Divorce in the classical period was easy.1 As marriage was based on consent, so the will of either of the consenting parties in free marriage to renounce it sufficed. As we have seen, until Marcus Aurelius, ‘consenting party’ included the pater.

From the point of view of the man, particularly in the wealthier classes, ease of divorce, allowing for serial marriage, was one strategy for securing an heir to his property. In pre-industrial societies, the probability of a man dying without a son to survive him was high. How common a motive this was for divorce among Romans we cannot easily determine. In England, in the period between about 1670 and 1857, when full divorce permitting remarriage was obtainable only by Act of Parliament (and therefore virtually confined to the wealthier classes), only about 200 such Acts were passed, nearly all instituted by men.2

Roman jurists mention sterility among typical causes of divorces by consent; however, a certain asymmetry is evident. Although male sterility was recognised as a possibility, almost all our evidence relates to divorce on the grounds that the wife had not produced children. Given the state of medical knowledge and techniques in the ancient world, it was not an unnatural assumption, except where the husband was actually impotent, that the deficiency lay with the wife. In any case, the man had a more direct interest in the production of heirs. We have already seen that by the time of Hadrian the law prescribed a procedure relating to a wife pregnant at the time of divorce; though liable on default to be obliged to recognise the child, the husband had the option of denying paternity. If he wished to claim the child as his, he could insist on the pregnancy and confinement being under surveillance, to avoid the possibility of substitution or a supposititious child. A generation later, because of a test case brought before the urban praetor, the emperors Marcus Aurelius and Lucius Verus issued a rescript specifying what should be done in the opposite situation, where a husband claimed that his divorced wife was pregnant and she denied it. If the husband insisted, the wife must be examined by three reliable midwives, chosen by the urban praetor, at the home of a respectable matron. However, the emperors issued a warning that, if the husband’s claim turned out to be false, his reputation would suffer.3

Among upper-class Romans, politics, as well as finance, influenced divorce, especially in the late Republic; moralists of the early empire were inclined also to blame the dolce vita of high society. These factors are less likely to have weighed among the humbler people. Divorce is not easy to detect, however; inscriptions recording persons married more than once do not necessarily specify whether the earlier marriages were ended by death or divorce. From the city of Rome, 23 inscriptions have been found where a dead woman is commemorated by two living ‘husbands’. Although marriage is claimed, some at least of these unions began and ended as contubernium or had one slave partner; the remainder accounts for about two-thirds of the total. Even these people were relatively well-off, able to afford an inscription; among the poor, the desirability of maintaining the household as a viable economic unit might have been a disincentive to divorce.4

‘Turia’, we are told, offered her husband divorce, so that he could remarry and try to obtain an heir, but he refused. One of the more sensational instances preserved in anecdotal literature concerns the orator Hortensius (114–50 B.C.) and Cato the Younger. Hortensius in the first instance asked Cato to dissolve his daughter’s marriage to Bibulus, so that he could marry her and have a son by her. When this was rejected, he persuaded Cato and his father-in-law to consent to Cato’s divorcing his wife Marcia, who was then married to Hortensius. As Plutarch tells the story, politics played as much part in Hortensius’ calculations as progeny, since he already had a surviving son. Cato’s reluctance may not have been very strong; when Hortensius died a few years later, he left Marcia a very rich widow, and Cato remarried her. Marcia’s opinions are not recorded; but if she was now rich in her own right, she must by then have become sui iuris, and freely consented to the remarriage. She and Cato had had surviving children by the earlier marriage.5

Augustus’ legislation, while doing little or nothing to discourage divorce, encouraged remarriage and the production of children.

Divorce was a private act and as such was subject to no limitation by law, with one exception, introduced by Augustus. A freedwoman married to her patron could not divorce him without his consent.6 This is the exception that proves the rule. The reason for the exception lies in the patron’s rights (as heir on intestacy) over his wife’s property. In other words, patron-freedwoman marriage was the one type of free marriage in which there was not complete separation of property as between husband and wife. Moreover, the Augustan ius liberorum meant that, if a patron-husband released his wife to marry another man, he risked losing his inheritance rights over her property, to the benefit of another man’s children.

Divorce in Marriage with Manus

We have little direct evidence for divorce in manus-marriage.7 Since the wife was part of the familia of the husband, then steps must be taken to exclude her from the familia. Originally, the wife could not take the initiative and obtain release from manus. This meant, in effect, that a husband could divorce his wife, but she could not divorce him.8

Plutarch comments on the severity of this. However, from what he says, it appears that in the early Republic at least (if not actually by a ‘law of Romulus’) divorce was permitted to the husband only on a few specific grounds. Because of the original absence of punctuation in classical texts, these grounds could be read as ‘poisoning of children, substitution (or theft) of keys, wine-drinking’ or ‘poisoning, substitution of children or keys, wine-drinking’. Whatever the significance of the earlier part of the list, the last two items show a concern for the safety of the household property and, perhaps, for the sexual restraint of the woman. There appears to have been a provision in the Twelve Tables about divorce, but its precise nature cannot now be recaptured. At all events, it seems that in the early Republic divorce was permitted to the husband only for a few specific causes, and he was liable to forfeit his property, paying half to his ex-wife and half to the goddess Ceres, if he divorced her for any other reason.9

The results, according to Dionysius of Halicarnassus (2.25.7), were excellent. There were no divorces at all for the first 520 years of Rome’s existence. This takes us down approximately to the date of the divorce by Spurius Carvilius Ruga; and Dionysius is perhaps wrong.10 Women were led to behave themselves with modesty and decorum and, having no other recourse, to conform themselves to the will of their husbands. Their rewards were to be mistress in the house to the same degree as the husband was master, and after his death to succeed to his estate as a daughter to a father (Dionysius omits to add—sharing the estate with any children by that or other marriages, and only if the husband had not provided otherwise by will).

Carvilius divorced his wife, not for any of the specified causes, but because she had failed to produce children. Watson (1965b) argues plausibly that Carvilius pleaded (successfully) against surrendering any property to his wife, and that this created a legal precedent. Henceforth, an innocent wife could be divorced and receive no financial compensation. The result was the development of pre-marital agreements to secure the return of the dowry, and also a legal procedure for its recovery, the actio rei uxoriae. In later Republican law, there were no fixed grounds for divorce.

Dionysius adds that marriage with confarreatio was totally indissoluble. That may already have ceased to be the case by the end of the Republic, since in Festus diffarreatio is explained as a ritual by which a man and woman may be sundered. It is unlikely to have remained indissoluble even for the flamen Dialis very long after Tiberius’ limitation of the effects of confarreatio; by the time of Domitian the flamen Dialis could obtain a divorce, by imperial dispensation. In the historical period few outside the priesthood are likely to have been married with confarreatio.11

Divorce in manus-marriage would require not merely a repudiation but the breaking ofmanus by ‘emancipation’ of the wife. The text of Gaius is defective at a crucial point, but it looks as though by this time, doubtless under the influence of free marriage, it was possible for a woman or her father to take the initiative in obtaining a divorce from a manus-marriage. A daughter, he says, cannot compel her father, but a wife, if she has sent a repudium (notice of divorce), can compel her husband (sc. to remancipate her), just as though she had never been married to him.12

Procedures for Divorce

The law prescribed no fixed procedure for divorce from a free marriage. It is sometimes alleged that the Twelve Tables, as cited in Cicero Phil. 2.69, did lay down a fixed formula of words, at least for divorce by the husband, the only type then available. What Cicero says is: Illam suam suas res sibi habere iussit ex XII Tabulis clavis ademit exegit. However, the meaning changes, depending on whether one punctuates after iussit or after Tabulis: (i) ‘He told his wife to keep her own belongings; in accordance with the Twelve Tables, he took away the keys and drove her out’; (ii) ‘In accordance with the Twelve Tables, he told his wife to keep her own belongings; he took away the keys and drove her out’. The second version would mean that the Twelve Tables specified a formula of words, tuas res tibihabeto. An (unconvincing) attempt has been made to relate these words to a legal procedure for regulating the property and recovering the wife’s dowry.13 Certainly, some such phrase seems to have been traditional. Gaius says merely that in divorces, ‘that is, in the announcing of the divorce’, this phrase is acceptable, and so also is res tuas tibi agito; he does not say that they were either essential or compulsory.14

The stories of the man with a wife in Spain and another in Rome, and the aftermath of Messalina’s wedding to Silius (see Chapter 3, notes 53 and 95), indicate that a divorce could be held to have taken place, even without a formal declaration, if there appeared to be the intent to divorce. In the latter case, Tacitus specifically says that Messalina had sent no notice of divorce; however, those closest to Claudius at court treat the matter as a political crisis and assume that Silius is aiming at supplanting the emperor. Narcissus, trying to rouse the emperor to a sense of the danger, excitedly tells him that he is divorced. The former case, Cicero tells us, aroused considerable dispute among jurists. The man himself was dead and could not be questioned as to his intention. Later, we find jurists in the empire implicitly assuming the primacy of intention, when they say that if a couple divorced and then came together again this should be counted as one and the same marriage. A particular instance is mentioned, that of Maecenas and Terentia, whose onoff marriage gave lawyers food for argument about the legal validity of gifts that passed between them.15

Paul, in a commentary on the lex Julia de adulteriis, said, ‘No divorce is valid unless witnessed by seven adult Roman citizens besides the freedman of the person (qui—masculine) divorcing.’ This shows that Augustus did make this procedure a legal requirement and is apparently confirmed by Suetonius’ remark that Augustus imposed on divorce a set form (modus). ‘Limit’, another possible meaning, would not make sense here, since Augustus’ legislation tended to encourage divorce, rather than the reverse.16 All that this shows, however, is that Augustus provided for a procedure in the case, namely, the adultery of a wife, where divorce was compulsory and it was important for the husband to be able (pimp). It does not show that a formal procedure was instituted for to establish the fact of divorce, on penalty of prosecution as a leno divorce in general.

A formal declaration by the divorcing party to the other partner in the marriage does appear to have been usual; he or she nuntium remisit, ‘sent notice of divorce’.17 Although no classical text actually says so, the declaration of divorce, rather than its reception by the other party, terminated the marriage; otherwise (especially given the nature of communications in the ancient world) the power to divorce and remarry could in some circumstances have been severely limited, as for instance in the prolonged absence of a husband, whose fate was unknown. A mistaken presumption of his death could give rise to an ‘Enoch Arden’ situation, if there were no possibility of divorce by simple declaration.18 However, a woman married to her patron would be in a particularly awkward position, since his consent was necessary to a divorce. Simple absence did not dissolve the marriage. Our earliest evidence for a woman being allowed to remarry comes from the reign of Constantine; it concerns the wife of a soldier (permissible by that time), who had had no news of him for four years. She is allowed to remarry if she has first applied without result to the appropriate military officer. Someone insane could not be said to have ‘received’ the notice, but could nevertheless be divorced.19

Capacity to Divorce

Persons sui iuris in free marriage could terminate marriage. Those in potentate had to have the co-operation of the pater;20 the pater could himself intervene to end a marriage, but, as we have seen, from the time of Marcus Aurelius not if the marriage was harmonious.

Augustus’ lex Julia et Papia forbade a freedwoman to divorce her patron against his will. Since marriage rested on consent, this ban obviously could take the form only of a lex imperfecta, that is, it could not prevent or penalise the act itself, it merely prevented its legal consequences from coming into effect. So, the woman would not be able to reclaim her dowry, nor would she have conubium with anyone else, and she would be unable to marry again.21

If a woman became insane, she could not divorce her husband, nor was her curator allowed to do so; if she were in potestate, of course, her father could obtain a divorce for her. Since insanity did not invalidate previous acts, but merely removed legal capacity for any further acts, a freedwoman whose patron became insane, or a daughter whose father did, was effectively deprived of the ability to obtain a divorce.22

Dissolution of Marriage other than by Divorce

Marriage could be terminated, other than by divorce, if either party lost conubiwn or lost civitas. The former occurred if men became senators when married to women in forbidden categories, or if they became soldiers.

Certain kinds of’capital’ sentence on convicted criminals involved loss of civitas or enslavement. Condemned persons of high rank were more likely to receive the former type of sentence, lower orders the latter.23 The commonest form of penal slavery resulted from condemnation to work in the mines, and then only if the sentence were for life; if it was merely for a term, free status was retained, and children born of women so condemned were free. Women were usually sentenced to help the mineworkers (ad ministerium metallorum) rather than actually to work in the mines.24 Citizen status was lost by a sentence aquae et ignis interdictio (effectively banishment) or deportation; this was a capital sentence, the condemned person being, so far as the law was concerned, dead. According to late interpolations in the Digest, deportation did not dissolve marriage, but in classical law it did. The dowry either remained with the husband of a condemned woman or could be claimed by her father (if she were still in potestate); but if she had been condemned of certain statutory offences, namely maiestas (treason), armed assault, parricide, poisoning and assassination, it was confiscated by the state.25

Captives in the hands of a foreign power were the slaves of the state which captured them, and were usually sold to private owners. As slaves, they had none of the rights of Roman citizens. Since captives could and did sometimes return, their rights were regarded as being suspended; should someone die in captivity, by a legal fiction he (or she) was deemed to have died at the moment of capture, so that his estate could be settled as that of a free person. If he returned, then by the principle known as postliminium former rights and status were restored, with one major exception—marriage.26

A slave could not be validly married to a Roman since there was no conubium, and capture therefore dissolved marriage. When the captive returned, however, marriage was not restored by postliminium, but it could be renewed by consent.27 Corbett (1930:213) takes this ‘to signify merely that the effective operation of postliminium was subject to the condition of the other consort’s consent’. This is not what the text says; marriage was renewed by consent. Postliminium normally operated automatically; the rights of the captive had merely been in suspension. This could not be done in the case of marriage, since during a marriage it was open to either partner at any time to terminate the marriage by giving notice of divorce. If one partner was taken captive, there could be no automatic retention of rights, since this would conflict with the liberty of the other partner to end the marriage at will.

There is an exception that proves the rule. Suppose husband and wife were taken captive together, and children were born to them in captivity. If both husband and wife subsequently returned, the children would be treated as legitimate. The reasoning here must be that, as neither partner had been in a position to exercise a wish to end the marriage, its continuance had not been dependent on the will of anyone other than the captives, and so postliminium could apply.

This ruling was given in a rescript by the emperors Severus and Caracalla. However, for the children to be regarded as legitimate, both parents must return; if only the mother returned with the child, it would be regarded as illegitimate. This is not discrimination against the mother; the texts, all dependent on the original rescript which was given to a provincial official, presumably in relation to an individual case, do not specify the other cases, of a child returning with the father only, or alone, but the same would doubtless apply. Since the law assumed that the dead parent died at the moment of capture, the child could not be regarded as born in wedlock.28

Marital Fault

English law has only recently abandoned the concept of marital offences as the grounds for divorce; the sole ground is now irretrievable breakdown of the marriage, and in the Divorce Reform Act of 1969 adultery and behaviour such ‘that the petitioner cannot reasonably be expected to live with the respondent’ are listed along-side separation or desertion for specified periods as Tacts’ to be proved in support of a claim that the marriage has broken down. Consequently, the concept of marital fault has disappeared in the Marital Proceedings and Property Act of 1970, except for an instruction to the court to arrange matters for the parties ‘so far as is practicable and, having regard to their conduct, just’.29

In Roman law, both pre-classical and post-classical, divorce was permissible only on the grounds of certain specified offences committed by the other partner. Divorce by consent was finally abolished by Justinian. Increasingly severe penalties were imposed in the Christian empire on persons seeking a divorce for any other reason. To loss of dowry and gifts before marriage were added temporary or permanent loss of conubium, deportation and, eventually, confinement for life in a monastery and forfeiture of property.30

In classical law, in contrast, there were no specified grounds for divorce, and the sanctions for frivolous or apparently unjustified divorce were social rather than legal. In 50 B.C., Caelius retails to Cicero among the gossip of the city that the sister of Valerius Triarius has divorced her husband, who was away on duty in a province, for no reason, just the day before his return. He adds that she is about to marry Decimus Brutus. Cicero himself divorced his wife Terentia four years later. Plutarch’s account of the affair lists a whole series of complaints on Cicero’s part (all denied by Terentia) ranging from serious matters, such as removal of household property and incurring of debts, which if true could have formed the basis of a series of lawsuits, to grumbles about her parsimony in supplying the travelling-expenses of his daughter Tullia when she met him after his return from exile. Plutarch comments that Cicero’s real motives soon became apparent when he married the young and wealthy Publilia. He subsequently divorced her because he thought she was pleased at the death of Tullia. All this sounds like the gleanings of gossip among their social circle.31

Augustus, as we saw, instituted penalties for a husband who did not divorce his adulterous wife. The restrictions on a freedwoman’s right to divorce had nothing to do with the grounds of divorce.

Nevertheless, though divorce without matrimonial offence was permitted, the concept of fault did still find expression, in two ways.

First, in classical law a wife guilty of adultery could be penalised by a praetorian action brought by the divorcing husband, the actio de moribus. The husband had to provide security, Gaius tells us; the reason doubtless was to discourage unfounded accusations made from dishonest motives. If successful, he could retain one-sixth of the dowry. Conversely, if the wife were suing the husband for return of dowry, he could put up a defence of retentio propter mores, i.e., that he had a right to retain the same amount on the grounds of her adultery. For less serious misconduct, he could keep only oneeighth. If the fault was on the husband’s side, then instead of being allowed to return the dowry, as usual, in three annual instalments, he was given only six months to pay or, in the case of graver misconduct, had to pay up at once.32

These procedures were already available in the Republic. In 100 B.C. a certain C.Titinius of Minturnae divorced his wife Fannia, and tried to retain the dowry, alleging immoral behaviour on her part. The judge was C.Marius; when it was discovered that Titinius had known about his wife’s character all along, even before he married her, Marius tried to persuade Titinius to give in and return the dowry. When he refused, Marius awarded him from the wife’s dowry the derisory sum of one sesterce, saying that it was perfectly clear to him that Titinius had deliberately chosen to marry an unchaste wife because he had designs on her patrimony. Under the empire, Titinius would not have got off so lightly. As husband of an adulterous wife, who had not prosecuted her within the time prescribed by the lex Julia, he would not only have lost the dowry but would himself have been liable to prosecution as her pimp. The jurist Scaevola said that in such a case the husband should not be allowed to keep any of the dowry, since he himself had caused or connived at her immorality.33 The penalties, besides divorce, imposed by the lex Julia on an adulterous wife will be discussed later.

Secondly, ‘fault’ (culpa) occurs in contexts where the meaning is not misbehaviour of any kind but rather ‘responsibility’; it merely indicates the party who initiated the divorce. Ulpian says that the husband could retain one-sixth of the dowry in consideration of the children of the marriage (retentio propter liberos), up to a maximum of three children, if the divorce had been through the culpa of the wife or her father. Clearly, this does not mean any moral transgression on the part of the wife’s father, but merely that he had sought the divorce. Cicero makes the same point but, confusingly, using culpa in the moral sense: If a divorce occurs through the fault (culpa) of the husband, then even though the wife sent notice of divorce nothing should be kept for the children.’ If the divorce had been the woman’s ‘fault’ as initiator, and the husband was innocent of misconduct, he could keep part of the dowry for the children; but if he was guilty of misconduct, then, irrespective of who initiated the divorce, he could not.34

Although there were other grounds on which a husband could base a claim to retain part of the dowry, this penalty on a wife or father who initiated a divorce applied only where there were children and was meant to make some provision for them. Otherwise, the whole dowry had to be returned. Nevertheless, Cicero himself was in a dilemma. Before he had finished paying off the instalments of Tullia’s dowry, her relations with Dolabella had deteriorated to a point where Cicero was thinking of divorce. Besides, he himself would like to get back the amount of dowry already paid. In July 47 B.C. he tells Atticus: ‘I think I should send notice of divorce (and you agree). Perhaps he will ask for the third instalment. Consider whether we should wait for him to raise the matter, or act first.’35 Cicero had been dilatory in paying the first instalment of the dowry, the second had been paid over a year before, and the third was now overdue. Tullia had had a weakly seven-month baby in May 49 B.C., but no more is heard of it, and it had probably died. There were as yet no further children of the marriage, so Cicero could get back the whole dowry in the event of divorce, but that could take three years; and unless he initiated a divorce soon, he would meantime have to get together the money for the third instalment due to Dolabella. He alludes in the letter to various delinquencies of Dolabella’s, apparently including infidelity. This could perhaps have been used to oblige Dolabella to return the dowry more promptly; but in the current political situation Cicero was not anxious to make an enemy of Dolabella.36

A year or two later, the divorce had occurred. In January 45 B.C., Cicero tells Lepta that he is trying to get Dolabella’s agents to pay the first instalment of the dowry.37 Tullia had just had a baby. A few weeks later she was dead, and the child probably did not long survive her.


Rome was a monogamous society. One woman could not be married to two men simultaneously, nor could one man have two wives.38 Given the absence of any legally compulsory formality for the constitution either of marriage or, apparently, divorce, there might on occasion be difficulty in establishing whether either had taken place. Some modern writers have supposed, largely on the doubtful basis of Cicero’s story about the man from Spain, that a second marriage ipso facto dissolved the first, so placing themselves in difficulty when they try to explain how, in that case, the Romans had any juridical conception of bigamy.39

To the Romans the legal position was quite clear; so long as an earlier marriage remained in existence, no other marriage could be contracted. The lawyers’ problem concerning the man from Spain arose purely from this: that it had become important, for the settlement of his estate, to establish whether the first marriage had legally been terminated or not, and they lacked any proof.

That a second marriage did not of itself dissolve the first is shown, for example, by texts relating to the situation in which a woman marries again, wrongly supposing her husband to be dead. Quintilian makes this the basis of one of his Declamations, in which the husband, returning, kills the wife and her second husband and is himself tried for murder. His defence is that he was avenging himself on adulterers caught in the act. ‘Marriage’, he says,

Only gold members can continue reading. Log In or Register to continue