN CARAWAN 355 
 

 

drachmas. But you took no notice of this clause, violating the 
contract from the outset by not loading goods of any value.”  

For the linkage between penalty and security, Dem. 35, 
Against Lacritus, is especially instructive as it gives the most de-
tailed report of the contract, along with an inserted document 
(one that is quite competent, perhaps largely authentic).27 This 
is, again, a plaintiff’s speech arguing against a paragraphê; the 
basis for the paragraphê is that there is no symbolaion binding 
upon the defendant. In this case, however, the argument is that 
defendant Lacritus is not bound by the contractual obligation 
of his brother recently deceased, as he has renounced the in-
heritance. On this legal issue, our plaintiff, Androcles, has very 
little to say;28 he insists upon the requirements of the contract 
that were not met. And chief among these is the security ar-
rangement. 

For a loan of 3000 drachmas at a steep rate (rising if the 
voyage is prolonged), the brothers of Lacritus had agreed to 
load 3000 jars of wine in the Chalcidice, to sell that cargo in 
Bosporus (or farther along the coast) and then return with a 
cargo of marketable goods to Athens. These cargoes are again 
pledged as security, and the contract made clear that the cargo 
was to be solely pledged to Androcles and his partner, free and 
without encumbrance (11–12):  

They pledge these (assets) as security, owing no money on them 
to anyone else, nor will they borrow (in future) … they will pro-
vide the security unencumbered (anepapha) under the lenders’ 
control until such time as they pay the money accrued according 
to the contract. If they do not pay in the prescribed period, the 
lenders themselves have the right to offer the cargo as security or 
sell at the going rate. And if there is any shortfall of what is owed 

 
27 E. Drerup, “Ueber die bei den Attischen Rednern eingelegten Ur-

kunden,” Jahrbücher für klassischen Philologie Suppl. 24 (1898) 315–322. 
28 Beauchet, Histoire, suggested that Lacritus was held liable as surety be-

cause his role is persistently described as énad°xesyai (8, 15–16). Partsch, 
Bürgschaftsrecht 101–104, concluded that this term is intentionally misleading: 
if Lacritus were named as surety, Androcles would have used the legal term, 
§gguçsyai. But there was no mechanism to bar the jury from holding 
Lacritus liable (even if he was not surety); therefore Androcles insists that 
Lacritus has benefited from the transaction that he “sponsored.” 



356 THE ATHENIAN LAW OF AGREEMENT 
 

 

to the lenders according to the contract, the lenders have the 
right to execute their claim (¶stv ≤ prçjiw) against [the debtors] 
and all their property, on land and at sea, wherever it may be, 
just as in (seeking) judgment for debt or default, (suing) indi-
vidually or together. 

In the speaker’s summary this obligation to provide real secur-
ity is put forward as the chief grounds for the suit. And he 
indicates, incidentally, how that obligation translates into a 
penalty of twice the principal (18): 

First it is written that they borrowed thirty minae from us on 
security of 3000 jars of wine, with the assumption that they 
would hold security to the value of another thirty minae (on the 
return voyage), so that the realized value of the wine would 
amount to one talent … So it is written in the contract … but 
these men, rather than 3000 jars, did not load even 500 … They 
had no aim or intention to load 3000 jars in the vessel in keeping 
with the agreement. 

Here we have more clearly expressed what seems to be implied 
in the earlier case ([Dem.] 34), that the double indemnity de-
rives from the value of the cargo, twice converted. That such 
were the decisive terms of the contract is confirmed in the 
argument: Androcles insists that shirking the security arrange-
ment was the beginning of the fraud; they pledged security of 
such value, free and clear, and then ignored this crucial ob-
ligation.29  

It is in regard to this particular fault that he invokes the law 
of agreement (37–38): Lacritus has argued that he and his 
brothers made other arrangements in Pontus, loaning out the 
money to others, and he was bound by law in that transaction; 
Androcles protests, “The contract does not say this but de-
mands that they ship a return cargo to Athens, not loan our 
money to whomever they choose but provide (the cargo) un-
 

29 “Concerning the quantity of wine that they were required to ship, they 
made this arrangement and then immediately, from the first clause, began 
to violate and not perform the written obligations. It stands next in the con-
tract that they are pledging these assets free and clear and owing nothing to 
anyone and that they will make no additional loan on this security. So it is 
expressly written” (21). Cf. 24: pending payment, they were to provide the 
goods unencumbered, under the lenders’ control.  



 EDWIN CARAWAN 357 
 

 

encumbered (anepapha) until we receive the money we loaned.” 
Thereupon he has the contract read again and he demands, 
“Does the contract order them to loan our money … or to 
bring back a return cargo to Athens, in plain view (phanera) and 
unencumbered (anepapha)? For the contract does not admit 
anything to be more decisive (kyriôteron) than the written terms, 
neither to adduce any law or decree or anything else in opposition to the 
contract” (38–39).30 Lacritus must show “either that he did not 
receive the money or … that he repaid it, or that maritime 
contracts must not be valid or that one must put the money to other 
uses than that for which they received it under the contract” 
(43). 

These passages are especially instructive for the way in which 
further obligations are treated as part of a real transaction: a 
cargo is pledged in exchange for the money received. And that 
pledge may appear to us as a pure promise: at the moment of 
“agreement” the borrower does not have in hand the cargo 
that will serve as security. It is merely by his promise that the 
borrower is responsible for as much again as he actually re-
ceived. But that promise is cloaked in the fiction of a real 
exchange: the deal proceeds as though the cargo is already 
loaded and bound to double in value. This may approximate 
the sort of arrangement that inspired the Roman commentator 
to describe Greek contract as a sort of legal deceit: in syngraphis 
etiam contra fidem veritatis pactio venit.31 That description was ex-
plained by Mitteis from Hellenistic contracts with accrued 
interest: upon renewing the loan, the lender affirms that he has 
now loaned a sum much greater than the actual principal. The 
fourth-century loans illustrate a similar device.32  

 
30 On this rule see Cohen, in Gagarin/Cohen, Cambridge Companion 299. 
31 Ps.-Asconius on Cic. Verr. 2.1.91. L. Mitteis, Reichsrecht und Volksrecht in 

den östlichen Provinzen des römischen Kaiserreichs (Leipzig 1891) 459–475, argued 
that ps.-Asconius refers to a contract litteris (474) but that the two parties de-
posit the agreement with a Gewahrsmann as in classical syngraphai.  

32 This sort of accumulating obligation suggests another explanation for 
the much discussed horos SEG XXXIV 167: here apparently the owner (or 
buyer) has mortgaged the whole of the property for half the value; as often 
supposed, he has received title with half yet to pay. As E. Harris pointed out 
(“When is a Sale Not a Sale?” CQ 38 [1988] 351–381, and “Apotimema,” CQ  
 



358 THE ATHENIAN LAW OF AGREEMENT 
 

 

These clauses of the maritime loans pose perhaps the closest 
approximation to consensual contracts (in the strong sense), 
and yet, even here, the Athenians themselves do not seem to 
have regarded the enforceable obligation as a matter of binding 
promises. Instead, by conventional reasoning, they treated the 
deal as binding because one thing of equivalent value is owed 
for another.  

But there are cases where the law-bound agreement is not 
secured by any transfer of money or goods. These are best con-
sidered in two sets: first (section 3) in passages where the law of 
agreement is invoked only to validate the postponement of an 
obligation; and then (section 4) in settlements of prior disputes. 
3. Agreement to postpone an obligation: 

kur¤aw e‰nai tåw prÚw éllÆlouw ımolog¤aw, ìw ín §nant¤on  
   poiÆsvntai martÊrvn (Dem. 42.12) 
kÊria e‰nai ˜ ti ín ßterow •t°rƒ ımologÆs˙ (Dem 47.77) 

It is surprising to find, a generation after written testimony 
became standard, two passages which nonetheless invoke the 
law of agreement for commitments that are nowhere written 
down. Instead, in one instance we find expressly asserted what 
Pringsheim thought implicit in the other cases: that the agree-
ment must be witnessed. So one might be tempted to assume 
that we have here a relic of what was once standard: a verbal 
commitment that is binding so long as it is witnessed. Neither 
case involves a real transaction (loan, lease, sale, or service); 
both have to do with purely procedural requirements, and the 
bond appears to be strictly consensual—there is no quid pro 
quo.  

But the agreements in question appear to be unenforceable. 
Each passage has to do simply with the postponement of a 
prior obligation: in one case (Dem. 42) it is the listing of prop-
erty for antidosis; in the other ([Dem.] 47), it is the payment of 

___ 
43 [1993] 73–95), the terminology of Greek securities does not always 
convey clear distinctions: this transaction is called a sale with option to 
repurchase, prasis epi lysei, but may be essentially the same as an ordinary 
loan upon security, hypothêkê. It may be that the mortgagee has borrowed 
thirty minae but the loan is secured with a property of twice that value to 
cover the tokos or other accumulating obligations. 



 EDWIN CARAWAN 359 
 

 

court-awarded damages. And in each case the adversary has 
(allegedly) failed to do as he promised, yet there seems to be no 
penalty or prescribed remedy for that violation. The effect of 
the law is not to enforce promises but to bar enforcement of old 
obligations that have been waived. 

In Dem. 42 the plaintiff asks the jury to assign Phaenippus to 
his place on the board of three hundred responsible for bur-
densome public expenses. The agreement in question is given 
considerable emphasis throughout the speech (1–2, 12–13, 31) 
though it is only tangential to the question of Phaenippus’ net 
worth: he had agreed to deliver the inventory of his property a 
few days after the date prescribed by law, but then failed to 
meet that deadline. It is, of course, characteristic of Athenian 
court arguments that the litigant treats the case broadly, as 
embracing secondary issues that contribute in some way to the 
gravity of the charge. In this as in other misdeeds, Phaenippus 
was attempting to conceal his wealth.  

Thus, supposedly, Phaenippus has fabricated some liens—
and put up mortgage markers—on his most valuable holding (a 
large farm), markers that were not in evidence when the plain-
tiff inspected that property (5). Phaenippus has also broken the 
seals upon storage areas and made off with grain and timber. 
But first he coaxed an agreement from our plaintiff, to post-
pone the date when he must deliver a complete listing of his 
property; he then failed to deliver the list for some weeks, until 
just a few days before the trial. It was that delaying tactic that 
allowed him to conceal his assets.  

The breach of promise is not in itself grounds for a lawsuit; it 
is simply added in the scale of justice to offset whatever advan-
tage Phaenippus gained by his scheme. Strictly speaking, the 
question before the jury is not whether to punish Phaenippus 
for failing to meet his deadline; and it is quite clear that the 
plaintiff had no specific remedy for that violation.33  

In the earlier instance, [Dem.] 47, we find much the same 

 
33 See esp. 13, where he is most vehement: “If anyone should think that 

mutual agreements must be non-binding, you would hate him like the worst 
sort of shyster.” If there were a prescribed remedy or increment to the pen-
alty (say, doubling the damages), our plaintiff would have mentioned it.   



360 THE ATHENIAN LAW OF AGREEMENT 
 

 

tactic: the injustice is greatly aggravated by a violated agree-
ment which has nothing to do with the main issue.  

The essential question before the court is whether to con-
demn two men, Evergus and Mnesibulus, for false testimony in 
an earlier case of assault. The assault had come about when the 
unnamed plaintiff, as trierarch, went to the home of Theo-
phemus to seize property in payment for the ship’s equipment 
that Theophemus had failed to return (as in IG II2 1623.109–
112). At trial Theophemus charged that our trierarch struck 
the first blow; he called Evergus and Mnesibulus as witnesses to 
testify that he had challenged the trierarch to have the matter 
decided by slave-torture, but he refused—which testimony was 
a lie (as witnesses now attest, 40). In the midst of things, before 
the suit for false testimony could come to trial, the trierarch 
had asked Theophemus to put off payment, and he agreed (49–
50, 73). Nonetheless, in violation of the agreement, the de-
fendants had gone to the trierarch’s home and seized property, 
on the very day when he paid the money to Theophemus (ap-
parently unaware of the intrusion; cf. 51–52, 65–66, 77).  

The broken agreement is not a small matter. When Theo-
phemus’ agents invaded the trierarch’s home they injured a 
family servant who later died of the injury (68–73). Yet the 
trierarch was left with no legal recourse. Again, the law of 
agreement appears to be a rule without a remedy. 

The first instance also suggests that binding agreements must 
be witnessed, but that does not necessarily reflect the language 
of a statute. It may simply represent common knowledge about 
what legal “agreement” means: if it is to have any force, there 
must be some way of proving it. In both instances the speakers 
may be referring to a law that was not meant to apply to 
procedural arrangements such as these. Homologein of course 
describes all manner of statements that accept some legal re-
sponsibility—acknowledging a public duty, confessing a crime, 
or concluding a treaty, as well as making a business contract. 
There is no juristic authority at Athens competent to determine 
that the phrasing of a law on one sort of obligation does not 
properly apply to another. Accordingly, in these two instances 
where the agreement has nothing to do with a real transaction 



 EDWIN CARAWAN 361 
 

 

and seems to have no legal consequence, to invoke it may be 
merely an artful tactic of the speechwriter.34  

But if so, it is an entirely appropriate borrowing, as the un-
derlying obligation is well within the broad reach of symbolaion. 
In the law on dikai emporikai (section 2) we saw that the rule for 
binding agreement was attached to a statute requiring that a 
suit be based on some symbolaion, in the sense of transaction or 
real obligation. That rule for dikai emporikai was a restatement of 
the more general one to which Isocrates referred. And, as the 
following section will suggest, that general statute probably em-
braced symbolaia in the widest sense, including the obligations 
that arise from old liabilities, unpaid debts, and delicts (as in 
Dem. 42 and 47)—what the Greeks sometimes called symbolaia 
akousia.35 After all, symbolaia idia include unintended liabilities 
for wrongful acts as much as the debts men willingly incur.  
4. Settlement contracts (diallagai): [Dem.] 48, Isaeus 5, Isoc. 18 

That broad sense of symbolaion, found in both delicts and 
contracts, links the next set of cases: these are passages in which 
the law of agreement is said to apply to the settlement of dis-
putes, in and out of court. These agreements pose a special 
difficulty for the realists: here we have binding promises that 
are not secured by any exchange of money or goods, not even 
fictitious cargo. In order to explain what makes the promise 
binding, Wolff invoked the force of oath as a constraint beyond 
the law.36 That part of his theory would seem to solve one 
 

34 Perhaps indicative in this regard is Dem. 42.13: “Who does not know 
that the date prescribed in the law and the one agreed upon by the ad-
versaries are equally binding? For often where the thirtieth day is prescribed 
in the laws, we come to an agreement and set another date, and the magi-
strates, in all jurisdictions, postpone lawsuits and court decisions for the 
adversaries as they have reached mutual agreement.” He does not say that 
the magistrates were expressly instructed (by statute) to arrange such post-
ponements or to hear complaints when those agreements were violated.  

35 Kussmaul, in Schäublin, Catalepton 32–38. The Aristotelian division be-
tween symbolaia akousia and hekousia probably served as the basis for Gaius’ 
division of obligations into delict and contract: A. M. Honoré, Gaius (Oxford 
1962) 97–100; followed by Zimmerman, Law of Obligations 11. 

36 Wolff, ZRG 74 (1957) 26–72, assuming that stipulatio/sponsio derived 
from promissory oath; cf. Zimmerman, Law of Obligations 71, with Carawan, 
in Sommerstein/Fletcher, Horkos. Mitteis persuasively derived sponsio from 
 



362 THE ATHENIAN LAW OF AGREEMENT 
 

 

puzzle by introducing another: it explains what is practical and 
well-attested by invoking a psychological and often silent mech-
anism. But neither is the consensual model entirely adequate. 
What makes the obligation binding is not the promise in itself 
but the old liabilities, claims to property or compensation, that 
are hereby settled. If either party violates the settlement, the 
other has a remedy for damage to an asset that is already his, 
not for breach of a promise. The settlement oath serves to 
secure that disposition of property, as a boundary that one 
must respect.37 

So in the suit that Wolff took as the paradigm for oath-bound 
promises, [Dem.] 48, the plaintiff Callistratus and his brother-
in-law Olympiodorus had made an agreement settling their 
quarrel. At the death of their kinsman Comon, Callistratus and 
Olympiodorus initially quarreled over their respective rights to 
the estate but then quit their grievances (éphllãghmen) and 
drew up a written contract, concluding with an oath to bar any 
further dispute .38 In essence their deal is an absolute division of 
the property. In the businesslike terms of the contract itself, this 
is spelled out: “not to take any advantage, one over the other, 
in regard to any asset that Comon left behind, but to search out 
all assets jointly and always to act in concert, taking counsel 
together for what is needed.”39 They deposited that instrument 
with an arbiter, Androclides. And from that moment, arguably, 

___ 
surety: “Über die Herkunft der Stipulation. Eine Hypothese,” in F. Bern-
höft et al., Aus römischem und bürgerlichem Recht (Weimar 1907) 109–142; cf. 
Jolowicz, Historical Introduction 293–294. 

37 As P. Kussmaul found (Synthekai. Beiträge zur Geschichte des attischen Obli-
gationsrechtes [Basel 1969] 34), the law(s) of agreement seem especially ad-
dressed to such pacta.  

38 “I gave judgment for him and he for me, that each of us receive half of 
what Comon left (tå ≤m¤sea •kãteron ≤m«n labe›n œn kat°lipe KÒmvn), and 
that there be no unpleasantness thereafter” (8). Later he refers to this agree-
ment in principle as “equal shares” (isomoirein) and not taking advantage 
(pleonektein); see below, nn.39–43. 

39 ka‹ metå taËta sunyÆkaw §grãcamen prÚw ≤mçw aÈtoÁw per‹ èpãntvn, ka‹ 
˜rkouw fisxuroÁw »mÒsamen éllÆloiw ∑ mØn tã te Ípãrxonta fanerå ˆnta kal«w 
ka‹ dika¤vw diairÆsesyai ka‹ mhd' ıtioËn pleonektÆsein tÚn ßteron toË •t°rou 
œn kat°lipen KÒmvn ka‹ têlla pãnta koinª zhtÆsein, ka‹ prãjein met' él-
lÆlvn bouleuÒmenoi ̃  ti ín ée‹ d°˙ (9). 



 EDWIN CARAWAN 363 
 

 

each has an equal share.40 
Olympiodorus honored the agreement for a time, to the 

point of sharing a small sum that a slave revealed under threat 
of torture; but then, without consulting his partner, he tortured 
the slave and recovered another seventy minae which he re-
fused to share. Before this could be resolved, other claimants 
came forward and the estate was awarded to them. Olympio-
dorus later reopened the suit and succeeded in winning back 
the whole of the estate. Callistratus appeared as a rival claim-
ant in the later proceedings but he now claims that he should 
recover his share under the original agreement.41  

The case is based upon a law for binding agreements: though 
the statute itself is not preserved, at 11 Callistratus calls for the 
text to be read. At 46 he protests that if Olympiodorus held the 
agreement invalid, he should have gone to the trustee and de-
manded that he cancel the document.42 And then in his epi-
logue he invokes the familiar wording of the law: “How could 
[Olympiodorus] be sane, if he thinks he has to do none of the 
things on which he agreed and made covenant willingly?” (54).  

But Callistratus is in an awkward position: he must argue 
that his role as a rival is proof that he was cooperating with Olym-
piodorus; for he made no effective claim and thus adhered to 
their agreement to act in common.43 Setting out upon that line 
 

40 Cf. 12–17, regarding the initial disposition of the property including 
the first sum of money discovered from the slave (katå tåw sunyÆkaw tåw 
keim°naw parå t“ ÉAndrokle¤d˙ tÚ m¢n ¥misu §g∆ ¶labon); 18, the first quarrel 
over the seventy minae, “that each of us have his fair share according to the 
oaths and covenants that we made to each other for equal division”; 32, 
Olympiodorus prevailed and kept all the property though he made cov-
enant to share equally, by the contract lodged with Androclides; further to 
the same effect, 38, 46, 48. 

41 For such partnership (koinônia), sharing an estate and defeating other 
claimants, cf. Isae. 11.20–21; Dem. 41 (Spudias) 5, release from all claims, 
allowing no further enmity on condition that future inheritance would be 
divided in a certain way. 

42 éjioËn énaire›syai tåw sunyÆkaw parå toË ÉAndrokle¤dou …w paraba¤-
nontow §moË ka‹ ténant¤a prãttontow •aut“ ka‹ oÈk°ti kur¤vn oÈs«n t«n 
sunyhk«n (46). 

43 E.g., pãnta koinª zhtÆsein, ka‹ prãjein met' éllÆlvn bouleuÒmenoi (9); 
koinª bouleuÒmenoi; 18, Ùmvmok∆w koinª zhtÆsein ka‹ prãjein (10); regarding 
the first round of litigation, §skopoËmen pãlin ka‹ §bouleuÒmeya koinª … 
 



364 THE ATHENIAN LAW OF AGREEMENT 
 

 

of argument, he appeals to the jurors, to be “reconcilers and 
benefactors” (2), and he returns to this theme in the epilogue. 
But this is not a process in which jurors would actually negoti-
ate various options. Like most of the contract disputes, this is a 
suit for damages (dikê blabês) and the jury will give judgment 
either for the plaintiff or for the defendant; they do not de-
liberate over conciliatory solutions. What then does the plaintiff 
mean by calling upon them to be reconcilers?  

The plaintiff’s case is that the defendant must honor the 
original agreement, “according to the covenants which [they] 
entrusted to Androclides” (12). A verdict in the plaintiff’s favor 
would validate the old agreement, enforcing the division of 
property and authorizing the arbiter to dispose of any dispute. 
In this sense the jury is called upon to “reconcile” the two par-
ties, as their verdict would settle the dispute by validating their 
old reconciliation. If they judge for Callistratus, Olympiodorus 
must submit the claim for seventy minae (and any other assets 
Callistratus might claim) to the arbiter Androclides, for him to 
dispose of as provided in the covenants.44 Thus Callistratus has 
arranged for Androclides to be present, to testify that an agree-
ment was made and entrusted to him. By the terms of the 
agreement the actual covenants cannot be divulged without the 
consent of both parties. So, as he brings his argument to a 
close, Callistratus challenges Olympiodorus to do just that: let 
him call for the document and have it read out to the court.45 It 
is this maneuver that leads to the epilogue, beseeching the 
jurors “to persuade Olympiodorus” to do the right thing (58). 
What he seems to mean is that the jury should demand, with 
their thorubos, that Olympiodorus not proceed without reading 
the contract. 

Of course if he accepts this “reconciling” Olympiodorus as 

___ 
˜ntina trÒpon toÊtvn ti pãlin komioÊmeya (29); suney°mhn prÚw s¢ koinª 
prãjein ëpanta, ̃  ti ín dokª §mo‹ ka‹ so‹ bouleuom°noiw b°ltiston e‰nai (42). 

44 Wolff, ZRG 74 (1957) 69 n.199, discounted the “Scheinvertrag,” as R. 
Maschke characterized it, Die Willenslehre im griechischen Recht (Berlin 1926) 
168. 

45 When the case came before the public arbiter, Callistratus tried a sim-
ilar maneuver, challenging Olympiodorus to go with him to Androclides 
and make copies of the covenants to deposit in the echinos (48).  



 EDWIN CARAWAN 365 
 

 

much as concedes the case. His position must be that the 
original agreement is invalidated by the second round of litiga-
tion in which he won the whole estate: in that suit Callistratus 
posed as a rival claimant46 and his opposition voided their 
agreement. He would be on firmer ground if he had (as Cal-
listratus suggests) demanded of Androclides that the contract be 
destroyed “as no longer valid.”47 If he now recognizes its valid-
ity, he has no case. For it treats the division of property as a fait 
accompli and calls upon the parties to submit any dispute to the 
arbiter, who will dispose of the assets as prescribed in the cov-
enants.48  

Here again the law of agreement especially applies to en-
forcement clauses, the terms of the contract that prescribe what 
to do in the event of some default or dispute. And, to our way 
of thinking, these enforcement clauses amount to an exchange 
of promises—what the parties will do in certain eventualities, 
not in payment of a quid pro quo. But the Athenians did not 
frame the issue in quite that way. Callistratus insists upon his 
proprietary right to a share in the estate (it is already his); the 
terms of agreement that are particularly at issue are those that 
foreclose any further dispute. What the jury must decide is 
 

46 Thus 32–38, Callistratus emphasizes how he offered no real opposition 
to Olympiodorus, as the agreement remained in effect in the hands of 
Androclides. But Olympiodorus contends that Olympiodorus broke the 
agreement by acting against him: when any of our relatives asks him “why 
he will not pay, though he has sworn to share equally and the covenants are 
still in effect, he says that I violated the agreements … and ends up claiming 
that I spoke and acted in opposition to him—that is his pretext” (38). 

47 énaire›syai tåw sunyÆkaw parå toË ÉAndrokle¤dou …w paraba¤nontow 
§moË ka‹ ténant¤a prãttontow •aut“ ka‹ oÈk°ti kur¤vn oÈs«n t«n sunyhk«n 
ka‹ t“ ÉAndrokle¤d˙ t“ ¶xonti tåw sunyÆkaw diamartÊrasyai, ˜ti aÈt“ oÈd°n 
§stin ¶ti prçgma prÚw tåw sunyÆkaw taÊtaw (46).  

48 This kind of settlement appears characteristic of diaitai epi rhêtois, as in 
Isoc. 17 and 18. This matter will require detailed treatment elsewhere but it 
is at least clear in these two instances that the role of the arbiter is to decide 
whether the parties have complied with their agreement or a stipulated 
penalty should apply. A similar role is also indicated in the later dispute 
between Callistratus and Olympiodorus, when the matter of the seventy 
minae arose: Callistratus first (20) demanded that Olympiodorus surrender 
half and then (34) issued a challenge for Olympiodorus to submit the matter 
to Androclides.  



366 THE ATHENIAN LAW OF AGREEMENT 
 

 

whether that closing remains valid or not. 
The finality of the agreement, represented in the enforce-

ment clause, is again at issue in Isaeus 5, against Leochares On 
the Estate of Dicaeogenes (ca. 390). The essential details are these 
(18–24): At the death of the elder Dicaeogenes (II), apparently 
in 412/1, his estate was divided between his sisters and his 
adoptive heir Dicaeogenes III. Twelve years later the estate 
was again subject to litigation: Dicaeogenes III had laid claim 
to the whole estate, taking advantage of the disruption follow-
ing civil conflict (7–8).49 Ten years thereafter the sons of the 
sisters renewed their claim only to be met with a diamartyria by 
Leochares; he swore that the estate was not open to litigation as 
Dicaeogenes was the adoptive heir to the whole. Leochares was 
brought to trial for false-witness and the jury voted to convict 
him; but then the parties conferred and, before the votes were 
counted to decide the penalty, they agreed that the verdict be 
set aside on condition that Dicaeogenes “vacated … and 
agreed to deliver without dispute” two-thirds of the estate: 
éf¤stato … ka‹ …molÒgei énamfisbÆthta parad≈sein (20). To 
enforce that division of property, Leochares was named as 
surety, and now, in the case for which Isaeus wrote the speech, 
Leochares is charged with renegging on that obligation. 

The plaintiff, Menexenus, faces considerable difficulty. He 
argues that the clause énamfisbÆthta parad≈sein was under-
stood to mean that Dicaeogenes would surrender two-thirds of 
the original estate, free and clear; this would require Dicaeogenes 
to recover any property he had sold or mortgaged. But if that 
was the honest understanding, there was better language to 
express it.50 Instead, “without further dispute” suggests the 

 
49 Cf. W. Wyse, The Speeches of Isaeus (Cambridge 1904) 402–415. The 

implication is that no one disputed the original division while ordinary court 
proceedings were available (oÈs«n dik«n) but when civil conflict disrupted 
the courts Dicaeogenes put in his claim and it was only later decided at trial 
(efiw d¢ tÚ dikastÆrion efiselyÒntew) after litigation resumed, 401/0. As D. 
Whitehead points out, “Athenian Laws and Lawsuits in the Late Fifth Cen-
tury B.C.,” MusHelv 58 (2002) 3–28, the evidence does not support the 
common assumption that suits were suspended under the Thirty. 

50 As we saw in the mercantile cases (above, nn.28–29). The future clause 
énamfisbÆthta parad≈sein is not necessarily the wording of the oral “agree-
 



 EDWIN CARAWAN 367 
 

 

standard closing: the disposition of property is final and Dicae-
ogenes cannot renew his past claims. From the moment of the 
agreement he vacates the property that properly belongs to his 
cousins and cannot thereafter make any claim upon it. Menex-
enus, of course, argues that Dicaeogenes must also assure ac-
cess to the property; if not, Leochares is liable. 

But what he has difficulty articulating seems to be a natural 
presumption. The agreement is an acknowledgement of the 
reality at hand, not a promise waiting on future performance: 
there is no delivery date before which the properties belong to 
Dicaeogenes and after which to the cousins. When the cousins 
are denied access to those properties they resort to suing Leo-
chares, the surety who guaranteed their rights (not to dunning 
Dicaeogenes). Indeed, the clause naming Leochares as surety is 
the essence of the binding agreement: if there is any default or 
dispute, Leochares is answerable.  

On that understanding—and without further ado—Menex-
enus and his partners had proceeded to seize a property that 
Dicaeogenes had sold to Micion, believing that Dicaeogenes 
would not “confirm title” (bebaiôsein) to Micion for the property 
he had vacated in court (of course, Dicaeogenes would have to 
settle with Micion). Dicaeogenes instead supported Micion’s 
claim, and so Menexenus and partners ended up owing Micion 
forty minae for their intrusion (23), protesting that Dicaeogenes 
has done “the opposite of what was agreed.”51 That loss weighs 
in the balance: the plaintiff has at least an argument from 
fairness, that an honest understanding should outweigh the 
letter of the document. But, just as we saw in the case against 
Athenogenes (section 1), that “argument from consensus” runs 
counter to the law; there is nothing to suggest that homologia em-
braces the deeper understanding.  

What seems to be particularly “binding” (kyria) about the 
___ 
ment”; it probably belongs to the synthekai, the written terms. In any event, 
as we saw in regard to the “agreements” of trierarchs (section 1), such com-
mitments are not stand-alone promises but integral to the acknowledgement 
of some debt or assets at hand. 

51 Here we see how little the Greek obligation corresponds to common-
law “consideration”: Dicaeogenes cannot be held strictly liable for what his 
cousins lost in “reliance” on his promise; cf. Wolff, ZRG 74 (1957) 69–70. 



368 THE ATHENIAN LAW OF AGREEMENT 
 

 

law-bound agreement is the decision to make an end. Thus, in 
the paragraphê speeches ([Dem.] 37.58–60, 38.21–22), a binding 
settlement is sometimes described as a “boundary marker” 
(horos); in an inheritance dispute (40.39), the binding decision is 
a telos or peras. That principle of finality is crucial to the para-
dox with which we began in Isocrates 18. 

The context is complicated but revealing. The paragraphê is 
based on three connected contentions: (1) Callimachus has 
brought suit in violation of the “oaths and covenants”; (2) of 
course he is lying about certain facts; (3) he has brought a 
second suit in a matter settled in arbitration. The first and last 
contentions depend on the same principle of finality that we 
have discovered in the other cases. Both the Reconciliation 
Agreement, as a contract between the two parties at Athens, 
and the arbitration agreement between Isocrates’ client and 
Callimachus involve a decisive closing: the parties cannot re-
vert to the claims that are hereby resolved.52 In the Recon-
ciliation this closing was sealed with the oath mê mnêsikakein. In 
the private agreement it took the form of a diaita epi rhêtois. The 
latter is not an arbitration in the usual sense, in which a third 
party decides a dispute on the merits. Instead, much as we saw 
in the case against Olympiodorus, the designated arbiter will 
simply invoke the fixed terms of the agreement to decide 
whether obligations are fulfilled or the stipulated penalties 
apply.  

In both areas—the city’s Reconciliation and the private diaita 
—the parties are bound by certain commitments for the future. 
But this is not because promises are binding per se, for some 
future transfer of assets; it is because a certain finality must be 
given to formalized decisions about the past. It is in this regard 
that Isocrates introduces the law of agreement.  

After concluding the narrative, he begins his argument 
(18.11–19) with the violated settlement: “Initially [Callima-
chus] abided by our agreemnent (§n°meine to›w …mologhm°noiw), 
but now … he brings suit for 10,000 drachmas. When I put 
forward a witness [to the agreement, to affirm] that the suit was 

 
52 See E. Carawan, “The Athenian Amnesty and the Scrutiny of the 

Laws,” JHS 122 (2002) 1–23, esp. 5–12. 



 EDWIN CARAWAN 369 
 

 

inadmissible because a diaita had been concluded (…w oÈk efis-
ag≈gimow ∑n ≤ d¤kh dia¤thw gegenhm°nhw), he did not prosecute 
the witness” (as he should have done if he disputed that claim). 
But later, apparently after the new archon took office, Callim-
achus brought the same suit. And now at trial (12) he intends 
“not only to lie about the charges, but even to deny the diaita”: 
he insists he would never have settled for 200 drachmas; and he 
would never have entrusted the diaita to Nicomachus whom he 
knew to be a crony of the defendant. But (the defendant argues) 
the jury must consider that it was not an arbitration of some 
matter in dispute but one they entrusted, by agreement, to an 
arbiter “on stated terms” (oÈk émfisbhtoËntew éll' §p‹ =hto›w); 
so, arguably, he had no reason not to choose Nicomachus. 

This last turn of the argument is dubious (Nicomachus is still 
in a position to favor his friend), but what it assumes about the 
nature of such diaitai is significant for our inquiry: this diaita 
appears to be essentially a way of closing the agreement against 
further dispute, by appointing an arbiter to judge compliance. 
In the usual way of litigants and speechwriters, Isocrates has 
probably misrepresented the case for his adversary. Callima-
chus must have had some basis on which to challenge the 
finality of the settlement: he may have argued that the diaita 
was not epi rhêtois; or that the defendant defaulted on his debt 
and Nicomachus failed to impose the prescribed penalty; or 
perhaps that the “stated terms” were rather different.53 But the 
essence of Isocrates’ first argument is that Callimachus should 
be bound by the agreement and is therefore barred from bring-
ing suit on the same issue. And the second contention, that 
Callimachus will “lie about the charges,” probably has as much 
to do with the settlement agreement as with the original 
offense: Callimachus is lying about the diaita. 

The validity of this enforcement clause is crucial to the case. 
 

53 As Isocrates’ client goes to such lengths to defend his version of the 
diaita, it is perhaps possible that the agreement itself gave some opening for 
the second suit: our defendant construes it as a charge of “informing and 
denouncing,” but in a claim for damages it may have been more a question 
of how much he profited from the confiscation (perhaps Callimachus settled 
on the assumption that defendant got little or nothing, but later found that 
he had gotten the informant’s share).  



370 THE ATHENIAN LAW OF AGREEMENT 
 

 

The defendant has called witnesses to confirm his narrative of 
the earlier events, and he probably called witnesses, perhaps 
Nicomachus himself, to confirm that there was a diaita epi 
rhêtois. That this defendant does not appear to call for the clerk 
to read out the actual terms of agreement—and gives them 
small scope in his argument—does not diminish their im-
portance. As we saw in the case against Olympiodorus ([Dem.] 
48), it may be that the agreement itself barred disclosure with-
out consent of both parties. Or it may be (quite likely) that our 
defendant has much to lose from close scrutiny: the letter of the 
agreement may give Callimachus too much leverage. That 
would make the paradox all the more effective: if the jury de-
cides for Callimachus, they may indeed respect the rule making 
private agreements binding, but they will undermine the city’s 
Covenants.  

Whatever case Callimachus can make on the private settle-
ment, Isocrates has a trump to play against it. Callimachus (he 
contends) is violating the Covenants in two ways: (1) he is now 
charging our defendant as the informant, who instigated the 
confiscation, where the Reconciliation Agreement specifically 
shields from prosecution “informants and denouncers”; (2) he is 
also formally in violation of the rule that “dikai and diaitai under 
democracy shall be kyriai”—legal decisions and arbitrated 
settlements shall be final. Even if Callimachus has a reasonable 
argument to overturn the settlement or hold Isocrates’ client in 
default, to take his side is to undermine the more vital principle 
of the Covenants.  

From the way Isocrates links the private settlement to the 
larger issue of the Covenants, he expects the jury to see the two 
agreements as creating the same kind of obligation: the finality 
of such agreements is precisely what this case is about. So he 
frames the law of agreement in an a fortiori argument to amplify 
that essential principle. He has the “oaths and covenants” read 
to the jury and illustrates with two high-profile examples.54 The 
 

54 The endeixis against Philon was rejected; Thrasybulus and Anytus have 
not prosecuted the apograpsantes who listed their property for confiscation. 
Both cases comply with specific rules: apagôgê and endeixis to the Council for 
official wrongs under the oligarchy were disallowed (Andoc. 1.91); returnees 
could reclaim their real property (cf. Lys. Hippotherses [fr.1 Gernet] 35–46) 
 



 EDWIN CARAWAN 371 
 

 

comparanda serve to trivialize Callimachus’ complaint: it 
would be absurd to discount momentous wrongs in order to 
abide by the Covenants, only to ignore them over petty 
litigiousness (24). Then he turns the tactic around, arguing 
from the case at hand to the more weighty concern: if you 
enforce the private agreement (in Callimachus’ favor), you will 
break the bonds of the polis.  

Both agreements serve the principle of finality in disposing of 
prior claims, not a particular purpose. That distinction is made 
clear in the next turn of the argument (25): the Athenians 
thought it imperative to abide by their Agreement when it was 
uncertain whether it would be to their advantage (efi ka‹ mØ 
sun°feren énagka›on e‰nai to›w …mologhm°noiw §mm°nein); now 
that it has proven to be their salvation, it would be absurd to 
violate it. The Agreement has preserved the polis—and it must 
be defended on that score as well—but its binding force did not 
derive from that objective.  

The rule homologiai kyriai is introduced into political discourse 
in this context.55 It certainly encompassed ordinary dealings for 
lease or loan or the like, but the original context suggests that 
the rule especially applies to settlement contracts, the small-
scale version of the city’s Covenants. To be sure, both types of 
agreement, for new business and old, were encompassed in a 
single rule. And in the “paean to Covenant” (27–28) Isocrates 
emphasizes the broad reach of such agreements: trusting in syn-
thêkai Greeks and non-Greeks alike carry on commerce abroad; 
and among themselves the Greeks “conduct transactions”56 
___ 
but were barred from prosecuting those who denounced the property for 
confiscation (Isoc. 18.20, 23).  

55 See E. Carawan, “Oral ‘Agreement’, Written Contract, and the Bonds 
of Law at Athens,” in C. Cooper (ed.), The Politics of Orality (Leiden 2006) 
321–341. On the dating of Isoc. 18 see Whitehead, MusHelv 58 (2002) 3–28, 
followed by Carawan, “Amnesty and Accountings for the Thirty,” CQ 56 
(2006) 57–76, at 72–76. 

56 tå sumbÒlaia … poioÊmeya. Nicely translated by Mirhady (D. Mirhady 
and Y. L. Too, Isocrates I [Austin 2000] 103) “conduct our transactions.” In 
“Contracts in Athens,” in D. Cairns and R. Knox (eds.), Law, Rhetoric and 
Comedy in Classical Athens (Swansea 2004) 51–63, Mirhady demands a formal 
consistency in translating symbolaion as “contract” but emphasizes the real 
and transactional character of the term. To this extent his reading is 
 



372 THE ATHENIAN LAW OF AGREEMENT 
 

 

and resolve both personal enmities and wars that involve the 
whole community.  

At this juncture (ca. 402), presumably, any business contracts 
made under democracy, for shipping goods or leasing property, 
were subject to the rule homologiai kyriai. But in its immediate 
context the rule seems especially important in resolving past 
liabilities—settling personal enmities such as the very case at 
hand between Callimachus and Isocrates’ client. That impli-
cation, that the same rule covers both business contracts and 
settling old liabilities, is also indicated in the closest contem-
porary testimony, Andocides 1.87–88.  

In his defense On the Mysteries Andocides tallies up the pro-
cedural rules that took effect under the new democracy. In the 
list he cites the law that private suits and arbitrations concluded 
under democracy must be valid: tåw m¢n d¤kaw … ka‹ dia¤taw 
kur¤aw e‰nai, ıpÒsai §n dhmokratoum°n˙ tª pÒlei §g°nonto.57 
And then he explains the aim of the law: “so that debts should 
not be canceled and lawsuits not be reopened but that trans-
actions be carried out” (˜pvw mÆte xr°vn épokopa‹ e‰en mÆte d¤kai 
énãdikoi g¤gnointo, éllå t«n fid¤vn sumbola¤vn afl prãjeiw 
e‰en). Of course symbolaia encompass non-commercial obliga-
tions, liabilities recognized through lawsuits and arbitrated 
settlements.58 In other words, the intent of the law was that all 
such legally binding decisions, whether incurred under the old 
democracy or the new, shall remain binding (whereas rulings 
under the Thirty are invalid). That reading at least agrees with 
what Isocrates says about the law of agreement, as recently rec-
ognized in 402/1; for, as we saw, he introduces that principle 
as parallel to the binding authority of the Reconciliation Agree-
ment as a settlement contract (diallagai). 

Considering the overlap in their testimony, I suggest that 

___ 
consistent with Kussmaul, in Schäublin, Catalepton 31–44, though the latter 
prefers “Obligation.” 

57 By contrast, official decisions under the oligarchic regime are held in-
valid (Dem. 24.56); and public prosecutions—graphai, phaseis, endeixeis, and 
apagogai—apply only “from Euclides.” Cf. Carawan, JHS 122 (2002), esp. 
12–19. 

58 Cf. Kussmaul, in Schäublin, Catalepton 40, treating idia symbolaia in Isoc-
rates (but not this instance in Andocides). 



 EDWIN CARAWAN 373 
 

 

Andocides and Isocrates are referring to clauses of the reform 
legislation that are closely connected, if not, indeed, one 
statute. For Andocides to explicate his text in terms of symbolaia 
idia probably indicates that those words were in the law itself. 
As corollary to the clause validating “private suits and arbitra-
tions,” there would be a general rule recognizing contractual 
obligations of both types, to this effect: ka‹ sumbÒlaia ‡dia, ̃ sa 
ên tiw ımologÆs˙, kÊria e‰nai. The original intent was probably 
two-fold: (1) it would recognize the finality of settlement agree-
ments (like the one between Callimachus and Isocrates’ client); 
but (2) it would also enforce commercial obligations, so long as 
there had been some formal acceptance of the lease, loan, or 
other transaction. Phrased in this way the rule would en-
courage the settlement of unresolved issues from “involuntary 
transactions”: if a neighbor had appropriated an exile’s prop-
erty or caused some injury to his household—just as if he had 
left a debt unpaid—rather than go to court, they were urged to 
settle and then be bound by the specific terms of their agree-
ment.  

The law of agreement enacted ca. 402 was not a law to en-
force promises per se (as in Roman stipulatio). Among the known 
cases, to be sure, we find an argument from consensus, that a 
buyer should be bound by what he knowingly undertakes (sec-
tion 1). But the argument itself suggests that such consensus 
was not the ordinary sense of homologia in the law: indeed, the 
plaintiff has to argue from what is just or equitable, against a 
strict reading of the law. In commercial loans (section 2) the 
law of agreement especially applies to what we might treat as 
promises, further obligations beyond the quid pro quo. But the 
Athenians themselves did not treat these further obligations as 
binding on the grounds that promises are inherently binding. 
Rather, they construe the arrangements for security, surety, or 
penalty as practical extensions of the real transaction: the 
debtor pledges the cargo as collateral and, if he withholds that 
security, he owes twice as much because that is what the cargo 
would have fetched. Where the law of agreement is invoked 
without any real transaction (section 3), it simply waives a prior 
obligation; the promise is unenforceable. The model for these 
rules, the law of the early restoration, seems to have recognized 
what was probably a principle of long standing which the 



374 THE ATHENIAN LAW OF AGREEMENT 
 

 

threat of conflict now made all the more compelling: after 
receiving another party’s money or property—with his consent 
or without it—a formal agreement binds the receiver to repay 
by whatever terms he specifically accepts (section 4). Because 
the law was framed on that historic principle, it was not easily 
adapted to the consensual commitments that make commerce 
more productive. Indeed, it is probably more than an accident 
of the evidence that the commercial cases in which homologiai 
kyriai are invoked all date to the latter half of the fourth cen-
tury, after the rule was restated in the statute for dikai emporikai, 
fifty years after Isocrates’ paradox. Before ca. 350 the rule was 
certainly viable in commercial suits but apparently not so 
readily invoked. Through much of the fourth century, the law’s 
viva voce “agreement” is not heard as an exchange of promises 
—creating new obligations for the future—but as a mutual de-
cision about assets at hand and claims from the past.  
 
July, 2006        Modern and Classical Languages 
         Missouri State Univ. 
         Springfield, MO  65897 
         ECarawan@missouristate.edu