THE DISTRIBUTION OF WEALTHY ATHENIANS IN THE ATTIC DEMES
James Kierstead and Roman Klapaukh
1 INTRODUCTION
This chapter investigates the spread of wealthy citizens across the 139 demes of Attica, the constituent villages and boroughs of the Athenian state. We look primarily at the number of citizens (or councillors, a proxy for the number of citizens), and the number of wealthy citizens, in each of the demes, and see how closely they are correlated. A close correlation would suggest that one sort of economic opportunity was smoothly spread; a man’s origins in a particular deme would be a poor way of predicting whether or not he was wealthy. A weak correlation would suggest that some demes did give their citizens a higher chance of being wealthy.1 We also consider demes with an unexpectedly high number of wealthy citizens, and ask whether these outliers had anything in common.
2 LITERATURE REVIEW
A number of scholars have anticipated some of our methods and results. Bresson offered a regression analysis of the number of magistrates and wealthy citizens from the demes of Hellenistic Rhodes.2 Osborne produced a ‘Wealth Index’ for most of the Athenian demes by ‘taking the number of wealthy men attested per deme from Davies’ work and dividing this by the number of bouleutai provided by that deme’.3 Taylor provided a regression analysis of known officials against liturgists in the demes.4 How does our analysis add to theirs? In several ways, we hope.
First, we use a slightly improved dataset. In particular, where our predecessors took paired demes such as Upper and Lower Lamptrai together, we treat them as two separate demes.5
Second, we offer a more fine-grained analysis. We look at the distribution not only of the total number of wealthy citizens identified by Davies (1971), but also of various subsets of wealthy citizens. We reflect on the differences in the distribution of wealthy citizens from these different categories. We also think a lot about the differences that the data throws up between individual demes. In particular, we will think hard about demes that are statistical outliers. This goes against Osborne’s assertion that ‘the figures for individual demes are clearly too subject to the effect of individual chance attestations to be taken too seriously’.6 We discuss the reliability of individual results on a category-by-category basis.
The last way in which we depart from Osborne’s work is the most important. Osborne’s belief that differences between demes were not statistically significant led him to group demes by area, and then to come to conclusions about the wealth of various parts of Attica.7 According to Osborne,
the pre-eminence of the Mesogaia [to the east of Mount Hymettos] reinforces belief in the landed basis of much Athenian wealth, while the comparative poverty of the Thriasian plain [near Eleusis, between the Parnes and Aigaleos ranges] may indicate the degree to which men from other demes, resident in Athens, had a stake in its exploitation.
In addition, two groups of demes – mining demes and those with significant religious sanctuaries – are poorer than might be expected from their population.8 The picture that emerges from Osborne’s analysis is thus one in which there is considerable variation in the wealth of various parts of Attica.
To anticipate our results, we disagree with Osborne about the amount of variability in wealth in Attica. We try our best to argue the point by looking at the variation in our figures for economic opportunity (that is, what we call normalised names) across the demes. As we will see, most demes have very similar figures on this metric. We also disagree about which areas of Attica were relatively privileged.
3 EVIDENCE AND METHODOLOGY
We rely on a dataset compiled by Ober and Teegarden from Davies’ classic work Athenian Propertied Families.9 Davies compiled a register of wealthy Athenians, and included in each entry the most important facts that were known about each man. Ober and Teegarden counted the number of men from each of the 139 demes that appeared in Davies’ register, and put the results into a spreadsheet.
Davies took himself to be defining ‘the Athenian upper class’ in his work, and he believed that there was ‘a useable basic criterion for . . . membership’ in this group. This was the performance of liturgies.10 For Davies, the class of citizens referred to in Athenian texts as ‘the wealthy’ (the plousioi or euporoi) was, to all intents and purposes, co-extensive with the set of men who undertook liturgical services.11
How much wealth did a typical liturgist have, and what proportion of the Athenian citizenry could depend on that level of wealth? Davies’ answers to these questions have met with widespread (if not universal) acceptance. An estate which would have allowed its owner to pay for a liturgy probably brought in about three or four talents of silver a year. And there were probably only three or four hundred estates of this sort in Attica.12 In the fourth century BCE, when the citizen population was only around 30,000, liturgists can thus quite accurately be described as ancient Athens’ ‘one-percenters’.13
Davies also believed that the liturgical class always consisted of three or four hundred men throughout the classical period. Though Periandros’ naval law (c. 357) widened the trierarchy to 1,200 citizens, this was effectively reversed by Demosthenes’ reform in 340, which placed the burden of payment on the 300 wealthiest men.14 For Davies, the failure of Periandros’ arrangements is an indication that the figures of 300 or 400 liturgists were not ‘arbitrarily selected’. Instead, they were ‘in some sense natural and inescapable’.15
Davies’ main criterion for including men in his register consisted in ‘evidence of performance of the military and festival liturgies of the state’.16 By ‘evidence’ here he means ‘direct attestation’, and in referring to military liturgies he is referring to the trierarchy, in which wealthy citizens paid for the upkeep of a warship for a year.17 As for ‘festival liturgies’, there were about a hundred different types of these, but the most important was the chorêgia, in which rich citizens funded a chorus in one of the city’s dramatic festivals.18
In seeking to produce an accurate picture of the Athenian upper class, however, Davies also moved beyond his main criterion for inclusion. In addition to those who are directly attested as liturgists, individuals from four additional categories were included in Davies’ register: (1) those whose estates were large enough for the breeding of horses; (2) those known to be pentakosiomedimnoi (the ‘Five-Hundred-Bushel Men’ of the Solonian census classes); (3) those whose names are present on the diadikasia inscriptions from around 380 (on which more in the next section); and (4) those attested in none of these categories of evidence, but who we have other reasons to believe were extremely wealthy.19
Ober and Teegarden counted the number of men in each deme who were (1) in Davies’ register; (2) directly attested as trierarchs; (3) directly attested as festival liturgists; (4) on the diadikasia inscriptions; and (5) thought by Davies to have been ‘upper-class’ in terms of wealth for some other reason.20 We will discuss what we might learn from these different categories of evidence in more detail in the next section. But first we need to deal with some methodological issues.
The first methodological issue concerns the evidence. As its compiler himself stressed, Davies’ register is not a complete list of Athenian liturgists. It is simply a list of liturgists for whom evidence happens to have survived.21 The fact that this represents only a small sample of the total number of citizens who paid liturgies is not very problematic in itself, at least not for our purposes.22 This is because all that matters for our analysis is that the sample is reasonably representative.
What would compromise our results is if the Davies sample was biased in some systematic way; in other words, if certain demes were more likely to make it into the register for reasons which had nothing to do with their number of wealthy citizens.23 There are two possible sources of bias we need to discuss: Davies’ criteria for including wealthy citizens in the register (and Ober and Teegarden’s arrangement of the data in their spreadsheet), and the differential survival of the literary and epigraphic evidence.
Let us begin with Davies’ criteria for including citizens in his register. It is a possibility that some or all of these criteria are biased towards some demes at the expense of others. But it quickly becomes clear that this is not the case. Davies’ main criterion was payment of liturgies, and the liturgical system was designed to identify wealthy men from all over Attica. Because of the antidosis procedure, in which a man could get out of paying a liturgy if he could find another man rich enough to pay it in his place, citizens were incentivised to gather information about who the richest Athenians were.24 (And the strength of these incentives would not have varied because of geography.) This criterion accounts for most of the names in Ober and Teegarden’s spreadsheet (609 out of 687, or 89 per cent).
The other two columns in Ober and Teegarden’s spreadsheet are for citizens who appear on the diadikasia inscriptions, and citizens who are included in Davies for miscellaneous other reasons. The diadikasia was a legal procedure that involved men who contested their status as elite tax-payers, and there is no reason to believe that men from certain areas of Attica would have done this more than others.25 Now for Ober and Teegarden’s ‘Other Wealthy’ category, which incorporates three of Davies’ four additional criteria for inclusion (i.e. criteria other than direct attestation as a liturgist): first up are men who had estates large enough for horse-breeding. We cannot rule out that these men were more likely to come from rural areas, although Athenians regularly owned estates in other parts of Attica.26 Next up are the pentakosiomedimnoi, a census category that was probably assessed by the central state, and is therefore unlikely to have been skewed towards any particular region.27 There is a slight risk that certain demes had not yet been incorporated into the Athenian polis in the period in which most of the men attested in this category were alive (and thus that outlying parts of Attica are under-represented); but this risk is not a strong one.28 Finally, there is Davies’ ‘Miscellaneous’ category (the register’s category 4 above), but since he does not offer much detail on why these men were included, there is not much that we can say about possible bias in this case.29
So there is some slight risk of selection bias in Ober and Teegarden’s ‘Other Wealthy’ category. At the same time, the reasons to be cautious in each individual case are hardly overwhelming. Moreover, these risks affect only a small proportion of our data points (75 out of 687, 11 per cent). So selection biases produced by the criteria chosen by Davies, Ober and Teegarden should not fatally undermine the reliability of our results. We can now move onto the possibility that biases have resulted from the differential survival of the evidence.
Davies drew on both epigraphic and literary sources for this register.30 In other words, he drew on inscriptions that happen to have survived and texts that happen to have been transmitted to the present day. The question for us is whether the factors affecting the survival of these types of evidence would have been slanted towards certain regions of Attica.
Whatever processes affected the survival of inscriptions (invading armies, re-use of inscribed stêlai by later inhabitants), they seem to have affected different parts of Attica in a similar way. Jones has determined that the number of documents that survive from demes does not vary much between different areas of Attica.31 And given that demes observed similar political procedures, it seems unlikely that the number of inscriptions originally produced by demes in the classical period varied a great deal either.32
As for the literary evidence, although we might think that some genres (such as legal oratory) would be more likely to preserve mentions of city-dwellers, others (such as comedy) seem to have a bias towards rural demes.33 In the end, opposite tendencies in different types of writing probably cancel each other out.
So though our evidence as a whole contains some possibilities of bias that are worth bearing in mind as we move to our results, these are not very significant. We should now turn to our second major methodological issue.
This arises from the fact that many Athenians moved from deme to deme during their lives. This is now well established, thanks mainly to studies of funerary inscriptions, though there continues to be controversy about how we should characterise this migration.34 Our spreadsheet lists the home demes of wealthy Athenians; but these wealthy Athenians may well have moved out of their home deme at some point in their lives. And that is surely a problem if our main aim is to illustrate the geographical distribution of wealthy Athenians.
Now, some number of citizens probably did stay in their home deme all (or most) of their lives. But it seems a fair bet that many wealthy individuals spent a significant amount of time in the urban core. Strictly speaking, however, we are not investigating the geographical distribution of wealthy men in Attica, but their social distribution. In other words, we are asking not which demes rich Athenian were more likely to live in, but which demes rich Athenians were more likely to come from and be members of.
A citizen remained a member of the same deme his whole life, no matter how long he lived in other parts of Attica. Demes were one of the primary sites for socialisation, and bonds between demesmen were among the strongest an Athenian citizen forged outside of his family.35 In view of this, it would not be too surprising if some demes gave their members a significantly higher chance of being wealthy. Our work is intended to examine this very real possibility.
Our final methodological issue concerns chronology. Davies’ work covers the period 600–300 BCE. We think our results are most interesting when they are taken as relevant to questions concerning the nature of Athenian democracy. Our parameters for Athenian democracy are 508/7 and 322 BCE.36 So our evidence comes from a longer span of time than the period we are most interested in.
This is a concern, but not a fatal blow to our project. This is because Davies’ evidence is drawn mainly from within the classical period. Davies himself produced a breakdown of liturgists by generation, and of the 779 men in his sample, 559 or 72 per cent come from between 501 and 332.37 (It is likely that even more come from between 508/7 and 321, but the relevant generations that Davies uses do not coincide perfectly with our start- and end-dates for the democracy.)38 So we think our results can be taken, with a few caveats, to apply to the classical Athenian democracy.
4 RESULTS AND ANALYSIS
Our main task in this chapter is to examine the relationship between the total number of citizens and the number of wealthy citizens in the various demes. At this point we need to admit that nobody really knows how many citizens lived in the Attic demes. Because of this, we have had to make use of a proxy. Our proxy is the demes’ so-called bouleutic quota, the number of citizens that each deme sent to the central Council (boulê).
We assume that this number reflected a deme’s population – and that it did so in some constant way across all the demes, throughout the classical period. We do not need to worry about precisely how many citizens a deme must have had per councillor. As long as that number stayed roughly the same for most of the demes for most of our period, the bouleutic quota will provide a usable proxy for population.39
In what follows, we will almost always be seeking to determine whether there is a correlation between the number of wealthy citizens in the demes and their bouleutic quotas (that is, their citizen populations). We will also be looking out for exceptional demes – those which have many more rich men than we might expect from their populations. We will begin with the distribution of wealthy citizens as a whole, before breaking down the results according to the various categories of evidence.
Our most important category of evidence is the most comprehensive: the total number of names in Davies. In Figure 13.1, we have plotted the total number of men in Athens’ upper class in the various demes (according to Davies) against the demes’ bouleutic quota (a proxy for population). The result suggests a positive correlation between the two variables (r = 0.77). This implies that the more populous a deme was, the more liturgists it contained. That this model explains the data reasonably well is indicated by the coefficient of determination (r2 = 0.59). Most of the variance in the number of liturgists in the demes (at least, the ones Davies knew of) can be explained by the variance in the populations of the demes (or at least, their bouleutic quotas). And this result is extremely robust, there being less than one in a thousand chance that it was produced randomly (p = 0.000).40
A few of the individual demes, such as the super-large deme of Acharnai, in the top right-hand corner of Figure 13.1, are right where we would expect them to be if population and the number of super-rich were tightly linked. At the same time, we can see that there are a number of significant outliers, especially on the left-hand side of the figure. Three demes which have slightly more wealthy citizens than we might expect are Leukonoion, Kephisia and Kydathenaion.
We should note here that in earlier versions of Figure 13.1, Upper Paiania and Upper Lamptrai appeared as extreme outliers. But this was simply a result of the way Teegarden had entered the data for a small number of exceptional demes, those which fall into pairs such as Upper and Lower Paiania and Upper and Lower Lamptrai.41
Figure 13.1 Number of names vs. bouleutic quota
Teegarden simply put all the attested liturgists into the upper deme in each case.42 This produced an improbably large number of liturgists for those upper demes. We redistributed the attested liturgists among both these demes according to their bouleutic quotas, and when we had done so the upper demes no longer appeared as outliers.43
Figure 13.2 Normalised names heat map (readers should ignore the minus figures, which are an artefact of the heat map)
We can also look at the number of wealthy men per capita in each deme in a more direct way, by simply dividing the number of liturgists each deme had by its quota of councillors. This gives us a figure that tells us how many liturgists each deme had per councillor. The higher the figure, the more likely it was for a deme’s citizens to be found in the top wealth class. Since this figure is simply the number of names in Davies per councillor, we are referring to it as ‘normalised names’. This will be a crucial number for us in the next two figures.
In Figure 13.2, we drew a heat map of Attica, in which demes that have a high number of normalised names are shown as red-hot, and demes which have a low number of normalised names are cold and blue. The higher the number of liturgists per councillor, the brighter the deme. This map suggests once again that, on the whole, there is not a huge deal of variation within Attica in terms of how much opportunity a citizen had to become super-rich.
It also suggests once again that there are some significant outliers. These include Amphitrope, Angele, Kydantidai, Upper and Lower Pergase and Cholleidai, and Lakiadai. But the map also enables us to examine whether there is any sort of geographical pattern to the exceptions: are the hotspots of economic opportunity clustered in a particular part of Attica? The answer would appear to be no. In particular, this way of presenting the data does not appear to back up Osborne’s calculation that the Mesogaia has an unusually high number of liturgists per citizen.
Figure 13.3 Number of demes vs. normalised names
Figure 13.3 is a histogram that plots the number of names in Davies per councillor against the number of demes that have that number of normalised names. It could be described as a bar graph which shows how many demes have a given number of liturgists per councillor. So we can see, for example, that there are no fewer than 30 demes which have between 1 and 1.5 rich men per councillor, and, curiously enough, that there are no demes at all that have between 5 and 6 rich men per councillor.
Once again, this graph suggests that the variance between demes is not enormous: 95 of the 139 demes, that is, 68 per cent of them, have a number of normalised names of between 0.5 and 3. And once again, we can see that there is a small number of demes that have an unusually high figure for normalised names, including Leukonoion and Kydantidai.
Figure 13.4 Liturgists vs. bouleutic quota
Now that we have looked at the overall picture of the spread of the Athenian upper class across Attica, we can examine individual categories of evidence one by one. Figure 13.4 shows the number of festival liturgists that are attested per deme against the demes’ bouleutic quotas. Here we can see immediately that the positive correlation is very moderate (r = 0.53) and that a deme’s bouleutic quota is only a poor predictor of its number of recorded festival liturgists (r2 = 0.28). Once again, the result easily passes the standard tests of statistical significance (p = 0.000).
Unsurprisingly there are quite a few outliers, including some rather striking ones. The two most striking of all are Kydathenaion and Lower Paiania, at the top centre of the graph, with no fewer than 7 recorded festival liturgists each. It occurred to us that the high number of festival liturgists from Kydathenaion might have something to do with one of the most famous members of that deme, the comic playwright Aristophanes.44
Aristophanes might have led to a high number of recorded festival liturgists in his home deme, we supposed, because the transmission of his plays led to the survival of a number of associated texts which identified his chorêgoi. And these men were more likely to have come from the same deme as Aristophanes.
As it turned out, this supposition was incorrect. Only one of the seven festival liturgists from Kydathenaion listed by Davies is known to have served as a chorêgos for comedy – and that was in 447/6, before Aristophanes was active as a playwright. But the high number of recorded festival liturgists from Kydathenaion can be explained by a different circumstance.
The evidence for the chorêgiai of six out of our seven festival liturgists from Kydathenaion comes from just two inscriptions, both records of victors in dramatic and musical competitions from the tribe Pandionis (of which Kydathenaion was part).45 One of these inscriptions, it turns out, also accounts for four of the seven recorded festival liturgists from Lower Paiania (which was also in the tribe Pandionis).46
These demes are outliers, it would seem, mainly because of the accidental survival of a single list of victorious chorêgiai.47 (To our knowledge, no comparable inscription exists for other tribes.)48 If this inscription had not survived, these two outliers might not have been outliers, and we might have found a closer correlation between festival liturgists and deme populations. At the same time, the correlation would still have been relatively low, especially when compared to the relationship between trierarchs and deme populations.
This relationship is illustrated in Figure 13.5. It is immediately apparent that the correlation in this case is closer (r = 0.75), and that a deme’s population is a better predictor of its number of attested trierarchs (r2 = 0.57). In fact, the positive relationship here is very nearly as strong as for the category of attested liturgists as a whole (Figure 13.1); and the degree of statistical significance is just as high (p = 0.000). As ever, there are a few outliers: Lower Paiania is the most spectacular (with 21 trierarchs to its 11 councillors), but Anagyrous, Kephisia and Erchia also have more trierarchs than we might expect considering their populations (over a dozen each, despite having only half a dozen councillors each).
Figure 13.5 Trierarchs vs. bouleutic quota
What might account for the contrast between Figures 13.4 and 13.5? Why does the correlation between trierarchs and population appear to be so much closer than the correlation between festival liturgists and population? In trying to answer this question, it makes sense to look at the differences between the two main types of liturgy. (Though part of the difference may be accounted for by the higher number of data points we have in the case of the trierarchy: 490 of our 687 total instances, or 71 per cent, versus 119 instances, or 17 per cent, in the case of the festival liturgies.)
Significant differences between the two main types of liturgy certainly existed. The trierarchy appears to have been more expensive: from the speeches of Lysias and Demosthenes, we have figures for the cost of trierarchies ranging from 2,700 to 6,000 drachmas, whereas the most expensive chorêgia we hear of cost 3,000 drachmas (and most of the other festival liturgies were considerably cheaper).49 Festival liturgies – especially the chorēgia – may also have had a more aristocratic profile, partly because of the trierarchy’s association with the largely lower-class navy.50
In view of this, we might conclude that the distribution of very wealthy liturgists (those who served as trierarchs) was more in line with population than was the distribution of moderately wealthy festival liturgists. Or that Athenians who liked to think of themselves as aristocrats were less evenly spread around Attica than Athenians who were simply exceptionally wealthy.
However, there are a couple more differences between the two main types of liturgy which we think offer a better explanation for the patterns in Figures 13.4 and 13.5. These concern the way festival and military liturgists were selected, and the type of evidence about them that was preserved.
Crucially, the trierarchy seems to have been obligatory, whereas the festival liturgies were, strictly speaking, voluntary.51 Citizens who volunteered for festival liturgies (or agreed to serve after being nominated) could, of course, reasonably expect to win a great deal of honour from their fellow citizens – the kind of honour that, alongside its other benefits, might one day help their case in court. And this meant that there were considerable incentives for the rich to serve as festival liturgists. But it seems clear from several arguments in the orators that not serving as a festival liturgist was always an option, even for those who could afford it.52
By contrast, those liable to trierarchies appear to have been compelled to serve them by Athens’ military apparatus. Because of this, while ‘lists of trierarchs certainly existed . . . no such lists were kept of those liable to agonistic liturgies’.53 The evidence we have of those who paid festival liturgies tends to come in the form of victory monuments, erected either by the victorious chorêgoi themselves or by their deme or tribe.54
There are two obvious features of choregic victory monuments that it is important to highlight here. First, monuments that honour victorious chorêgoi do not provide us with an exhaustive list of chorêgoi (or even a random sample); they only honour those chorêgoi who happen to have been victorious.55 Second, victorious chorêgoi were usually honoured with their knowledge and consent (and very often with their active support).56
How this impacts the interpretation of our Figures 13.4 and 13.5 will immediately be clear. The festival liturgies were a matter of voluntary competition for honour, and our knowledge about those who paid them consists overwhelmingly of those who did well at them. Because of this, Figure 13.4 does not really describe the relationship between population and wealth. Instead, it describes the relationship between population on the one hand, and some combination of public-spiritedness, desire for glory and dramatic success on the other.
For our purposes, it is less troubling if those qualities are unevenly distributed. If some demes happen to have nourished a disproportionately large number of wealthy citizens who succeeded in distinguishing themselves in the city’s festivals, that need not say anything negative about the distribution of economic opportunity in Athens. It may, instead, simply reflect a perfectly natural variation in citizens’ eagerness to participate in public festivals.
We now turn to the diadikasia inscriptions. The diadikasia was a type of legal procedure concerned with ‘claims either to enjoy a privilege or to be exempt from an obligation’.57 Diadikasiai could be employed in cases of disputed inheritance, but (more relevantly here) were also used by men who wanted to avoid paying certain taxes. The main types of diadikasia appealed to in such cases were the skêpsis (‘excuse’) in which a citizen claimed he should be exempt; and the antidosis (‘exchange’) in which he argued that another citizen was richer than he was and should pay the tax in his place.58
The inscriptions used by Davies in compiling his register are all from around 380, and concern a body of men probably somewhat larger than the usual liturgical class of 300–400 men.59 Since the diadikasia was designed to give a final ruling on men whose status as elite tax-payers was disputed, it seems probable that the men whose names appear on these inscriptions can be described as part of the larger wealth elite.
Figure 13.6 shows the number of names attested on the diadikasia inscriptions against our usual proxy for population, bouleutic quota. Here the weakness of the correlation (r = 0.26) is quite striking, as is the powerlessness of the bouleutic quota in accounting for the number of liturgists (r2 = 0.07). Along with a single spectacular outlier (Kedoi), there are three moderate outliers (Athmonon, Kephisia and Euonymon). Although three of these demes are from the tribe Erechtheis, this is not simply an effect of the way the names are arranged on the inscriptions; they do not seem to be listed by tribe at any point.
There is a number of reasons to be especially cautious about this category of evidence. As we have seen, the men involved in the diadikasiai mentioned here were very probably more numerous than the typical set of liturgy-payers. As we have also seen, the men whose names appeared on these inscriptions had their status as payers of an elite tax disputed. Even though most of them will have possessed great wealth relative to most Athenians, some of these men were not really one-percenters.60
Figure 13.6 Names on diadikasia inscriptions vs. bouleutic quota
But these discrepancies fail to explain the weakness of the correlation in this case. Even if this category included many men who would not otherwise have made the cut for Davies’ ‘upper class’, this would not in itself explain the different pattern visible in Figure 13.6. Why would including slightly more – and slightly less wealthy – wealthy citizens produce such a markedly weaker correlation? There is no reason to believe that those on the edge of the liturgical class were distributed more unevenly than those whose place within it was secure.
Another reason for caution may be more significant. This is the low number of data points in this category of evidence. We have the names of only 61 citizens on the diadikasia inscriptions, less than 9 per cent of the 687 wealthy citizens who were included in Ober and Teegarden’s spreadsheet. (And in fact, this result fails the conventional test for statistical significance: p = 0.109.) The small size of the sample in itself makes it unlikely that any pattern observed in it will be representative of the distribution in the broader population, or indeed of the distribution in a larger sample – a principle that the contrast between our Figures 13.1 and 13.6 seems to instantiate.
The divergence of this case from the general pattern, though, when combined with a peculiarity of this category of evidence, may be of further interest. What the diadikasia inscriptions offer us that no other category of evidence does is a snapshot of a single year. Because of this, though these results might not tell us much about the distribution of liturgists throughout the classical period as a whole, they might tell us something about the annual variability in this distribution. In other words, is the pattern of distribution in one year, maybe 380, typical of the overall pattern?
We have just gone through a number of reasons to be cautious about these particular results, and we should not abandon them now. At the same time, although this sample is almost uselessly small for questions about all those who served as liturgists in Athens through the classical period, it might not be too small to tell us something about the smaller set of elite tax-payers from around 380. That group, as we have seen, was larger than the usual 300–400 liturgists; still, 61 names out of no more than 1,000 citizens who underwent diadikasiai around this time is not a disaster.
What might this single slice of evidence tell us about the interannual variation in the distribution of liturgists in Attica? If we can trust our Figure 13.6, it would seem that the pattern for 380 was very different to the overall pattern for the classical period. This, of course, would not be surprising. On the contrary, what would be surprising is if our annual slice of evidence reproduced the relatively close correlation that we have observed for the classical period as a whole.
This would be surprising because such an effect – with an appropriate number of liturgists being drawn every year from the appropriate demes – would be extremely unlikely to occur simply by chance. If it did occur, this would suggest that it was artificial; in more precise terms, that there was some mechanism that fixed the number of liturgists from each of the demes to their populations, similar to the way that the bouleutic quotas fixed the number of councillors drawn from the demes to their populations.
The pattern for the names on the diadikasia inscriptions would seem to exclude this possibility. The overall correlation between the demes’ populations and their numbers of wealthy citizens must, then, have been a result not of formal planning but of an egalitarian distribution of economic opportunities across the demes of Attica. (We regret to say that the question of how exactly this situation emerged and was maintained is outside the scope of this chapter.)
Figure 13.7 Other wealthy vs. bouleutic quota
We can now turn to our final category of evidence. This includes citizens who, though they are not directly attested as liturgists, and though they do not appear on the diadikasia inscriptions, Davies nonetheless included in his register for some other reason. It is immediately clear from Figure 13.7 that the correlation between the number of citizens in this category and bouleutic quota is again low (r = 0.2), and that population does virtually nothing to explain the incidence of these citizens in the demes (r2 = 0.04). The spectacular outlier in this case is Alopeke, and the most eye-catching of the other outliers are Lakiadai, Leukonoion and Kydathenaion.
As with the diadikasia inscriptions, the weakness of the correlation here may be largely due simply to the small number of cases in this category. At just over 10 per cent of the names in the spreadsheet (71 out of 687), this category is only slightly larger than the previous one, and this result too fails the conventional test for statistical significance (by a lot: p = 0.216). It should be remembered that the data points in this category, though weakly correlated when looked at on their own, contribute towards the strong correlation we found when looking at the totality of names in Davies’ register.
At the same time, it might be useful to consider whether the nature of the evidence that Davies drew upon in compiling this category might be linked to the relationship (or lack of a relationship) we can observe in Figure 13.7. Ober and Teegarden’s ‘Other Wealthy’ category, we recall, was drawn from three of Davies’ categories of evidence: (1) pentakosiomedimnoi, (2) men with estates capable of raising horses, and (3) men who we have some other reason to believe were extremely wealthy.
We might suspect that a large proportion of the men who qualified for Davies’ register under categories (1) or (2) lived in the earlier rather than the later part of the classical period. In the case of the pentakosiomedimnoi, this is because the Solonian census classes, though a live issue in the middle decades of the fifth century, had effectively ceased to matter by the later decades of the fourth.61 As for citizens who raised horses, much of Davies’ evidence for them seems to come from epinician poetry celebrating victors in the equestrian events at the Panhellenic games. And epinician poetry, as a genre, is slanted towards the first half of the fifth century.
It is also worth reflecting on another peculiarity of this category of evidence: that it consists largely of men we know about through their significant landholdings. Of course, we assume that the majority of liturgists throughout the classical period owned significant amounts of land.62 All the same, we know about most of the liturgists in Davies’ register because of reports about their payment of liturgies. In the case of this sub-category, on the other hand, we know about the liturgists because of indications concerning the size of their estates. Interestingly, all of our outliers are urban demes that are outside the city walls (with the exception of Kydathenaion, which is inside the city walls) – perhaps the perfect location for one’s estates to gain the attention of large numbers of Athenians.
We might, therefore, see Figure 13.7 as more heavily influenced than our other graphs by earlier liturgists, and by liturgists that we know about specifically because of their landholdings. But we should be very cautious with this category of evidence (as with the previous one), because the smallness of the sample makes any conclusions we might draw unreliable in the extreme.
5 CONCLUSIONS
The main aim of this chapter was to examine the relationship between the populations of the demes and their numbers of attested liturgists. Our most important conclusion is that there was a strong positive correlation between these two variables. We can conclude that an ancient Athenian’s deme affiliation is not a good predictor of whether or not he was wealthy. In this limited sense, there was an even spread of economic opportunity across the network of demes. Granted, we did not find a very strong positive correlation between population and liturgists, but that was hardly to be expected.
Still, that the correlation between wealthy citizens and populations in the demes is not very high might have something to tell us, especially when combined with the weak correlation we found for those who had undergone the diadikasia procedure in 380. As we have seen, this might suggest that the system was not designed to produce a correlation; there was no attempt to fix the number of liturgists for each deme. The egalitarian pattern we observe, then, is a ‘natural’ one – or, at least, it is a pattern that was created by something other than conscious oversight of how many liturgists came from each deme.
The ‘natural’ pattern we see is one in which there are moderately strong correlations, but no very strong ones. This brings us to the question of outliers. We identified several significant outliers in each of our categories of analysis. Do they have anything in common which might tell us something further about the distribution of economic opportunity in Attica?
Let us first recall which demes were outliers in our various categories of analysis. For Davies’ register as a whole, the outliers were Leukonoion, Kephisia and Kydathenaion. Our heat map threw up Amphitrope, Angele, Kydantidai, Upper and Lower Pergase and Cholleidai, and Lakiadai as outliers; and our histogram threw up Leukonoion and Kydantidai. Kydathenaion and Lower Paiania stood out in our analysis of festival liturgists, as did Lower Paiania, Anagyrous, Kephisia and Erchia in our discussion of trierarchs. Finally, Kedoi, Athmonon, Kephisia and Euonymon seemed untypical when it came to the diadikasia inscriptions, and Alopeke, Lakiadai, Leukonoion, and Kydathenaion when it came to the category of ‘Other Wealthy’ citizens.
Nothing in particular seems to unite these demes. But a few names appear several times in this list, and it might be worthwhile to look at these as a set. Kephisia, Leukonoion and Kydathenaion appear three times, and Kydantidai and Lakiadai appear twice. What do they have in common? Not much, though three of the five are urban demes and none are coastal. And if we exclude the intra-mural deme of Kydathenaion, the remaining demes cluster in two particular areas of Attica: just to the south of Mount Pentele (Kephisia and Kydantidai) and just to the east of Mount Aigaleos (Leukonoion and Lakiadai). This might modify Osborne’s picture of what privilege there was in Attica centring on the Mesogaia (east of Hymettus).63
If we had to point towards the most privileged part of Attica in terms of economic opportunity, in fact, we might opt for the area between the three mountains of Aigaleos, Pentele and Hymettos. But we would need to test this hypothesis statistically, which would involve putting the demes in that area in a single, artificial basket (even though the set would include both urban and inland demes).
It should also be said that our method of identifying outliers and seeing what they might tell us is imperfect. To really get to the bottom of what types of demes had an advantage (if any), we would need to consider whether there is a correlation between the number of normalised names and various further categories such as tribes and trittyes, as well as various indicators for location. But our preliminary explorations on this front do not suggest that tribes and trittyes, for instance, have a strong effect on a deme’s number of wealthy citizens.
Another significant finding of this chapter concerns the differences between various categories of evidence. First of all, there are differences in the distribution of attested wealthy citizens among these different sub-categories of evidence, and the patterns for individual sub-categories sometimes differ a great deal from the pattern for the whole set of wealthy Athenians.
Some of these differences, no doubt, are almost entirely a result of the small number of cases in some of the categories of evidence (and here we are referring to the diadikasia inscriptions and the ‘Other Wealthy’ category). But in the case of the directly attested liturgists, the difference in our figures seemed to reflect an interesting difference between the two main types of liturgy.
Festival liturgies were voluntary whereas trierarchies were mandatory; and whereas the names of festival liturgists tended to survive because they had been victorious, trierarchs could have their names recorded simply for fulfilling their obligations. The uneven distribution of festival liturgists, we concluded, reflects not an unequal distribution of opportunities to attain moderate wealth, but an unequal distribution of zeal for, and success in, various roles in public festivals. And an unequal distribution of these qualities was, we suggested, unproblematic in terms of the sort of equality that we tend to require of democracies.
ACKNOWLEDGEMENTS
A number of people helped us with this chapter. We are especially grateful to Sydney Shep, Josh Ober, Dave Teegarden, John Davies, John Shaver, Alain Bresson, Claire Taylor, Matt Simonton and Matthew Trundle.
BIBLIOGRAPHY
Anderson, G. (2003), The Athenian Experiment: Building an Imagined Political Community in Ancient Attica, 508–490 B.C., Ann Arbor.
Bourriot, F. (1976), Recherches sur la nature du génos: Etude d’histoire sociale Athénienne. Périodes archaïque et classique, 2 vols, Lille.
Bresson, A. (1988), ‘Richesse et pouvoir à Lindos à l’époque hellénistique’, in S. Dietz and I. Papachristodoulou (eds), Archaeology in the Dodecanese, Copenhagen: 145–54.
Canevaro, M. (2016), Demostene, Contro Leptine: Introduzione, traduzione e commento storico, Berlin.
Damsgaard-Madsen, A. (1988), ‘Attic funeral inscriptions: Their use as historical sources and some preliminary results’, in A. Damsgaard-Madsen and E. Hallager (eds), Studies in Ancient History and Numismatics Presented to Rudi Thomsen, Aarhus: 55–68.
Davies, J. K. (1967), ‘Demosthenes on liturgies: A note’, Journal of Hellenic Studies 87: 33–40.
Davies, J. K. (1971), Athenian Propertied Families, 600–300 BC, Oxford.
Davies, J. K. (1981), Wealth and the Power of Wealth in Classical Athens, Salem.
Gabrielsen, V. (1987), ‘The diadikasia-documents’, Classica et Mediaevalia 38: 39–51.
Gabrielsen, V. (1994), Financing the Athenian Fleet: Public Taxation and Social Relations, Aarhus.
Hansen, M. H. (1982), ‘Demographic reflections on the number of Athenian citizens 451–309’, American Journal of Ancient History 7: 172–89.
Hansen, M. H. (1983), ‘Rhetores and strategoi in fourth century Athens’, Greek, Roman, and Byzantine Studies 24: 151–80.
Hansen, M. H. (1985), Demography and Democracy: The Number of Athenian Citizens in the Fourth Century BC, Herning.
Hansen, M. H. (1991), The Athenian Democracy in the Age of Demosthenes, Oxford.
Hansen, M. H. (2006), Studies in the Population of Aigina, Athens and Eretria, Copenhagen.
Hansen, M. H., L. Bjertrup, T. H. Nielsen, L. Rubinstein and T. Vestergaard (1990), ‘The demography of the Attic demes: The evidence of the sepulchral inscriptions’, Analecta Romana Instituti Danici 19: 24–44.
Jones, N. F. (1999), The Associations of Classical Athens: The Response to Democracy, Oxford.
Kaiser, B. A. (2007), ‘The Athenian trierarchy: Mechanism design for the private provision of public goods’, Journal of Economic History 67: 445–80.
Kellogg, D. L. (2013), Marathon Fighters and Men of Maple: Ancient Acharnai, Oxford.
Lewis, D. M. (1955), ‘Notes on Attic inscriptions (II): XXIII. Who was Lysistrata?’, The Annual of the British School at Athens 50: 1–36.
MacDowell, D. M. (1978), The Law in Classical Athens, Ithaca.
Mossé, C. (1979), ‘Les symmories athéniennes’, in H. van Effenterre (ed.), Points de vue sur la fiscalité antique., Paris: 31–42.
Ober, J. (1989), Mass and Elite in Democratic Athens: Rhetoric, Ideology, and the Power of the People, Princeton.
Ober, J. (2000), ‘Review of Wilson 2000’, Bryn Mawr Classical Review 2000.10.02.
Ober, J. (2008), Democracy and Knowledge: Innovation and Learning in Classical Athens, Princeton.
Ober, J. (2015), The Rise and Fall of Classical Greece, Princeton. Ober, J. and D. A.
Teegarden (2006), ‘Deme data’, unpublished spreadsheet.
Osborne, R. G. (1985), Dêmos: The Discovery of Classical Attika, Cambridge.
Osborne, R. G. (1992), ‘Is it a farm? The definition of agricultural sites and settlements in ancient Greece’, in B. Wells (ed.), Agriculture in Ancient Greece, Stockholm: 21–5.
Ostwald, M. (1986), From Popular Sovereignty to the Sovereignty of Law: Law, Society and Politics in Fifth-Century Athens, Berkeley.
Raaflaub, K. A. (2007), ‘The breakthrough of dêmokratia in mid-fifth-century Athens’, in K. Raaflaub, J. Ober and R. Wallace (eds), Origins of Democracy in Ancient Greece, Berkeley: 105–54.
Ruschenbusch, E. (1978), ‘Die athenischen Symmorien des 4. Jh. v. Chr.’, Zeitschrift für Papyrologie und Epigraphik 31: 275–84.
Sinclair, R. K. (1988), Democracy and Participation in Athens, Cambridge.
Taylor, C. (2006), Participation in Athenian Democracy, PhD dissertation, University of Cambridge.
Taylor, C. (2011), ‘Migration and the demes of Attica’, in C. Holleran and A. Pudsey (eds), Demography and the Graeco-Roman World: New Insights and Approaches, Cambridge: 117–34.
Teegarden, D. A. (2006), ‘Haec notanda sunt’, unpublished notes on J. Ober and D. A. Teegarden, ‘Deme data’, unpublished spreadsheet.
Todd, S. C. (1993), The Shape of Athenian Law, Oxford.
Whitehead, D. (1986), The Demes of Attica: A Political and Social Study, Princeton.
Wilson, P. (2000), The Athenian Institution of the Khoregia: The Chorus, the City, and the Stage, Cambridge.
1 Though not necessarily of becoming wealthy; a few wealthy families may have monopolised wealth in their demes generation after generation, without allowing others to break into the top wealth class. Our work cannot exclude this; what we can show is whether some demes had more wealthy individuals than their populations merited. So when we ask whether economic opportunity was equably distributed, we are looking at whether two random citizens from different demes would have had a similar chance of being rich; not whether two citizens of the same deme would have had a similar chance of being rich.
2 Bresson 1988.
3 Osborne 1985: 45, 196–200, table 2 (a). He relies on Davies 1971.
4 Taylor 2006: 148. Taylor is looking at the relationship between wealth and politcal power, not wealth and population. Hence the number of councillors for her is simply one part of the total number of known officials. We have a different purpose for the number of councillors: as a proxy for population.
5 For the details of how we did this, see section 4.
6 Osborne 1985: 45.
7 Osborne 1985: 45.
8 Osborne 1985: 46.
9 Davies 1971; Ober and Teegarden 2006.
10 Davies 1971: xx.
11 Davies 1971: xx–xxi.
12 Davies 1971: xxiii–xxiv; 1981: 15–34. Few now doubt that there was a group of 300–400 super-rich citizens who paid liturgies. But scholars who follow Davies (e.g. Sinclair 1988: 64; Ober 1989: 128) think that liturgies were paid only by this elite group of super-rich men, others (Ruschenbusch 1978; Mossé 1979; Hansen 1991: 112–15) think that a wider group of 1,200–2,000 citizens (sometimes called ‘the leisure class’) also paid some liturgies.
13 Thirty thousand Athenian citizens in the fourth century: Hansen 1982; 1985; 2006.
14 Isocr. 15.145; Dem. 14.16; Hypereides F 134; with Davies 1981: 19.
15 Davies 1981: 24.
16 Davies 1971: xx.
17 On the trierarchy, see esp. Gabrielsen 1994; Kaiser 2007.
18 On the chorēgia, see esp. Wilson 2000. For other types of festival liturgies, see Davies 1971: xxi. For the number of festival liturgies in the fourth century (from around 97 in a typical year, up to 117 in a year of the Great Panathenaia festival), Davies 1967.
19 Davies 1971: xxv–xxvii.
20 Ober and Teegarden 2006.
21 Davies 1971: xxx: ‘The prosopographical survey of this class which follows in the Register is not to be taken as adequately representing this class, let alone as anything even remotely approaching a complete roster of upper-class Athenians.’
22 Davies 1971: xxviii–xxix compares the number of known liturgists with the number of actual liturgists in cases where the second number is available. Of the 800–1,200 estimated trierarchs of the Peloponnesian War, we know the names of 30 (3 per cent); of the 1,800 estimated trierarchs of 357 to 340, we can name 149 (8 per cent); of the 400 estimated trierarchs of 333 to 321, 158 are known by name (40 per cent); of the 3,300 estimated festival liturgists in each generation of the fourth century, we can identify between 21 and 30 (less than 1 per cent).
23 The kinds of bias that Davies himself was concerned about (1971: xxvii -xxviii) do not greatly affect our analysis. So he may be right that ‘speeches tell us less about the placid than about the litigious and the grasping’, but this is not the sort of bias that would lead to certain areas of Attica being over-represented.
24 On the antidosis procedure, see MacDowell 1978: 162–4; Todd 1993: 120. On the incentives it may have generated, Kaiser 2007.
25 For fuller discussion and references relating to diadikasia, see the next section.
26 E.g. Apollodoros the son of Pasion of Acharnai (for whom see Davies 1971: no. 11672) owned land in three different demes: [Dem.] 50.8.
27 The Solonian census classes are the subject of continuing controversy (see e.g. Raaflaub 2007: 128–36). But it seems reasonable to suppose that if they were assessed at all, they were assessed by the central state.
28 The conventional view is that most of Attica was incorporated into Athens in the Peisistratid period (561–510); see e.g. Ostwald 1986: 3–28. For the alternative view, see esp. Anderson 2003.
29 Davies’ description of this category is limited to a single sentence: ‘(iv) Lastly, there are a few men who are not attested as λῃτουργοῦντες, but who are known to have owned property of a value well in excess of 3 tal., or whose economic transactions presuppose such a property level’ (1971: xxvii).
30 See Davies 1971: passim.
31 Jones 1999: 100–7.
32 For political procedures in the demes, see Whitehead 1986: 86–148.
33 For the demes in comedy, see Jones 1999: 107–15.
34 Damsgaard-Madsen 1988 leads Jones 1999: 58 to point confidently to ‘overall trends’. For scepticism, see Taylor 2011, with Kellogg 2013: 51–5.
35 Whitehead 1986: 223–57.
36 In this we follow Ober (e.g. 2008: 30).
37 Davies 1971: xxvii.
38 If we divide the number of men in each of the relevant generations by the number of years in that generation, and then multiply the result by the desired number of years, we get 72 more liturgists. Adding this estimate to our total for the classical period gives us 631 liturgists, 81 per cent of the men in Davies’ register. Note that while there are 779 men in Davies’ register, there are only 687 in Ober and Teegarden’s spreadsheet; the difference seems to be due to the fact that we do not know the deme affiliation of every citizen who is known to have been wealthy.
39 Osborne 1985: 42–6 and Kellogg 2013: 36–4 both integrate this assumption into their analyses, along with a precise figure (65) for the number of citizens per councillor. At the same time, both note that the bouleutic quota should only be used to provide a ‘minimal figure’ (Osborne 1985: 43, quoted by Kellogg 2013: 40 n. 15) and that deme populations must have outgrown their quotas by the time of the reapportionment of bouleutic quotas of 307/6. All that is necessary for our analysis is that the ratio between populations and quotas did not vary much across demes before the end of the classical period. Hansen et al. (1990: 30) appear to be alone in supposing that bouleutic quotas changed in the classical period.
40 All p-values were calculated for us by John Shaver using SPSS. The programme calculates values to three decimal places only; hence values smaller than one thousandth are shown as ‘0.000’ or ‘< 0.001’.
41 On these ‘divided demes’ see esp. Jones 1999: 61–4. It looks as though such cases arose when Kleisthenes (for whatever reason) decided to divide a pre-existing village into two separate demes, after which ‘citizens of the two demes were not regularly distinguished as hailing from one or the other, rather both groups were simply called “Lamptrians”’ (Jones 1999: 62). Because attested liturgists appear in the sources simply as e.g. Ἀριστοκρ[άτης- – -]/ Λαμπτρε[ύ]ς (IG II2 1928, ll. 30–1), Davies lists such men under the heading Λαμπτρεῖς, without distinguishing between the two parts of the village. But since the Kleisthenic system treated the two sections of Lamptrai as two separate demes, we also tried to treat them as such.
42 A pragmatic move that he helpfully records in his appendix to the spreadsheet (Teegarden 2006).
43 More fully, we divided the number of liturgists in each category by the total bouleutic quota of both the twinned demes. Then we multiplied the result by each individual twinned deme’s bouleutic quota, giving us a better estimate of that deme’s number of liturgists.
44 Ἀριστοφάνης ὁ κωμωιδοποιὸς…τῶν δήμων Κυδαθηναιεύς (Anonymous, Life of Aristophanes = PCG T1, 1).
45 IG II2 1138 and 2318. We were helped here by Davies 1971: 602–24, index 1, a breakdown of liturgists by deme which also notes which kind of liturgy each man is known to have paid. We looked up the main entry in Davies’ register for every festival liturgist from Kydathenaion and Paiania, and noted down the source for each man’s festival liturgy.
46 IG II2 1138.
47 IG II2 2318 is, in fact, a fragment of 1138; it was identified as such by Lewis 1955: 17–24.
48 None is mentioned in the discussions of phyletic monuments of Jones 1999: 187–8 or Wilson 2000: 171–2.
49 For a full list of the evidence, see Davies 1971: xxi–xxii.
50 Wilson 2000: 109–43, esp. 111–16. Ober 2000 objects that the trierarchy was more costly, and more important to the Athenian state due to its military role. But we are not sure this affects Wilson’s claims about the aristocratic cultural profile of the chorêgia.
51 Davies 1981: 24–8.
52 E.g. Aeschin. 1.97; Dem. 21.154 with Davies 1981: 25.
53 Davies 1981: 25. Wilson 2000: 53 agrees that there were no lists of festival liturgists, though he notes the scepticism of Gabrielsen 1994 about the existence of lists of trierarchs.
54 On choregic monuments, see Wilson 2000: 198–264.
55 One such list is IG II2 1138, in which the tribe Pandionis, after resolving to honour one Nikias the son of Epigenes of Kydathenaion (Davies 1981: no. 10807), also decides to ἀναγράψαι δὲ καὶ εἴ τις ἄλλος νενίκηκεν/ ἀπ’ Εὐκλείδο ἄρχοντος (ll. 9–10) – but not unsuccessful chorēgoi.
56 On inscriptions on some of the monuments, the chorēgos states explicitly that he is responsible for the dedication, in the third person (IG II2 3055, l. 1, Νι[κ]ίας Νικοδήμου Ξυ[π]εταιὼν ἀνέθηκε νικήσας χορηγῶν Κεκροπίδι παίδων), or in the first (IG II2 3101, ll. 1–2, ἡδυγέλωτι χορῶι Διονύσια σ[ύ]μ ποτε ἐν[ίκων],/ μνημόσυνον δὲ θεῶι νίκης τόδε δῶρον [ἔθηκα]).
57 Todd 1993: 119.
58 For this distinction, and on the diadikasia in general, see first Todd 1993: 119–21 and the references he gives there.
59 Davies 1981: xxvi, n. 5, lists these inscriptions as IG II2 1928–32; Hesp. 7 (1938), 277, no. 12 and 306, no. 29; Hesp. 15 (1946), 160, no. 17. He notes (p. xxvi) that ‘the transactions thereon have been thought to concern the trierarchy, the proeisphora, and the chorêgia. I shall argue elsewhere (Davies 1981: 133–50) that this is probably not so, and that instead the transactions concerned membership of a body called the Thousand, which functioned as a panel of eisphora-payers until it was replaced by the symmory system in 378/7.’ Whatever the precise nature of this body of men (Gabrielsen 1987 thinks they are eisphora-payers), we follow Davies in regarding ‘the persons named on these inscriptions as members of the upper or liturgical class’ even though ‘in number they probably exceeded the normal size of the liturgical class’ (pp. xxvi–xxvii). As for the date, ‘the preserved archon-dates in the surviving fragments are 383/2, 381/0 and 380/79’ (IG II2 1930, l. 2; IG II2 1931, l. 3; Agora I 4689, ll. 2–3: Davies 1981: 134 with n. 5).
60 They seem to have numbered, in total, around a thousand (hence Davies’ preferred name for the group). Cf. also his comment that ‘their title for inclusion [in the register] is . . . slightly better than that of the 1200 members of the trierarchical panel as it functioned between 356 and 340’ (p. xxvii).
61 That the Solonian census classes were still a live issue (if they ever were) in the mid-fifth century is suggested by the opening of the archonship to the zeugitai in 457, a move that seems to be connected with the reforms of Ephialtes ([Arist.] Ath. Pol. 26.1–2). That they had no real relevance by the late fourth century is implied at 7.4 of the same text.
62 That is, that many liturgists came from the 7.5–9 per cent of citizens who owned 30–5 per cent of the land in Attica (Osborne 1992; Ober 2015: 91). In most cases, even those who made their money from other sources also owned land: e.g. Isocrates the orator (Davies 1981: no. 7716) had 100 students that paid him 1,000 drachmas each in tuition (according to Dem. 35.25, 42); but he also owned land worth around 3 talents] ([Dem.] 34.23).
63 Osborne 1985: 45.
