¹ John, J. Hector St. [sc. de Crèvecoeur], Letter from an American Farmer (Philadelphia, 1793), 170–1Google Scholar.

² A shorter version of this paper was presented at the 1991 meeting of the Association of Ancient Historians.

³ This is, roughly speaking, the period during which slavery reached its greatest extent in these states and before its rather precipitous dissolution in the aftermath of the Revolutionary War and increasing immigration from Europe. This essay will focus more closely on slavery in Connecticut, as I am most familiar with this state's history.

⁴ Cf. Osborae, R., Demos, the Discovery of Classical Attika (Cambridge, 1985), 15–63Google Scholar; idem, Classical Landscape with Figures: The Ancient Greek City and its Countryside (London, 1987), 37–40Google Scholar.

⁵ Few would now accept the picture presented by Ehrenberg and others that of Attika growing and exporting grapes and olives and feeding itself on imported grain (Ehrenberg, V., The People of Aristophanes: A Sociology of Old Attic Comedy [New York, 1962], 73–75Google Scholar). Imported grain, rather, fed primarily the non-agricultural urban population, which also relied on the surplus production of rural Attika; imported grain also helped feed the rural population in years of dearth when domestic production fell short.

⁶ Subsistence farming stands in contrast both to plantation farming and to the more modern model of farming practiced in the United States and Europe. Plantation farming also produced food needed for domestic consumption, but such production was a means to an end, that is, the disposal of specialized products in an external market. Modern farmers also grow primarily for an external market but, in contrast to plantation farmers, also meet all or most of their domestic needs by purchasing from the market. On subsistence farming in ancient Attika, see Jameson, M., “Agriculture and Slavery in Classical Athens,” Classical Journal, 73: 2 (1977), 124–6Google Scholar; on the Attic farmer's ideal of autarkeia (self sufficiency), ibid., 129; and on his limited recourse to the market, ibid., 130–1. On the continuity of small-scale farming into the fourth century, see Andreyev, V. N., “Some Aspects of Agrarian Conditions in Attic in the Fifth to Third Centuries B.C.,” Eirene, 12: 1 (1974), 18–25Google Scholar.

Because there were few large-scale producers in the northern colonies, the region's export trade in agricultural products, including supplying provisions to the Continental Army during the Revolutionary War, generally depended on accumulating the comparatively small surpluses of individual subsistence farmers. Exporters purchased these surpluses either directly from the farmers or indirectly from country storekeepers who regularly accepted produce as payment in kind for the manufactured goods they supplied to the farmers, a model which, mutandis mutatis, may help to explain how Attic farmers disposed of their individual surpluses and where at least some of Athens' exports in oil and wine came from.

⁷ As is argued, for example by Starr, C. G. in “An Overdose of Slavery,” Journal of Economic History, 18: 1 (1958), 23–24CrossRefGoogle Scholar; cf. Starr, 's The Birth of Athenian Democracy: The Assembly in the Fifth Century B.C. (New York and Oxford, 1990), 38 and 73–4, n. 28Google Scholar.

⁸ For North America, see, for example, de Tocqueville, Alexis, Democracy in America, Lawrence, G., tr. (New York, 1969), 352–3Google Scholar; for ancient Attika, see Wiedemann, T. E. J., Slavery (Greece and Rome New Surveys in the Classics, no. 19, Oxford, 1987), 25Google Scholar. He bases this assertion on the observations that grain-growing does not require year-round labor and that slaves employed in growing grain would thus be productive for only part of the year and a drain on the owner's resources for the rest (in contrast to the cultivation of other crops, such as sugar cane or grape vines, which require year-round labor). This assertion is certainly valid if we think only of plantation agriculture with its large gangs of slaves but is much less so for small subsistence farms, where one or two slaves supplement the farmer-owner's labor, which is required year-round with varying intensity for all the other chores, from milking to repairing equipment. Note also in reference to Attic agriculture, that most farmers, whether slaveholders or not, supplemented their grain growing with viticulture, which does require some year-round attention.

⁹ Wood, E. M., Peasant-Citizen and Slave: The Foundations of Athenian Democracy (London and New York, 1988)Google Scholar.

¹⁰ For example, most recently Jameson, , “Agriculture and Slavery,” 140Google Scholar; for a refutation of Jameson, see Woods, E. M., “Agricultural Slavery in Classical Athens,” American Journal of Ancient History, 8: 1 (1983), 6–15Google Scholar, although Woods's principal argument, that democracy protected small-scale Athenian farmers from exploitation by their wealthy neighbors and by the state, merely shows that these small-scale farmers did not need slaves to meet such exploitative demands. It hardly proves that they did not own slaves for other reasons (see further below, note 44).

¹¹ Cf. for classical Athens, the remark attributed to Socrates by Xenophon (Memorabilia 2.3.3) that those who can do so purchase slaves to have “fellow workers” (sunergous).

¹² This is particularly clear from probate records which inventory estates room by room.

¹³ Piersen, W. D., Black Yankees: The Development of an Afro-American Subculture in Eighteenthc Century New England (Amherst, MA, 1988), ch. 3Google Scholar.

¹⁴ Crèvecoeur had been a backwoods farmer in Chester (Orange County) New York and may perhaps have been a slaveholder himself (Allen, G. W. and Asselineau, R., St. John de Crèvecoeur: The Life of an American Farmer [New York, 1987], 40–41Google Scholar). For similar portraits of northern family slavery, see, e.g., Dwight, T., D. D., , Greenfield Hill: A Poem, in Seven Parts (New York, 1794), pt. 2, verses 193–214, pp. 36–37Google Scholar; and Cooper, J. F., Satanstoe (New York, 1845), vol. 1, 74–75Google Scholar.

¹⁵ Or they were called negroes or blacks. We shall touch on the effect of ethnic differences later in this paper.

¹⁶ Cf. Fowler, W. C., “The Historical Status of the Negro in Connecticut,” The Historical Magazine, 24 (ser. 3, vol. 3) (1874), 81Google Scholar.

¹⁷ See the discussions by Garlan, Y., “Le travail libre en Grece ancienne,” Non-Slave Labour in the Greco-Roman World, Garnsey, P., ed., Cambridge Philological Society, Suppl. 6(1980), 6–22Google Scholar; Mossé, C., “Les salariés à Athènes au IVe siècle, Dialoguesd’histoire ancienne, 2 (1976), 97–101Google Scholar; and Zimmerman, H.D., “Die freie Arbeit in Griechenland während des 5. und des 4. Jahrhunderts v.u.Z,” Klio, 56: 2 (1974), 336–52CrossRefGoogle Scholar. These authors focus primarily on non-agricultural workers; for agricultural workers, see especially Zimmerman, , “Die freie Arbeit,” 240–1Google Scholar.

¹⁸ On this point see more generally Wood, , Peasant-Citizen, 48–51Google Scholar.

¹⁹ On this last point, see further Cazzaniga, G. M., “Stratificazione sociale, rapporti di dipendenza e forme servili nel mondo antico,” Schiavitù antica e moderna: Problemi storia istituzioni, Sichirollo, L., ed. (Napoli, 1979), 154–6Google Scholar. Note also that Aristotle's discussion of slavery in Book 1 of the Politics (1235^b1ff.) is set within the framework of the household (oikia), the master-slave relationship being grouped with those of husband-wife and father-child as “parts of a household” (merê oikias).

²⁰ Politics, 1260^a 39–40.

²¹ Although census figures are inevitably only approximations, as we know from recent experience, they are nonetheless valuable as long as we do not ask of them the kind of accuracy they cannot be expected to provide. For some of the difficulties associated with colonial census figures, see Greene, L. J., The Negro in Colonial New England 1620–1776 (New York, 1942), 72–73Google Scholar.

²² The negro population in 1774 was almost entirely dependent, whether legally slave or free; indeed, in data as late as the 1790 federal census (United States Department of Commerce and Labor, Bureau of the Census, Heads of Families at the First Census of the United States Taken in the Year 1790: Connecticut [Washington, 1908]Google Scholar), virtually all negroes still lived in households headed by whites.

²³ The Public Records of the Colony of Connecticut, vol. 14 (Hartford, 1887), 485–92Google Scholar.

²⁴ In the case of Windham County the difference in ratios reflects an unexplained surplus of negro females above the age of twenty in the two towns of Mansfield and Windham. When these two towns are discounted, the cumulative ratios for the other towns in Windham County follow the same pattern found elsewhere in the state, with a greater surplus of males to females in the negro population above twenty (1.69:1) than in the negro population under twenty (1.52:1), although the ratio of negro males to females in the younger population is still substantially higher than the male:female ratio of the corresponding white population (1.09:1).

²⁵ McManus, E. J., Black Bondage in the North (Syracuse, 1973), 38–39Google Scholar.

²⁶ Sometimes an owner sold or bequeathed his slave to the mate's owner, so that the two could live together, but that did not happen often enough to have slaves develop the expectation that they would normally leave after marriage the household of his or her owner to reside in that of their mate's owner.

²⁷ Sinclair, R. K., Democracy and Participation in Athens (Cambridge, 1988), 9CrossRefGoogle Scholar. Though Sinclair does not say so, I would assume that most of these prisoners were non-Greeks from areas under Persian control just beyond the frontiers of the Greek-speaking world. Adult Greeks could easily escape from Athens and make their way home, unlike non-Greek barbarians. On most Athenian slaves being non-Greek (or, I would add, of non-Greek parentage), see Jameson, , “Agriculture and Slavery,” 140–1, n. 93Google Scholar.

²⁸ This point is made by Wiedemann, , Slavery, 25Google Scholar.

²⁹ Exposed children, both slave and free, were also an important source of new slaves. Quantification is impossible, but the enslavement of exposed children almost certainly played an important role in supplementing natural reproduction as a source of slaves, lessening substantially the need to import slaves in order to maintain a slave population which was still predominantly male and thus unable to totally reproduce itself. On the exposure of unwanted children, see further Boswell, J. E., “Expositio and Oblatio: The Abandonment of Children and the Ancient and Medieval Family,” American Historical Review, 89 (1984), 10–33CrossRefGoogle ScholarPubMed.

³⁰ “Talasiourgoi and Paidia in IG 2² 1553–78: A Note on Athenian Social History,” Historia, 38: 3 (1989), 365–70Google Scholar.

³¹ A quick reading of the probate inventories for eastern Fairfield County in the last half of the eighteenth century leaves the impression that male slaves were generally worth more than female slaves of the same age, that young males just entering adulthood were worth most of all, and that older male children were often worth more than mature male adults. I know of no corresponding data for Attic slaves, but the similar conditions discussed earlier suggest that the relative worth of male versus female and of old versus young child should be roughly the same in both slave regimes.

³² For publication data, see above, note 22.

³³ In the published returns this census category is labelled “All other free persons,” but inasmuch as Indians were not counted in the census, “all other free persons” must be “free negroes.” On Indians not being included in the 1790 census, see Heads of Families … in the Year 1790 (above, note 22) page 4, and page 6 for the text of the law authorizing the census. Free negroes living in households headed by whites are likely to have been ex-slaves who continued to live with their former owners.

³⁴ Fairfield also had a surprisingly low number of free non-whites compared with other towns, suggesting that its owners were more reluctant to manumit their slaves than those in other towns were. Because of this reluctance, the 1790 census figures give a better sense of the structure of the slave population of Fairfield than they do for other more liberal towns.

³⁵ The argument goes as follows. According to the 1774 census, slightly over 50 percent of the negro population of Fairfield was under twenty. Applying this same ratio to Fairfield's 1790 slave population, we estimate that approximately 102 of the 203 slaves in this population were under the age of twenty and approximately 101 above twenty. Most of the slaves under twenty were likely to live in the same household with at least one parent, and few were likely to be living in households with only one slave. Conversely, most of the 50 slaves living in households with only one slave were likely to be adults over the age of twenty. Assuming for the moment that all 50 were adults, if we subtract these 50 from the total slave population of 203, we will have approximately 51 slaves over twenty and approximately 102 slaves under twenty remaining.

Further, if we subtract the 50 households with only one slave from the total of 96 slaveholding households, we find that approximately 51 slaves over the age of twenty and approximately 102 slaves under the age of twenty lived in 46 different households. There is necessarily some fuzziness in these calculations for two reasons. First, slaves in their late teens (and in some cases perhaps as young as fourteen) worked as adults. Second, some even younger children could have been “bounded out,”" as it were, in a situation analogous to apprenticeship. Yet even if we were to assume a ratio of two-thirds functioned as adults (i.e. adult slaves, slaves in their late teens, bounded-out youth, etc.) and only one-third functioned as “children,” which is a very generous allowance, there would still be 85 functional adults and 68 children spread out over the 46 households with more than one slave. The general conclusions of this argument are confirmed by baptismal, marriage, and death records which allow us to reconstruct, albeit only partially, many slave families, showing the large number of children in them.

³⁶ MacMullen, R., Roman Social Relations: 50 B.C. to A.D. 284 (New Haven and London, 1974), 14Google Scholar.

³⁷ Mutual assistance from kin, neighbors, and friends was the normal recourse of poorer fanners; but this type of assistance was unreliable as a labor source, at least compared with chattel slavery. It also placed the fanner under reciprocal obligations which drew upon his most precious resource, his own labor at peak periods of agricultural activity.

³⁸ In contrast to a hired hand, a slave had to be purchased (or at least reared from birth), and the initial investment could easily be lost through early death from accident or illness, a factor of which our sources for northern slavery were certainly aware. One often reads that slavery in agriculture was inefficient off the plantation because slaves had to be fed, etc., year-round even though agricultural work, especially the raising of grain, was seasonal, with a few short periods of high activity and relatively long periods of inactivity. The argument is, however, probably overstated in view of the numerous other chores which subsistence agriculture required year round (see above, note 8). It is also far from certain that slaveholding farmers were all that concerned with their slaves’ “down time” (which would coincide with their own). Even in the oftquoted anecdote of the Staten Island slaveholder who complained that his hogs and corn were raised only to be consumed by his household during the winter, so that “he had nothing left but the Negroes” by the time spring arrived, the slaveholder still had no intention of selling these surpluses (for the anecdote, see McManus, , Black Bondage, 53Google Scholar, which quotes Morris, Ira K., Memorial History of Staten Island, vol. 2 [New York, 1900] 36Google Scholar). A general shortage of labor left few alternatives, and in such circumstances, even a bad deal (using slaves) was preferable to no deal at all.

³⁹ On the availability of land on the frontier, the consequent shortage of labor, and the importance of the latter for the development of slavery in New England, see Greene, , The Negro in Colonial New England, 100–1Google Scholar.

⁴⁰ Cf., for example, Aristotle, Rhetoric 1367^a33 (“for it is the mark of a free man not to live dependent upon another”) and Finley's elaboration: “The free man was one who neither lived under the constraint of, nor was employed for the benefit of, another; who lived preferably on his ancestral plot of land, with its shrines and ancestral tombs” (Finley, M. I., Ancient Slavery and Modern Ideology [New York, 1980], 90Google Scholar). For a discussion of some of the ancient evidence for this attitude, see de Ste. Croix, G. E. M., The Class Struggle in the Ancient Greek World from the Archaic Age to the Arab Conquests (Ithaca, 1981), 181–2Google Scholar.

⁴¹ This obviously oversimplifies what was in reality a very complex process, but I remain convinced that, in the end, the key factor in the rise of slavery in the north was the shortage of available labor and that the key factor in its demise was the end of that shortage. Note, for example, that Connecticut's first restriction of slavery, a ban on the importing of new slaves into the state, was enacted in October 1774 to protect employment for poor whites against further competition from slaves (“Whereas the increase of slaves in this colony is injurious to the poor and inconvenient …,” Connecticut Public Records, vol. 14, p. 329Google Scholar; see further, Fowler, , “Historical Status,” 17–18Google Scholar, who inexplicably dates the law to 1769). “In Massachusetts, John Adams recalled, the opposition of white labor assured the extinction of slavery for it ‘would no longer suffer the rich to employ their sable rivals so much to their injury’” (Litwack, L. F., North of Slavery: The Negro in the Free States, 1790–1860 [Chicago and London, 1961], 6Google Scholar). See further Zilversmit, A., The First Emancipation: The Abolition of Slavery in the North (Chicago, 1967), 46–47Google Scholar. The ideology of freedom fostered in the American Revolution also played some role in the gradual elimination of slavery in the north, but the failure of the same ideology to produce a similar effect in the southern states would seem to show that the role of ideology was, at best, limited in comparison with economic and social considerations (for arguments in favor of the importance of ideological concerns, see Litwack, , North of Slavery, 3–12Google Scholar, and Zilversmit, The First Emancipation).

⁴² By underemployed, I mean that they had the time, the ability, and (most important) the desire to work more. I am not using the term as it is used in the debate as to whether peasants are, by the nature of the agriculture they practice, underemployed, at least in comparison with modern models of efficiency.

⁴³ That some, and perhaps many, people in Attika worked for others, e.g. as hired hands, does not mean that a labor shortage did not exist, just as the existence of indentured servants and other forms of dependent labor does not mean that there was no shortage of labor in the northern colonies. A labor shortage exists whenever whatever the labor available is insufficient to meet current demand.

⁴⁴ “The freedom of the [Athenian] peasantry in various ways encouraged the growth of an alternative form of dependent labour, chattel slavery, and that in one way or another the freedom of the citizen and the degradation of the slave were but two sides of the same coin” (Wood, , Peasant-Citizen, 110Google Scholar). Wood's statement is true, but perhaps not in the way she intended it. In Wood's view most societies are formed of peasants and those who exploit them; but in Athens democracy protects the peasants, forcing the exploiters to turn elsewhere, viz. to chattel slavery, to extract their profits (ibid. 110–5). Even if it is true that Athens'; exploiting rich could not appropriate the peasants' surpluses, it does not follow, as Wood seems to argue, that the peasants will not dispose of those surpluses for their own benefit and indeed increase them if possible, for example, through the use of additional, slave labor. In effect, peasants can also be expropriators. Wood seems reluctant in general to admit this possibility, although she does at one point concede that “some smallholders would be able to afford a slave or two, whose principal functions would probably be in the house, but who might lend a hand in the fields” (78–79). Wood, of course, has no basis for the claim that these slaves' “principal functions would probably be in the house”; besides, as we noted earlier, the distinction between field and house slaves is meaningless in the kind of family slavery considered here.

⁴⁵ Thompson, L. A., Romans and Blacks (Norman and London, 1989), 63–67Google Scholar.

⁴⁶ Diller, H., “Die Hellenen-Barbaren-Antithese im Zeitalter der Perserkriege,” Grecs et barbares (Entretiens sur l’antiquité classique, 8 (Vandoeuvres-Genève, 1961), 40Google Scholar, with the remarks of Baldry, H. C., Grecs et barbares, 69 and 74Google Scholar. This would apply, of course, only to imported slaves, not to slaves born in Attika, who could be expected to leam their masters' language.

⁴⁷ See Tocqueville, , Democracy, 341–4 and 350–1Google Scholar, on the sad condition of free negroes in the north. On the isolation of the northern negro, Tocqueville (p. 351) points out that northern negroes were in effect aliens in a white population which continued to expand through immigration, unlike the negro population for whom immigration had come to an end with the elimination of the slave trade, which had been the only way for negroes to come to America. See also McManus, , Black Bondage, 180–8Google Scholar, and, for the period down to 1860, Litwack, , North of Slavery, 153–86Google Scholar.

⁴⁸ This argument will be even stronger if, as I suspect, a large proportion of metics (resident aliens), virtually all of whom lived in the urban demes and in Peiraieus, were in fact freedmen or their descendants.

⁴⁹ (Hartford, 1788), 21–25.

⁵⁰ Humphreys, , Israel Putnam, 22–23Google Scholar.