8
FAMILIES IN EARLY PTOLEMAIC EGYPT
Dorothy J. Thompson
When Alexander of Macedon took Egypt from the Persians in 332 BC his conquest marked the start of a strong and long-lasting Greek presence in Egypt. But Egypt was not alone. Elsewhere in the former Persian Empire those Macedonians and Greeks who accompanied or followed Alexander settled and made their new homes. In considering Egypt, I am really involved in a case study, since Egypt is one of the few areas where we can begin to trace Greek impact on and reaction to the existing culture and society of this new hellenistic world. The context, then, for this study of the family is a broad one. The study itself is one in detail, involving a demographic investigation of tax material from the third century BC.
In her study of Families in Classical and Hellenistic Greece, Sarah Pomeroy based her careful picture of the hellenistic family in Egypt primarily on the texts of one particular tax-man’s dossier and, in doing this, she acknowledged the possibility that the publication of further evidence might change the picture.1 In what follows, some of what I argue runs counter to her conclusions on the nature of Greek family structure among the settler families of Ptolemaic Egypt in the third century BC. Nevertheless, the pioneering importance of her work in this area should be acknowledged.2 In bringing some new material into the discussion, I am also here concerned as much with the Egyptians as with the early Greek settlers in Egypt. In the main, however, it is documents which have already been published – in the two main languages of Ptolemaic Egypt, in Greek and in demotic – that I exploit to investigate family size and structure, and possible ethnic difference between the two main groups of the population – the Egyptians and Greeks – in the first hundred years of Ptolemaic rule that followed Alexander’s death in 323 BC.
First, the material which lies at the base of this study. The documents used here will shortly be published (or republished) as part of larger project, as the first part (P.Count) of W. Clarysse and D.J. Thompson, Counting the People. P.Count consists of a group of texts produced in connection with the Ptolemaic census and the collection of the salt-tax, the main personal tax which was levied on the basis of this census.3 These texts have all been preserved as mummy casing, as papyri, that is, mixed with lime and recycled as a form of papier mâché used to provide covering for a mummy, as head-pieces, pectorals, and even shoes.
There are two main forms of register. First, there are household listings in which individual households form the organizing principle and, secondly, there are occupational registers in which household information has been subordinated to occupational categories.4 In this form, our registers are those of adult taxpayers, with names, relationships and family totals recorded; in just one exceptional case the ages of some of the taxpayers are also provided.5 That information was standard in the Roman period but earlier, as the system was developing, age indications are rarely found. What, then, Clarysse and I can do with our material is far less than Bagnall and Frier did with the Roman census material in their important demographic study on Roman Egypt.6 It is only through use of the comparative material from the Roman period that we can posit hypotheses on the age structure of the population, on the age gap in married couples, and on the fertility and life expectancy of different groups in the population. Nevertheless, by turning the names of our taxpayers into numbers and by classifying those registers where household totals survive, we can begin to build up a picture of family and household structure from what at first sight might seem quite unpromising material.
And whereas Bagnall and Frier had somewhat under 300 census declarations from a period of 250 years that formed the basis of their study, our database of Ptolemaic tax-households numbers 427, all from a few decades within one period – from the third century BC. Besides being limited in time, this material is further limited in its geographical scope. Given the damper climate of the coast, no papyri have survived from Alexandria or from the Delta, except in carbonized form. All our texts come from mummy-casing from Middle Egypt, from just two administrative areas – from the Fayum, known as the Arsinoite nome, and from the Oxyrhynchite nome; we lack comparable registers from the south. Our texts, as already mentioned, are in both Greek and demotic, and the collocation of these two bodies of material is an essential part of the enterprise. Within the database, when details are known, we tag our families as Greek or Egyptian.
Here we meet our first problem – the problem of the identification of ‘Greeks’ and ‘Egyptians’. This has nothing to do with the language of the text but is essentially the question of how far the nationality of names can be seen as an indication of the nationality of those that carried them. Earlier, it was generally accepted that in the first century of Ptolemaic rule, such an identification was possible.7 Recent work, however, has demonstrated that in some cases the same individuals might be known by either a Greek or an Egyptian name, depending on the context. Thus in one context a man may appear as Seleukos; in another he was known by his Egyptian name of Sokonopis.8
There is, further, the question of Hellenic status – to be Hellēn in Greek or Wynn in demotic was a favoured tax-status rather than an indication of ethnic origin – and it becomes clear from our texts that some at least who enjoyed this status, though they might be known by names that were Greek, in fact derived from families where other family members had names that were Egyptian.9 Along with a whole host of other ethnic indicators, names form just one identifier, and sometimes they may be misleading. So when I here use the terms ‘Greek’ and ‘Egyptian’ what I refer to is not primarily to do with origin but is shorthand rather for those who presented themselves and were accepted by others as belonging to one or the other sector of what was, in effect, a developing and quite complex society.
Nevertheless, as we shall see, there is sufficient difference to be documented between these two groups to allow the conclusion that most of those with Greek names were what we would call ethnic Greeks and, even if some of those with Egyptian names might actually enjoy Hellenic tax-status, most were probably of good Egyptian background.
The second necessary preliminary is to remind the reader that since this is tax-material with which are dealing, it is adults only that are listed. We lack the names or numbers of the children – a real disadvantage, and again one that contrasts poorly with the Roman data.
What comes out of this material? What can we learn of early Ptolemaic families? First, there was a notable difference in the average size of Greek and Egyptian families of the third century BC. (‘Families’ are what we call the units made up of related family members, without any extra non-kin members. ‘Households’, in contrast, are what we call the larger units, where non-kin family dependents and slaves are added to the family group.) The size of a family unit, as indeed that of a household, will differ according to a range of demographic, economic and cultural factors. The age at marriage of offspring with its related effect on fertility, whether or not the newly married couple form a separate menage or stay living at home, the accepted treatment of elderly parents, whether a family is urban or rural and the social and economic status of the household head are all of them factors likely to affect the size of a family unit. So too, it seems, from what we find, was the ethnic affiliation of the family – whether it was Egyptian or Greek.
On the figures of our database, the average number of adults to a family unit was 2.75. That is the figure for both Greeks and Egyptians together. To work from this relatively secure figure – insofar, that is, as any figures compiled for fiscal ends can ever be secure – to an average family size including children is far less certain. Without sure knowledge of the age-range of the tax-paying population covered in our registers, any multiplier adopted for this calculation involves an element of guess-work. If we adopt the multiplier of 2.909 derived from the Roman material and apply it to the figure of tax-paying males, we get the following results: 4.2 for an average family size, 4.0 for Egyptians and 4.4 for Greeks (see TABLE 1).10
TABLE 1. Full family size (third century BC).
| No. of males | No. of families | Average size | |
|---|---|---|---|
| All families | 618 | 425 | 4.2 |
| Egyptian families | 354 | 255 | 4.0 |
| Greek families | 248 | 162 | 4.4 |
Although only approximate, these figures are probably within the right range. From the Roman census material, where children are recorded, Bagnall and Frier reckoned an average of 4.3 for ‘principal resident families’. The most striking feature of these figures is the somewhat larger size of Greek families. This is a feature that we shall find recurs elsewhere in our material.
Greeks then – or those families where the name of the household head was Greek – lived in larger than average families as can be clearly seen in TABLE2 together with Fig. 1, where family totals in the database as a whole (black) are followed by those for Greek (white) and Egyptian (grey) families, divided according to the ethnic affinity of the name of the household head. Adults only are recorded here.
Table 2. Family size in the third century BC (adults only).
| Family size | All families | Greek families | Egyptian families | |||
|---|---|---|---|---|---|---|
| No. | % | No. | % | No. | % | |
| 1-adult families | 76 | 17.8 | 49 | 30.2 | 27 | 10.5 |
| 2-adult families | 167 | 39.2 | 44 | 27.2 | 119 | 46.5 |
| 3-adult families | 83 | 19.5 | 28 | 17.3 | 55 | 21.5 |
| 4-adult families | 50 | 11.7 | 20 | 12.3 | 29 | 11.3 |
| 5-adult families | 18 | 4.2 | 5 | 3.1 | 12 | 4.7 |
| 6-adult families | 15 | 3.5 | 6 | 3.7 | 9 | 3.5 |
| 7-adult families | 9 | 2.1 | 4 | 2.5 | 3 | 1.2 |
| 8-adult families | 4 | 0.9 | 2 | 1.2 | 2 | 0.8 |
| 9-adult families | 3 | 0.7 | 3 | 1.9 | ||
| 12-adult families | 1 | 0.2 | 1 | 0.6 | ||
| Totals | 426 | 162 | 256 | |||
Fig. 1. Family size in the third century BC (adults only).
Fig. 1 clearly shows two things: first that 2 adult-households were by far the most common form of family unit, and secondly that, whilst at the upper end of the scale no Egyptian family home contained more than 8 adults, Greek families listed in our tax-data might number up to 12 adults.
When non-kin family members are added to our data, then the contrast between Greeks and Egyptians becomes even stronger. TABLE3 records the average size of a household, including children (reckoned once again on the basis of adult males x 2.909):
TABLE3. Full household size (third century BC).
| No. of males | No. of households | Average size | |
|---|---|---|---|
| All households | 651 | 425 | 4.5 |
| Egyptian households | 355 | 255 | 4.0 |
| Greek households | 280 | 162 | 5.0 |
For families, there was an average of 2.75 adults to a unit; for households, the average stood at 2.97, with 2.7 for Egyptian households and 3.3 for Greek. When calculated to include children, the average size is 4.5, for Egyptians the number remains the same as for families (4.0) but the size for an average Greek household stands at 5.0.
The difference between these two sets of figures – for families and households – is accounted for by the non-kin dependents that are listed in many of the larger Greek households: the wet-nurses, household or workshop slaves, and a variety of other pastoral and agricultural dependents, who have the effect of making the larger families into even larger households, as illustrated in TABLE4 with Fig. 2 below:
TABLE 4: Household size in the third century BC (adults only).
| Household size | All households | Greeks | Egyptians | |||
|---|---|---|---|---|---|---|
| No. | % | No. | % | No. | % | |
| 1-adult households | 70 | 16.4 | 44 | 27.0 | 26 | 10.2 |
| 2-adult households | 163 | 38.2 | 40 | 24.5 | 119 | 46.5 |
| 3-adult households | 82 | 19.2 | 29 | 17.8 | 53 | 20.7 |
| 4-adult households | 50 | 11.7 | 18 | 11.0 | 31 | 12.1 |
| 5-adult households | 22 | 5.2 | 8 | 4.9 | 13 | 5.1 |
| 6-adult households | 13 | 3.0 | 4 | 2.5 | 9 | 3.5 |
| 7-adult households | 15 | 3.5 | 10 | 6.1 | 3 | 1.2 |
| 8-adult households | 4 | 0.9 | 2 | 1.2 | 2 | 0.8 |
| 9-adult households | 2 | 0.5 | 2 | 1.2 | ||
| 11-adult households | 2 | 0.5 | 2 | 1.2 | ||
| 13-adult households | 1 | 0.2 | 1 | 0.6 | ||
| 14-adult households | 1 | 0.2 | 1 | 0.6 | ||
| 15-adult households | 1 | 0.2 | 1 | 0.6 | ||
| 22-adult households | 1 | 0.2 | 1 | 0.6 | ||
| Totals | 427 | 163 | 256 | |||
From the evidence of all 427 households, it is clear that two adults still formed the most common unit, representing also the unit of habitation for the largest group in the population. At the upper end of the scale, however, the size of household units is noticeably larger than found for families only. And when these household figures are broken down according to the ethnic affinity of the household head, a more varied picture emerges. The larger households of the Greeks are striking. They illustrate well the settler position within third-century society, a position of predominance that was reinforced by the number of household slaves and other dependent staff who are documented for these households. With one household of 22 (it is unknown how many of these were actual family members), one each of 13, 14 and 15 adults, and two of 11 and 9, the larger Greek households stand out in contrast to those of Egyptians, where the largest households are those of 8 adults. Slaveholding is found to be primarily a Greek phenomenon, as too is the occurrence of other resident household staff: cowherds, shepherds, goatherds and agricultural workers.
Fig. 2. Household size in the third century BC (adults only).
Size of household is, however, just one way of documenting difference. Marked differences are also found in other areas – like the occupation of the household head. The larger Greek households are, not surprisingly, those of the military settlers. All Greek households made up of more than 10 taxpayers are headed by military men, by cavalry cleruchs settled with land in the Arsinoite and Oxyrhynchite nomes known from texts dated 230–229 BC. It was the cavalry cleruchs who, without a doubt, formed the elite of the rural landscape in the third century BC. These were the economic-ally privileged of Ptolemaic society, and their larger houses and households are just one more sign of this status. As we know from their further designation as 100-aroura cleruchs, at this date the immigrant cavalry settlers were also endowed with large plots of land – 100 arouras is 27.5 hectares or some 67 acres. The evidence of our salt-tax registers now allows us to recognize this wealth and status also in terms of the size of their households. And in these particular cases, the use of Greek names does on the whole appear coterminous with origin and ethnicity. These cavalry cleruchs were probably immigrant Greeks and, in the case of the Oxyrhynchite settlers of two of our texts, more specifically Greeks from Cyrene.11 Their domination in terms of both land and household size is a feature of Ptolemaic Egypt in this period, at least in this part of Egypt.
In the meantime, the Egyptian inhabitants of the country lived in far smaller households. Simple households of two adult taxpayers formed the most common unit, and many of these were of conjugal pairs. And in the smaller units of those with Egyptian names the reality of everyday life for the native population in the new society of Ptolemaic Egypt can be found reflected. How far life had changed from under the Persian overlords cannot be known. Nevertheless, it is clear that the smaller households of the Egyptian villagers form a notable measure of their lesser economic status in rural society. Their smaller plots of land belonged to the crown and, in contrast to the cleruchic land of the settlers, there were rents to pay on them. Their households were smaller; they lacked the family backup and the slaves that form a more regular feature of the larger homes of the settlers.
So far we have been considering simply the size of the different units within the population but our data allow us to go somewhat further. And if Ptolemaic Egypt is to be added to the demographic discussion of family history, then some analysis must be made of the types of families found. Was it primarily the nuclear family that is documented or perhaps the more extended and multiple families that have been seen as typical of pre-modern Mediterranean forms of domestic organization? The categorization of family types, the so-called ‘Cambridge typology’ developed by Peter Laslett and his colleagues for work on the family history in Europe, is the framework adopted by Bagnall and Frier in their study of census returns from Roman Egypt. Since this is by far the closest material for comparison with our data, it seems right to employ the same categorization here.
The central categories of this typology, in the simplified form used by Bagnall and Frier, are as follows:12
TABLE 5. Cambridge family typology.
| 1. Solitary persons; those who live alone, whatever their marital status. |
| 2. Multiple persons with no conjugal family present (mainly co-resident siblings). |
| 3. Simple family households; conjugal families in their various phases (from a married couple without children, through to a formerly married parent with unmarried children). |
| 4. Conjugal families extended through the presence of co-resident kin; groups of co-resident siblings with only one brother married. |
| 5. Multiple families, usually linked by kinship. This includes both households in which children remain after they marry and frérèches consisting of co-resident siblings, more than one of whom is married. |
| 6. Incompletely classifiable households. |
It is important to realize that this is primarily a family typology based on the kin group, those co-residential individuals who make up the family group. Non-kin dependents, whether slave or free, have no effect on the classification used to describe the family members. Their presence or absence, of course, as we have just noted, is crucial to any assessment of the economic and social role of the household, but in the categories used for classification that presence is not a visible one. With this proviso and with the reminder that our information is always for tax-households – for adults only – we may look at the composite picture of our Ptolemaic family forms. First (TABLE6 and Fig. 3) comes a summary typology of family types, with details for Greeks and Egyptians, both separately and combined, followed (in TABLE7 and Fig. 4) by the numbers of adults living in these different types of family.
TABLE 6: Family structure by type (third century BC).
| Types | All | Greeks | Egyptians | |||
|---|---|---|---|---|---|---|
| No. | % | No. | % | No. | % | |
| 1. Solitaries | 76 | 17.8 | 49 | 30.1 | 27 | 10.5 |
| 2. No conjugal family | 14 | 3.3 | 6 | 3.7 | 8 | 3.1 |
| 3. Conjugal families | 188 | 44.0 | 61 | 37.4 | 123 | 48.0 |
| 4. Extended families | 51 | 11.9 | 21 | 12.9 | 29 | 11.3 |
| 5. Multiple families | 75 | 17.6 | 19 | 11.6 | 54 | 21.1 |
| 6. Non-classifiable | 23 | 5.4 | 7 | 4.3 | 15 | 5.9 |
| Totals | 427 | 163 | 256 | |||
| 1+3 combined | 264 | 61.8 | 110 | 67.5 | 150 | 58.6 |
Fig. 3. Family structure by type (third century BC).
TABLE 7: Adults by family type (third century BC).
| Types | All | Greeks | Egyptians | |||
|---|---|---|---|---|---|---|
| No. | % | No. | % | No. | % | |
| 1. Solitaries | 85 | 6.7 | 57 | 10.6 | 28 | 4.0 |
| 2. No conjugal family | 37 | 2.9 | 18 | 3.4 | 19 | 2.7 |
| 3. Conjugal families | 467 | 36.7 | 187 | 34.8 | 272 | 38.7 |
| 4. Extended families | 191 | 15.0 | 97 | 18.1 | 9. | 12.8 |
| 5. Multiple families | 395 | 31.1 | 134 | 25.0 | 249 | 35.4 |
| 6. Non-classifiable | 96 | 7.6 | 44 | 8.2 | 45 | 6.4 |
| Totals | 1271 | 537 | 703 | |||
| 1 + 3 combined | 552 | 43.4 | 244 | 45.4 | 300 | 42.7 |
Fig. 4. Adults by family type (third century BC)
Several features immediately stand out. First to note (Fig. 3) is the large number of solitaries (type 1) among these families, especially among the Greeks for whom this type accounts for over 30% of all households containing nearly 11% of family members. In part, this imbalance derives from one specific list, where numerous singleton women are recorded in an army context.13 This is a particularly difficult and incomplete register which constantly skews our results. But the problem of solitaries is also in part the result of the composition of our data, recording adults only. Were children listed, some of the apparently single women would doubtless turn into mothers with children and move from type 1 to type 3; the presence of children would certainly have modified the picture. This is why, at the foot of TABLES 6 and 7 (and to the right of Figs. 3 and 4), types 1 (solitaries) and 3 (conjugal families) have been amalgamated. This gives a far fairer picture of reality than does the straight application of this family classification to records of adults only.
Next, among both groups, as is clear from Fig. 3, it is the conjugal family (type 3) which accounts for by far the largest family type. For all households, some 44% belong to this type, with (Fig. 4) almost 37% of all adults living in this type of menage. When the material is divided into the two main groups we find that such households were more common among Egyptians – 48% of all Egyptian households with almost 39% of adults – than among Greeks – 37% of households with almost 35% of adults.
The same material can be presented to show the gender divide, with adult numbers of males and females for the different family types. (In using ‘males’ and ‘females’, we adopt the bureaucratic jargon of the original documents where tax-persons, sōmata, are divided into male, arsenika, and female, thēluka). The gender breakdown for Greeks (TABLE8 with Fig. 5) and then for Egyptians, by both family and household (TABLE 9 with Fig. 6), is as follows:14
TABLE 8. Gender breakdown for Greek families and households.
| Types | Family adults | Household Adults | ||||||
|---|---|---|---|---|---|---|---|---|
| m. | f. | total | % | m. | f. | total | % | |
| 1. Solitaries | 24 | 25 | 49 | 11.0 | 27 | 30 | 57 | 10.6 |
| 2. No conjugal family | 7 | 8 | 15 | 3.4 | 9 | 9 | 18 | 3.4 |
| 3. Conjugal families | 89 | 67 | 157 | 35.3 | 102 | 84 | 187 | 34.8 |
| 4. Extended families | 48 | 35 | 83 | 18.7 | 58 | 39 | 97 | 18.1 |
| 5. Multiple families | 58 | 41 | 99 | 22.2 | 62 | 50 | 134 | 25.0 |
| 6. Non-classifiable | 22 | 20 | 42 | 9.4 | 22 | 22 | 44 | 8.2 |
| All | 248 | 196 | 445 | 280 | 234 | 537 | ||
Fig. 5. Gender breakdown for Greek families and households.
TABLE9. Gender breakdown for Egyptian families and households.
| Types | Family adults | Household Adults | ||||||
|---|---|---|---|---|---|---|---|---|
| m. | f. | total | % | m. | f. | total | % | |
| 1. Solitaries | 16 | 11 | 27 | 3.9 | 16 | 12 | 28 | 4.0 |
| 2. No conjugal family | 16 | 3 | 19 | 2.7 | 16 | 3 | 19 | 2.7 |
| 3. Conjugal families | 128 | 139 | 270 | 38.7 | 129 | 140 | 272 | 38.7 |
| 4. Extended families | 42 | 48 | 90 | 12.9 | 42 | 48 | 90 | 12.8 |
| 5. Multiple families | 129 | 117 | 246 | 35.3 | 129 | 120 | 249 | 35.4 |
| 6. Non-classifiable | 23 | 22 | 45 | 6.5 | 23 | 22 | 45 | 6.4 |
| All | 354 | 340 | 697 | 355 | 345 | 703 | ||
Fig. 6. Gender breakdown for Egyptian families and households
In considering this information, it is family rather than the household adults which are of greater interest. Once again, the solitaries category (type 1) causes problems and since knowledge of children in households would change this picture, let alone the problem of the one rogue register already mentioned, it seems best simply to ignore this group. As already noted, it is the conjugal family which accounts for by far the most common family type, somewhat more common among Egyptians than Greeks. In this particular group, where adult sons and daughters are quite often present in the nuclear (or simple family) household, it is the low number of females among the Greeks which is immediately striking (TABLE 8 with Fig. 5). Of Greeks in conjugal families (type 3), women formed only 43% of family members (67 females compared with 89 males), whereas for Egyptians (TABLE 9 with Fig. 6), the proportion of women in conjugal menages stood at 52% (139 females to 128 males). Since women married earlier than men one might expect a smaller number of females also among Egyptian conjugal families. This is not, however, the case. In this particular family type among the Egyptians, as just noted, females outnumbered the males. Besides the daughters still living at home, five two-female menages are in part responsible for this feature; at least four of these were of a mother and adult daughter. Further, two cases of bigamy added to the number of women in Egyptian conjugal families of type 3; both wives were listed within the same household.
For Greeks, overall, two features stand out. First, as just noted for conjugal families, there is the overall feature of the lower number of females (see TABLE 8). Whereas in family types 1, 2, and 6 women slightly outnumber men, in types 3, 4 and 5 the family sex ratio is noticeably elevated, with men outnumbering the women. Sex ratios are always expressed in relation to 100 women. A sex ratio of 105 means that there are 105 men to 100 women, a low ratio of 88 involves far fewer men, with just 88 or them to every 100 women. In conjugal families, then, in extended and multiple families (types 3, 4 and 5) among the Greeks, the sex ratio stood at 132.8, 137.1 and 141.5, with an average for these three groups of 136.4. Fuller details for this phenomenon is provided in TABLE 10 below:
TABLE10. Sex ratios in Greek families.
| Types | m. | f. | total | as % | sec ratio |
|---|---|---|---|---|---|
| 1. Solitaries | 24 | 25 | 49 | 11.0 | 96.0 |
| 2. No conjugal family | 07 | 08 | 15 | 3.4 | 87.5 |
| 3. Conjugal families | 89 | 67 | 157 | 35.3 | 132.8 |
| 4. Extended families | 48 | 35 | 83 | 18.7 | 137.1 |
| 5. Multiple families | 58 | 41 | 99 | 22.2 | 141.5 |
| 6. Non-classifiable | 22 | 20 | 42 | 9.4 | 110.0 |
| Totals | 248 | 196 | 445 | 100.0 | 126.5 |
There is overall, it is clear, an under-representation of females in these groups, with far fewer daughters than sons recorded.
It is this feature that I want to explore further – the apparent shortage of females among certain Greek family groups in our database. First, however, let us put some flesh and bones onto these facts and figures. In one of the census declarations made for the salt-tax, we find the following group: a military man Leptines from Pisidia, his wife Hedyle, their four sons Glaukias, Moirikon (or Myrikon), Nikandros and Theophilos, and their daughter Baia, together with an extensive household of slaves, most probably – given their numbers – workshop slaves in two different locations.15 The typicality of this Greek army family, with four sons but just one daughter, becomes clear only when placed in the context of the wider set of data just presented. For it is only when we can quantify our data that we can begin to think in terms of the typical ‘Greek’ or ‘Egyptian’ family.
How representative is the material of our database? A glance at what we know of contemporary sex ratios can perhaps be used to support the wider applicability of our material, at least for the mid-third century BC. Sex ratios are one of the most important, yet elusive, demographic factors that affect the changing structure of a population and its different household patterns. These ratios, in turn, are themselves the product of demographic factors – the sex ratio at birth, age at marriage, fertility levels, or differential mortality rates. All of these, of course, are influenced both by local and more general factors, but this is particularly the case for rates of mortality. In more recent Egypt, for instance, bilharzia has been a major underlying cause of disease and death with a greater risk to males, who work in the fields, than to females, whose contact with Nile water is less frequent; this seems likely to have been an ancient problem too.16 Males too bear the brunt of warfare, though in Egypt of the third century BC this particular hazard was one more likely to affect the minority Greek population than Egyptians. For women, in contrast, the dangers of childbirth and of disease that attacks the undernourished have always been serious problems. Above all, however, a differential sex ratio results from different cultural and social attitudes within a given society. If the most extreme example is that of contemporary China, where a one-child policy combined with a preference for sons has seriously affected the natural ratio, imbalances in the sex ratio are to be found in a wide range of societies in all historical periods. This is equally the case in our material.
TABLE11 presents the sex ratios surviving in our material. In this table, figures in italics are (reasonably safely) supplied and the final column to the right gives what is a rough guide to the major component of the population concerned: E(gyptian), G(reek) or M(ixed).
TABLE 11. Adult sex ratios from Ptolemaic tax-documents.
| Population with reference | males | females | ratio | pop. |
|---|---|---|---|---|
| Arsinoite nome: P.Count 1 (254–231 BC) | 28,512 | 30,197 | 94.4 | M. |
| Cleruchs?: P.Count 1 (254–231 BC) | 3,472 | 3,147 | 110.3 | G |
| Serving cavalry: P.Count 1 (254–231 BC) | 1,426 | 1,080 | 132.0 | G |
| Total army: P.Count 1 (254–231 BC) | 4,898 | 4,227 | 115.8 | G |
| Civilian adult population: P.Count 1 | 23,614 | 25,970 | 90.9 | M |
| Arsinoite villagers | 8,163 | 7,980 | 102.3 | M |
| Themistos tax-area: P.Count 2.475–477 (229 BC) | 5,245 | 5,631 | 93.1 | M |
| District A for Year 19: P.Count 2.470 | 1,174 | 1,174 | 97.0 | M |
| District B for Year 18: P.Count 3.1–5 (229 BC) | 1,289 | 1,229 | 104.9 | M |
| District B for Year 19: P.Count 2.471; 3.6–9 | 860 | 829 | 103.7 | M |
| District C for Year 18: P.Count 3.139–143 | 740 | 838 | 88.3 | M |
| District C for Year 19: P.Count 2.472; 3.144–147 | 727 | 852 | 85.3 | M |
| District D for Year 19: P.Count 2.473 | 1,266 | 1,399 | 90.5 | M |
| District E for Year 19: P.Count 2.474 | 1,218 | 1,351 | 90.2 | M |
| ‘Hellenes’ within this tax-area: P.Count 2.484 | 862 | 894 | 96.4 | G |
| Village families: P.Count 2.1–145 | 37 | 34 | 108.8 | E |
| Cavalry + veteran families: P.Count 2.278–434 | 53 | 30 | 176.7 | G |
| Cavalry + veteran households: P.Count 2.278–434 | 69 | 48 | 143.8 | G |
| Themistos meris: P.Count 11.28–31 (243–210 BC) | 8,795 | 8,253 | 106.5 | M |
| District A: P.Count 11.28–31 | 789 | 793 | 99.5 | M |
| District B: P.Count 11.28–31 | 782 | 713 | 109.7 | M |
| District C: P.Count 11.28–31 | 1,288 | 1,183 | 108.9 | M |
| District D: P.Count 11.28–31 | 1,514 | 1,644 | 92.1 | M |
| District E: P.Count 11.28–31 | 926 | 840 | 110.2 | M |
| District F: P.Count 11.28–31 | 917 | 752 | 121.9 | M |
| District G: P.Count 11.28–31 | 628 | 573 | 109.6 | M |
| District H: P.Count 11.28–31 | 574 | 513 | 111.9 | M |
| District I: P.Count 11.28–31 | 1,377 | 1,242 | 110.9 | M |
| Herakleides meris: P.Count 12.135–138 (243–210 BC) | 5,352 | 5,067 | 105.6 | M |
| Polemon meris: P.Count 8.1–3 (243–210), tax-district | 806 | 954 | 84.5 | M |
| Herakleopolite tax-area: P.Count 45.3–5 (243–210 BC) | 5,645 | 5,480 | 103.0 | M |
| Database adults: family figures without dependents | 618 | 551 | 112.2 | M |
| Database adults: household figures with dependents | 651 | 594 | 109.6 | M |
| Egyptian families | 354 | 340 | 104.1 | E |
| Egyptian households | 355 | 345 | 102.9 | E |
| Greek families | 248 | 196 | 126.5 | G |
| Greek households | 280 | 234 | 119.7 | G |
| Greek epigonoi, Oxyrhynchite: P.Count 47 (230 BC) | 151 | 223 | 67.7 | G |
| Lykopolite villagers: P.Count 53 (second century) | 181 | 178 | 101.7 | E |
The ratios of our database can be found towards the foot of the column. For families overall the ratio is 112 (110 for households). These are then broken down, where names are known, into Egyptians and Greeks. It is the figure of 126.5 males to 100 females for Greek family adults which is where we started this investigation.
Two features of these different ratios may be noted. First, not surprisingly, it is clear that the smaller the sample, the more variation there is likely to be in the sex ratio. A larger population, such as that for the Arsinoite nome, at the very head of the list, is in practice made up of many different families and mixed communities that individually exhibit a wide range of different ratios. This can be seen most clearly in the make-up of the civilian tax-areas recorded in P.Count 2–3 and P.Count 11, where apparently wild fluctuations between the different constituent districts may be charted. How are these differences to be explained? Either the quality of our records is responsible or real differences in the gender distribution are to be found in different sections of the population. Figures, however, for tax-areas and districts are still relatively large, and it is at a lower level that the greatest anomalies appear. So, among the community of the Greek epigonoi in P.Count 47, the high number of females, which has kept skewing our ‘solitaries’ figures, results in an exceptional adult ratio of just 68 males to 100 females; the total number here, however, is small (374) and the list is not complete. Fluctuations like this are interesting, but hardly of broader relevance.
Among Greek army families, the picture is different and distinctive. Indeed, the second feature of this material is the higher ratio found in all our data (with the familiar exception of P.Count 47) for the Greek sector of the community – for those households, that is, where the name of the household head is Greek. This appears most clearly in the army figures for the Arsinoite nome given at the head of the table (the fourth item down), where an overall army ratio of 116 is made up of a comparatively higher ratio of 132 for the serving cavalry, the misthophoroi hippeis, and of a somewhat lower ratio of 110 for the larger group, most probably that of the cleruchs. A similar ratio, as already noted, is found for the Greeks in our database (126.5), including both Arsinoite and Oxyrhynchite army families. The contrast with Egyptians in the database is striking; among Egyptian family members the sex ratio is 104.
What are we to make of this? An explanation for the marked difference in ratio between the two main groups of Egyptians and Greeks might be made in terms of differential social practices, in terms of either concealment of females by their families in registration or neglect of females by the recorders. The latter is less likely at this period than later under the Romans, when women were no longer liable to tax. Under the Ptolemies, when women were liable for the salt-tax in the same way as men, it was in the interests of both the wider administration and the tax-collectors that women too were recorded. The concealment of females in registration is, further, inherently unlikely, since girls were more likely to be at home than were their brothers when officials came round.
A further explanation for the difference between Greeks and Egyptians may lie in their different attitudes to the birth of daughters. Despite Pomeroy’s claims that it is first in Roman Egypt that good evidence is found for the exposure of children, what we find in this early Ptolemaic material points the other way.17 Classical writers were struck by the fact that the natural wealth of Egypt supported its population. ‘And [sc. among the Egyptians] of necessity they raise all the children born to them in order to increase the population … ’ is how Diodorus Siculus reported on the practices of Egypt, where a large population was considered the key to wealth and prosperity for both cities and countryside.18 And Strabo, who visited Egypt under the early empire, commented that ‘one of the customs most zealously observed among the Egyptians is this, that they rear every child that is born … ’.19 In Strabo’s observation, we should perhaps read an implicit contrast with practices elsewhere.
The higher sex ratios for Greeks that are clear in our tax-material, when compared with those for Egyptians, might seem to provide supporting evidence for the practice of selective infanticide (for femicide, that is) or – more probably – for exposure within this sector of the community. That this was standard practice is suggested by a first-century BC temple ruling, from the Greek city of Ptolemais in southern Egypt, which specifies a purification period of fourteen days for the partner of a woman exposing a child.20 Pomeroy attempts to discount this particular piece of evidence,21 but in my view it forms part of what is now a wider picture of the possible fate of daughters among the settler population of Ptolemaic Egypt. Of course, not all girls were exposed – Leptines did have one daughter Baia and this was not an uncommon case. Indeed, first daughters, especially if also first child, were more likely to have been reared, insofar as rearing was possible in such a society with high perinatal mortality for both mother and child.
If the imbalance between the two sexes that we have found within the Greek sector of the population is to be explained in terms of selective infanticide or exposure, we are left with the problem of the excess of males, not all of whom can have found Greek wives. Where would they find their future wives? Here again the evidence of our database can come into play. Here we find that, as always in such an immigrant situation, some intermarriage with native women by settlers was practised from the start. And, insofar as names can be used to signal ethnic background, the evidence is very telling. Out of the 85 household heads with Greek names in our database who are both male and married, 75 have wives whose names survive; of these, 68 wives (91%) were also Greek while just seven (9%) had names that were Egyptian. Some intermarriage of Greeks with the native population is clear. On the Egyptian side, however, much was unchanged. Not a single household head in our database whose name is Egyptian appears with a Greek-named wife. It is clear that poaching across the ethnic divide was a one-way matter. Within the traditional occupational groups of Egypt, to judge from the names, endogamy continued to be practised. And the recurrence of family names suggests that among Egyptians close-kin marriage was also quite common. And although there is no evidence in our tax-registers for brother-sister marriage, there is, as already mentioned, some for Egyptian polygamy.
To sum up. What I hope that I have shown in this chapter is that from Egypt the survival of papyrus tax-registers enables us, to some degree, to examine demographic questions which cannot be answered elsewhere in the ancient world. In the case of the first century of Ptolemaic rule, when the new Greek settlers, together with their new Macedonian pharaoh, made their home in this ‘antique land’, they brought with them family practices and ways transposed into an older traditional society. How far the differences that we have been looking at here – different sizes of household, different patterns of family living and different treatment of females – are a feature just of the different social and, more particularly, the different economic standing of the settlers and how far they are due to different cultural practices, or whether, indeed, these different strands can ever be divided, are questions that remain.
Notes
1 Pomeroy 1997, 229 n. 134.
2 See further, Pomeroy 1993, 1994, 1996.
3 Clarysse and Thompson 1995, on the salt-tax.
4 For examples, see Thompson 1997, 249–51.
5 P.Count 9 (after 251/0 BC).
6 Bagnall and Frier 1994; cf. Bagnall, Frier and Rutherford 1997.
7 e.g. Peremans 1980/81.
8 Clarysse 1992, 55.
9 See Thompson 1997, 247–8; 2001a, 310–11, for tax-Hellenes.
10 Given the fragmentary nature of some of the texts, the total for families is lower than that for households, and in one case (not included) the gender divide is unknown.
11 P.Count 46 and 47 (230 BC).
12 See Bagnall and Frier 1994, 59.
13 P.Count 47 (230 BC).
14 Discrepancies in these tables between the total figures and those for males and females are the result of illegible or incomplete data.
15 P.Lille I.27 = W.Chrest. 199 = Scholl, Corpus 87 (254–231 BC). Arsinoite; on this text, see further Thompson 2001b.
16 Omran 1973, 18, on contemporary Egypt; Contis and David 1996, 253–5, on the ancient evidence.
17 Pomeroy 1997, 226, discounting SEG 42.1131 (n. 20 below); but cf. 225, noting the lack of unmarried daughters in Greek tax-registers.
18 Diodorus Siculus 1.80.3; cf. Polybius 36.17.7–8, control of family-size as a symptom of decline.
19 Strabo 17.2.5 (C824).
20 SEG 42.1131 with Bingen 1993, 226–7, from Ptolemais (first century BC); Rowlandson 1998, 65, no. 40, for translation.
21 Pomeroy 1997, 226; but cf. eadem 1993, on the Delphinion inscriptions from near-contemporary Miletos which, she argues, show a similar picture of infanticide to that presented here among an immigrant group of new citizens.
Bagnall, R.S., and Frier, B.W.
| 1994 | The Demography of Roman Egypt, Cambridge. |
Bagnall, R.S., Frier, B.W. and Rutherford, I.C.
| 1997 | The census register P. Oxy. 984: the reverse of Pindar’s Paeans, Papyrologica Bruxellensia 29, Bruxelles. |
Bingen, J.
| 1993 | ‘La lex sacra SB I 3451 = LSCG. Suppl. 119 (Ptolémaïs, Haute-Égypte)’, Chronique d’ Égypte 68, 219–28. |
Clarysse, W.
| 1992 | ‘Some Greeks in Egypt’, in J.H. Johnson (ed.) Life in a Multi-cultural Society: Egypt from Cambyses to Constantine and beyond. Studies in Ancient Oriental Civilization 51, 51–6, Chicago. |
Clarysse, W. and Thompson, D.J.
| 1995 | ‘The salt-tax rate once again’, Chronique d’ Égypte 70, 223–9. |
Contis, G. and David, A.R.
| 1996 | ‘The epidemiology of bilharzia in ancient Egypt: 5000 years of schistosomiasis’, Parasitology Today 12.7, 253–5. |
Omran, A.R.
| 1973 | ‘The population of Egypt, past and present’, in A.R. Omran (ed.) Egypt: Population problems and prospects, 3–38, Chapel Hill, N.C. |
Peremans, W.
| 1980/81 | ‘Égyptiens et étrangers en Égypte sous le règne de Ptolémée I’, Ancient Society 11/12, 213–26. |
Pomeroy, S.B.
| 1993 | ‘Infanticide in hellenistic Greece’, in A. Cameron and A. Kuhrt (eds.) Images of Women in Antiquity, 2nd edn, 207–22, London. |
| 1994 | ‘Family history in Ptolemaic Egypt’, in Proceedings of the 20th International Congress of Papyrologists, 593–7, Copenhagen. |
| 1996 | ‘Families in Ptolemaic Egypt: continuity, change, and coercion’, in R.W. Wallace and E.M. Harris (eds.) Transitions to Empire. Essays in Greco-Roman history, 360–146 BC, in honor of E. Badian, Oklahoma series in classical culture 21, 241–53, Norman, Okla. |
| 1997 | Families in Classical and Hellenistic Greece: Representations and realities, Oxford. |
Rowlandson, J.
| 1998 | Women and Society in Greek and Roman Egypt: A sourcebook, Cambridge. |
Scholl, R.
| 1990 | Corpus der ptolemäischen Sklaventexte, Forschungen zur antiken Sklaverei Beiheft 1, 3 vols., Stuttgart. |
Thompson, D.J.