II. Family Disruption at Commercial and Court Sales

In the preceding section, I presented empirical evidence that individual sales predominated at South Carolina court sales. The South Carolina evidence was similar to data drawn from a group of Missouri slave sales that were mostly probate sales. I also demonstrated that this empirical evidence was consistent with the judicial preference for individual sale at court sales in North Carolina and Kentucky. The South Carolina and Missouri evidence was also consistent with Georgian Howell Cobb's observation that the sale of slaves in families was not the usual practice at sheriffs' sales. Although the legislatures of Alabama, Georgia, and Louisiana protected the relationship between very young children and their mothers--as long as a nonslave with a property interest did not object--the general rule and practice appears to have been the opposite: that slaves should be sold singly rather than as parts of family groups. Judges enforced the practice of single sale by holding those who conducted the sales personally liable if their sales of slaves in groups yielded less money than individual sales would have. For seven of the fifteen slave states--Alabama, Georgia, Kentucky, Louisiana, Missouri, North Carolina, and South Carolina--there is evidence of varying sorts that at court sales the preference was for individual sales of slaves. This is perhaps enough evidence to support the generalization that individual sale was the default rule for court sales of slaves in the antebellum South.

In this Part, I compare court sales of slaves with commercial sales. Specifically, I focus on the risk of family separations. I make a rough comparison of this risk at the two types of sales using as a variable the incidence of individual sale. After comparing the risk of family separations, I will, in Part III, compare the sales proceeds at these two types of sales.

By itself, the insistence of appellate justices that slaves be sold singly at court sales is some evidence that the frequency of individual sale at court sales was greater than at noncourt sales. The existence of such a rule implies, of course, that fewer slaves would be sold individually if no such rule existed. That in turn suggests that the default social or economic selling practice--that is, the manner in which slaves would be sold absent such a rule--included a smaller fraction of individually sold slaves. At best, though, this inferential evidence regarding the relative incidence of individual sales as between court and commercial sales is merely suggestive.

The most obvious sources of data to which one might turn for a less inferential conclusion regarding comparative rates of individual sale at commercial and court sales are bills of sale. These documents are well-known to historians of slavery, particularly economic historians. However, there are enough difficulties with the use of bills of sale that I believe they merit something of a technical discussion in the text of this Article. I would like to suggest that historians should use bills of sale very cautiously.

Bills of sale were documents that a buyer of property--any property, not just slaves--might choose to file with the state after purchase. Although a bill of sale might be recorded for any type of property, in South Carolina, nearly all (96.5%) of the bills filed with the state were for slave sales. The most important reason for a buyer to file a bill of sale was to provide certain evidence of title. A certain record of title benefited a buyer if a question ever arose regarding the ownership of property. In this way, bills of sale were like the systems that states and counties today use to record real estate or automobile sales. The bills-of-sale records were also significant to creditors. Slaves formed the basis of many Southern credit transactions, and a smart creditor would want to be sure that the debtor held recorded title to the slaves against whom the creditor would levy if the debtor defaulted.

One problem with using bills of sale as representations of commercial sale is that the bills include transactions that are not commercial sales. These noncommercial slave transactions included among the bills are of two types. First, the bills include court sales. I excluded those bills of sale that indicated that the seller was acting as an agent of the courts. These court-sale bills amounted to fifteen percent of the total number of bills, which represented twenty-three percent of all the slaves sold. The second group of transactions among the bills that were not exclusively commercial sales were transfers to or by trustees, usually trustees for married women. Many of these transactions were marriage settlements. In all, just over ten percent of all the slave bills of sale were transfers by trustees, with most being trustees for married women. The exclusion of court sales and transfers by trustees leaves seventy-one percent of the total number of slave bills of sale, which included fifty-nine percent or 2642 of the 4492 slaves among the bills that I examined.

I suspect that the noncourt bills of sale are largely for the category of sales that Tadman identifies as local, noncourt sales. Tadman divides slave sales into three types. The first type is interregional sales, which, as noted above, were Tadman's principal interest in Speculators and Slaves. The other broad category of sales that Tadman considers is local sales, divided into two types, court and noncourt sales. In terms of the numbers of such sales, Tadman regards court sales as the most frequent, followed by interregional sales, with local, noncourt sales being the smallest category. For all slave sales, I estimate that court sales comprised fifty percent; local, commercial sales sixteen percent; and interregional sales thirty-four percent. Because an interregional buyer of a South Carolina slave would have almost no incentive and little opportunity to file a bill of sale in the exporting state, the noncourt bills of sale are probably mostly for local, noncourt sales. As most court sales were also local, that is, not interregional sales, this group is most directly comparable to court sales.

Regarding separate sale, I draw my conclusions regarding commercial sale from the subset of bills that were not court sales. As noted above, a majority (fifty-two percent) of slaves sold at South Carolina's court sales experienced individual sale. The bills of sale indicate that at local, commercial sales nineteen percent of slaves were sold individually. Therefore, the risk of family separation due to individual sale appears to have been between two and three times higher at court sales than at commercial sales.

One additional problem concerning the use of bills of sale as representations of commercial sales merits consideration. For court sales, my empirical conclusions are based upon the sale records of the agents of law who actually conducted the sales. These records are the primary records of such sales, that is, the sellers made these records at the time of sale. In this way, the official sale books of the legal officials are analogous to ledgers or account books in which a commercial merchant might record sales. Unlike the officials' sale books, the bills of sale are not the primary records of sales. The bills of sale are instead secondary recordings of completed transactions. For this Article, the principal difficulty that the secondary nature of the bills poses is whether the number of slaves listed on an individual bill of sale reflected the number of slaves sold as part of any one transaction. For example, 495 of the 802 commercial bills that I examined were for individual slaves. Did the recording of bills reflect the conduct of sales? That is, were the bills that listed individual slaves all from sales of individuals, or did some buyers file an individual bill for each slave purchased as part of a group? Another possibility is that when a buyer bargained for and purchased a number of slaves individually, from the same seller, the buyer might file one bill of sale that included all of the slaves. The latter possibility--grouping individually purchased slaves onto one bill of sale--seems more likely than the former, as filing an individual bill for each slave purchased would not offer a buyer any greater security and would only heighten the costs and inconvenience of filing.

The bills themselves do not yield an answer to this query, although comparison with data drawn from the court-sale records provides some clues. Fourteen percent of the bills were for probate sales. The size of the groups included in the bills of sale for probate sales was substantially larger than the average size recorded in primary records of probate sales. For probate sales that were among the bills-of-sale records, the average number of slaves recorded on each bill of sale was 6.5; among the bills that were not for individual sales, the average group size was 9.5. These averages are much larger than the figures derived directly from the probate-sale records: a mean of 1.5 slaves per sale, with 3.4 slaves being the average group size. This greater average size of groups in the bills suggests that probate-sale buyers aggregated individuals or small groups of slaves into larger groups before filing bills of sale--perhaps in order to reduce transaction costs--and/or that buyers of larger groups felt greater need for the security that filing a bill of sale provided.

If the behavior of buyers at noncourt sales were similar to that of probate-sale buyers, then the bills of sale for commercial sales would skew from individual sales toward larger groups. However, since in commercial sales an official record of the sale did not already exist, buyers at these sales had more inducement to file bills of sale than did buyers at court sales. This means that for commercial sales, the bills of sale are probably more representative of the sales transactions themselves than are the bills for court sales. Just how much more representative the bills of commercial sales are one can only guess. There are no extant records with which to resolve this question, so I presume that this nineteen percent incidence of bills of sale with individual slaves reflects corresponding sale percentages. Thus, my use of these data in this way is not conservative with regard to my argument that the incidence of individual sale was higher at court sales than at commercial sales.

The court-sale data approach the patterns of family disruption that Tadman identified for interregional sales. As noted above, when comparing local commercial sales with interregional sales, Tadman has concluded that the local sales “were in general not so systematically oriented toward separation as was the interregional trade.” With regard to the comparison between local commercial sales and court sales, Tadman appears to be mistaken. His evidence is convincing, however, with regard to the comparison between local commercial sales and interregional sales. Tadman has estimated that about one-quarter of the slaves that interregional traders purchased were young slaves between the ages of eight and fourteen who were purchased and resold singly. He has also estimated that family fragments--mothers with children and husbands and wives, with or without children--comprised only about twenty-two percent of the slaves sold interregionally. Although Tadman does not specifically report the figures for individual sale in the interregional trade, this rate is likely as high or higher than at court sales. Thus, in terms of the risk of family disruptions, court sales--the largest overall category of sales--exceeded local commercial sales, but they, in turn, may have been eclipsed by interregional sales.

Court sales of slaves, then, included an incidence of individual sale that was between two and three times higher than for the commercial sales included among the bills of sale. This differential suggests a higher risk of family separation at court sales than at local, commercial sales. Unfortunately, the rather skeletal quality of the bills-of-sale data, not to mention the contested character of the figures for separation risk at commercial sales, makes impossible the derivation of a precise figure for this differential, but it is clear that law's agents conducted one-half of all slave sales and did so in a manner that magnified the risk of family separations.