Interpreting Demographics in Roman Egypt

This article looks at the demographics of pre-Christian Egypt to evaluate the claim that the presence of never-married adults is a Christian phenomenon. Roman legal and literary sources treat single adults as a special anomaly, such as Vestal Virgins or priests of Cybele. Augustine law encouraged marriage and even penalized potential heirs if not married. This applies only to the citizen class and specifically does not apply to those in the military, sex workers, and enslaved people.

In general, Roman Egyptian society followed the “Mediterranean marriage pattern” which involves early marriage for women and a significant age gap, with men being older, a universal expectation of marriage, and no lifecycle period of unwed labor for (free) women.

The data for this analysis comes from surviving census records from approximately 0 to 250 CE. Available records include 400 documents which record 250 households and nearly 500 individuals. Most were lower or middle class. Approximately 11% were enslaved people. Census records were completed by household, counting all those at the same residence, including names, relationships, and identifying characteristics. They do not indicate marital status directly, though it can be interpolated from other relationships, if the spouses are in the same household. Divorce was easy and common and often simply for incompatibility.

Rarely, the word “parthenos” occurs to identify an unmarried young woman, but it does not automatically mean “virgin” in the sexual sense. Pre-marital chastity was not required for pre-Christian Egyptian women, though it could be a concern for upper class Romans and became a general expectation in Christian society. Compare also to Jewish concerns for virginity. Marital status is referred to explicitly by widows to emphasize their legal vulnerability, but not typically by divorced or never-married women.

Marriage was created by cohabitation and ceremony, but not by legal formality. Widows might head their own household (with children) without remarrying. There are examples of widows protesting a daughter’s marriage when the daughter had been a business asset.

So what are we looking for in the census that would fit in the category “single.” Based on normative life stages, the author settles on defining “single” as anyone 15 or over, currently not married (defined as the absence of a spouse in the household), but regardless of whether the person has had children and regardless of the rest of the household composition. This will be a diverse group in terms of age and life situation. The evaluation excludes people who were not legally able to marry.

Out of 1116 people whose situation can be reconstructed from records almost half (554) were 15 or older. Of these, almost half (232) were single, as defined above. Unlike the “northern pattern” which included lifecycle employment for young unmarried people, there were no free domestic servants. The census identifies 253 households, of which 132 included at least one single person.

Looking at gender, of the 232 singles, 109 were women. And as women comprised roughly half of the 15-and-over population this meant women and men were equally likely to be single, and roughly half of all adults of either gender were single. Of these 109 single women, 49 had children living with them, and so may well have been married in the past. [Note: Given the reference to children typically living with their father’s extended family after divorce or parental death, this statistic seems to beg for some explanation. It would be interesting to know the ages of the children in this set. Was this primarily older widows living with their adult children? Is is the assertion that the children of disrupted families typically lived with the father’s family over-blown? See the comment about widows with children heading households.]

So 60 women-–about a quarter of all adult women-–have no evidence of ever being married (of those, some may have been previously married, but had no living children). This is a much higher proportion than for otherwise similar Mediterranean societies at various times that have been studied demographically.

Children created a significant social distinction. A widow with children could head a household while a childless widow or never-married woman would belong to a near relative’s household. (Typically a male relative, but presumably the adult daughter of a widow could belong to her mothers household and there is reference to this).

Therefore “single women” do not constitute a homogeneous group. But since children typically remain with the father’s family in the case of divorce or one parent’s death, male demographics also provide information regarding the proportion of never married people (which may be under-counted for women, when relying on the presence of children as an indicator.)

[Note: of course men and women aren’t directly comparable because one possible pattern would be for a smaller group of men sequentially marrying multiple women, resulting in lower never-married rates for women and higher rates for men.]

Men typically married younger women, giving women a good chance of outliving their husbands. But widows, especially older ones, often chose not to remarry. Men who married young would have wives of similar age (due to age limits for women to be considered marriageable), but those who married (or remarried) later were typically more than 10 years older than their wives. Overall this led to a surplus of single younger men and some men who never had the opportunity to marry.

Of the 283 adult men in the census (excluding enslaved people), 91 did not have a wife or their own children living with them, meaning about 1/3 of all men gave no evidence of ever having been married. (Though they may have had past childless marriages or ones where the children had died, similarly to the stats for women.) This also means that the vast majority of single men (3/4) had no living children, while ¼ were single with children.

[Note: given that we observed that about a half of currently-single women had children living with them, this suggests the possibility that currently-single men were more likely to have been never-married than currently-single women. But I think these numbers also raise questions about the claim that children typically went with their father's household, unless the vast majority of the single-with-children women were widows as opposed to divorcées. If so, this would have been a useful clue to apply to hypotheses about relative numbers of divorced versus widowed women.]

The pressure and opportunities for marriage varied according to gender and age. The normative pattern was for women to be married by 20 and for men to experience pressure if still unmarried by their mid-30s. But having been married, the pressure to remarry was lower, e.g. for widows with children. And the high proportion of singles reflect these differing pressures.

In Roman Egypt inheritance was not gendered – daughters inherited equally with sons, and children inherited from both parents. Spouses did not inherit from each other. The typical household structure involved multiple married couples related by the male line, with their children, including unmarried adult daughters and married sons. Women, when divorced or widowed, usually returned to their father’s household. Women with surviving parents and unmarried siblings tended to marry later (mid 20s) while those whose parents were dead or with married brothers tended to marry earlier. The author suggests that this tendency might reflect social dynamics where there was increased friction between a young never-married woman and her sisters-in-law when sharing the same household. Or that women who expected to receive a substantial inheritance, but whose parents were still alive, may have felt either enabled or pressured to postpone marriage.

By some statistics, 3/5 of women had married by age 20 and nearly all by their late 20s. How then do we explain the relatively higher rate of single women in their mid 20s? The author suggests it may represent childless divorcées or widows (whose children lived with their father) who had returned to their birth household. But interpreting these statistics involves guesswork and assumptions.

The article now presents some case studies.

1. A 24 year old woman, Senosiris, never married (apparently), living with her parents, a younger (but marriageable) sister, and an older married brother, with his wife and 2 infants.
2. 2. A 40-year-old woman, Tereus, lived with her parents and an 8-year-old brother. [Note: That’s quite an age gap for the siblings, so perhaps there’s a second marriage involved?]
3. 3. A three-person household consisting of a 56-year-old man and two 40-something sisters with no evidence of marriage or children for any of them. The 2 sisters—but not the brother—are listed as owning the house, and the household was not poor, as it included 2 enslaved men.
4. 4. More unmarried siblings forming a household: two 20-something women who owned their house.

Several more examples are given, focusing on single men living in extended households.

The overall conclusions are that, despite the social context that assumed marriage as the normative life, a significant proportion of the Roman Egyptian population was unmarried at any given time, either never married or not remarried after divorce or spousal death. The reasons in specific cases are nearly impossible to uncover, but personal circumstances could clearly affect both the ability and the desire to refrain from marriage. Yet the lives, expectations, and more informal liaisons of these singles are absent from letters and documents of the time, which helps provide the illusion of universal marriage.

This raises the question of whether the rise of religious singlehood in the fourth century under Christianity was a true demographics shift or simply a new option for reframing the context of singlehood.