As adopted by the Constitutional Convention, Article I, Section 2, Clause 3 of the Constitution mandated that the population numbers forming the basis for calculating representation in the House of Representatives would “be determined by adding to the whole Number of free Persons, including those bound to Service for a Term of Years, and excluding Indians not taxed, three fifths of all other Persons.” Contemporaries and historians have contested the meanings and consequences of the stipulation that three of every five enslaved people be counted. This formulation bolstered the House delegation size of states with considerable numbers of enslaved people and further demeaned those who suffered under slavery. We have paid much less attention to the Framers’ innovation of including two other groups in addition to free white men. They also required the counting of non-Native American free women and free children when calculating representation. In doing so, the Framers tilted the House balance slightly back in favor of states with fewer enslaved people. Remarkably, no one at the time seems to have noticed the significant, lasting, practical ramifications of including white women and children, in addition to white men, when calculating each state’s number of House members.
That representation at the national level would include the enumeration of white women and children was far from a foregone conclusion during the American Revolutionary period. The same summer that the Constitutional Convention deliberated in Philadelphia, in New York City the Articles of Confederation Congress passed the measure eventually known as the “Northwest Ordinance.” This law established procedures for the area roughly bounded by the Ohio River, the Mississippi River, and the Great Lakes to progress from unincorporated land to statehood. At least in implicit conversation with the Constitution, the law’s text demonstrates the Confederation Congress’s muddied thinking on whom should be counted to calculate representation, and for what purpose. On the one hand, it required that territorial governments be established when districts exceed “five thousand free male inhabitants of full age,” and that there should be one elected representative “for every five hundred free male inhabitants.” On the other hand, it mandated that districts could ascend to statehood upon reaching a total of “sixty thousand free inhabitants.” While there is evidence that in seventeenth-century New England the term “inhabitant” may have implied “landowner,” that connotation seems to have faded, with the Confederation Congress intending the term to encompass men, women and children. Thus the Confederation Congress both included and excluded free women and children when considering population for the purpose of representation.
State constitution writers were more clear: overwhelmingly, the state constitutions in force as of 1787 counted voting or taxable free men only for the purposes of future reapportionment. Rhode Island and Connecticut retained their colonial charters until well into the nineteenth century. Several states’ constitutional framers pegged reapportionment to the number of taxpayers, usually white men. That included Vermont counting “taxable inhabitants” and Massachusetts “ratable polls.” Pennsylvania, where a small fraction of the population paid taxes, was most expressive in this regard, asserting that representation “in proportion to the number of taxable inhabitants is the only principle which can at all times secure liberty, and make the voice of a majority of the people the law of the land.” New York, Maryland, Virginia, and Georgia used “electors” or other terms tying representation to the number of eligible white male voters. While New Jersey ratified the only constitution allowing women to vote, New Jersey’s framers punted on reapportionment. They left to future legislators how to interpret their directive that changes in representation be decided “on the principles of more equal representation.” South Carolina’s language most prefigured the federal arrangement, probably because, of all the states, it had perhaps the most similar problem to the federal one in being divided among jurisdictions with large enslaved populations and those with few. The great proportion of South Carolina’s enslaved population resided in the low country parishes, while upcountry parishes were predominantly white. South Carolina’s constitution called for future reapportionment “according to the particular and comparative strength and taxable property of the different parts of the same, regard being always had to the number of white inhabitants and such taxable property.” “Taxable property” in this case signified wealth in slaves. For all their experimentation, enumeration by counting the full population went beyond state constitution-makers’ thinking.
The Constitutional Framers blazed a path in prescribing reapportionment by a means other than by counting those of its residents theoretically eligible to bear arms and to vote (white men), or those who paid taxes (rateables and polls). The federal Constitution appears to be the first American document that explicitly tied future apportionment to an exact count of the total population, and certainly the first one that did so using such a complex formula. Perhaps this novelty contributed to lack of consideration of its mathematical complexities.
The Constitutional Framers generated their formula for the purposes of determining both states’ respective federal privileges, in the form of House representation, and their federal obligations, in the form of taxation. Consideration of the latter motivated the origin of what was proposed and adopted in the Constitution. According to James Madison’s notes of the Constitutional Convention, the committee first tasked with considering the number of legislative representatives reported their recommendation on July 9, 1787. That committee came up with a plan for a total of fifty-six representatives, based upon a formula of one representative for every 40,000 people. When pressed on what data they used, committee member Gouverneur Morris implied that the committee considered both free and enslaved inhabitants, and admitted that, as far as exact numbers, “the Report is little more than a guess,” although among William Paterson’s notes for his own June 9 speech were the Confederation Congress’s state-level population estimates as of 1785. Because the Convention delegates from several states were displeased with the number of representatives their respective states were allotted, the committee was reconstituted. The next day, the new committee reported with a plan that was ultimately adopted, calling for sixty-five representatives. No documents indicate what happened in that committee meeting, but the changes appear to have been designed to secure the support of states that gained representation, rather than based upon either different population estimates or a rationale for how representation was calculated.
Although that committee’s numbers ended up being adopted for the makeup of the first federal Congress, James Wilson produced the exacting and ultimately successful plan concerning the future divvying of representation. According to James Madison’s notes of the Convention, Wilson suggested that taxes and representation under the Constitution be calculated “in proportion to the whole number of white & other free Citizens & inhabitants of every age sex & condition . . . and three fifths of all other persons not comprehended in the foregoing description, except Indians not paying taxes.” He noted that such had been “the rule in the Act of Congress . . . for apportioning quotas of revenue on the States.” Wilson was referring to the Article of Confederation Congress’s 1783 proposed amendment to the Articles of Confederation for the purposes of raising revenue more reliably and being able to calculate states’ respective shares more easily than the prevailing method of relying on real estate evaluations. The Confederation Congress had decided that population could serve as a proxy for wealth. The three-fifths formulation resulted from a compromise among several possible ratios for the comparison of how much enslaved people contributed to wealth compared to free people, and had nothing to do with representation at that point. Although there were several attempts in the Articles of Confederation Congress to move to representation by population, representation at the national level remained one vote for each state until the establishment of the federal government.
Whether white women and children were counted in the Constitution may seem like a question that had no immediate practical consequences. One might guess that as long as the number of free men as a proportion of overall population did not vary much from state to state, whether or not white women and children were included would not affect the comparative size of different states’ congressional delegations. That assumption is wrong.
Still, the delegates to the Constitutional Convention appear to have operated under this misguided notion. Hypothetical scenarios with simplified numbers illustrate the point. If free men comprised a third of the residents of State A, and a third of the residents of States B and C, then counting their whole populations would not change the comparative sizes of their Congressional delegations. Table 1 shows this stable ratio to be the case. The additional inclusion of free women and children in Table 1 means more representatives, but the proportional relationships among the states remains intact. In other words, in this scenario, adding white women and children does not make a material difference to the states’ relative power in Congress.
Slavery’s existence, though, and the southern Framers’ desire to protect it, warped representation no less than it did all other facets of the early republic. The additional counting of three out of every five enslaved people rendered the inclusion of white women and children to be highly consequential. Let us look at another hypothetical scenario, this time with two fictitious states of equal population, State D half enslaved and half free, State E entirely free. In Table 2 we see that the ratio of representation between states does indeed vary according to whether white women and children, in addition to free men and a portion of enslaved people, would be included. Without white women and children counted, State D’s delegation exceeds that of State E; with white women and children counted, State E’s delegation proves larger than State D’s. In short, counting a larger proportion of the free population increases the relative power of those states composed of more free people, and could do so significantly. Just as this is the case in our imagined scenario, it was true of actual Congressional apportionment in the early republic.
The first practical application of the Constitution’s apportionment formula indicated the effects of including white women and children. As noted above, the Constitutional Convention somewhat arbitrarily set the initial number of House members for each state. In 1792, Congress faced the conundrum of how best to apportion House seats based upon the results of the 1790 census. That sparked a political conflagration, with heated debates both within the Washington administration and Congress, and initial legislation that drew President George Washington to wield his veto for the first time. Article 1, Section 2, Clause 3 of the Constitution of the United States mandated that “the Number of Representatives shall not exceed one for every thirty thousand” counted according to the formula that Wilson had proposed, with every state having at least one House member. Beyond that, it was silent, leading to two major questions. First, how large should the House of Representatives be? Second, by what mathematical formula would that number of seats be divided? The first of those questions appears more philosophical than practical, and the second technical. But, in combination, they could mean the difference between a smaller state being limited to one seat or granted two, or a bloc of states having a slight deficiency or edge.
Congress eventually passed, and Washington signed, a plan devised by Secretary of State Thomas Jefferson. Jefferson called for the enumerated population of each state to be divided by 33,000, with the fractional remainders ignored. In other words, if a state had an enumerated population up to 33,000, it was granted one representative. An enumerated population of anywhere from 33,001 to 65,999 also resulted in one representative rather than two, as fractions would not be counted. This method produced a House composed of 105 members. Perhaps not coincidentally, Jefferson’s Virginia gained a seat compared to the formula proposed by Jefferson archrival and New Yorker, Secretary of the Treasury Alexander Hamilton.
Table 3: 1792 House Delegations Excluding and Including Free Women and Children
*Enumerated population includes free white men and three-fifths of enslaved people.
Note: All population numbers taken from the 1790 U.S. Census.
With the inclusion of white women and children, the makeup of the Third Congress of the United States looked very different than if only free men and enslaved people had been counted. Without them, if we were to use Jefferson’s formula, the total number of seats would only have been 33, rather than 105. Some states would have lost influence with the inclusion of white women and children, while others would have gained leverage.
The inclusion of white women and children substantially affected the balance of power in the House of Representatives along a line that deeply concerned Revolutionary-era Americans. Not surprisingly, including a higher proportion of free people in the total count benefited what Americans called at the time “eastern” and what we now call the “northern” states as opposed to the Southern states, as Table 4 indicates.
The inclusion of free women and children, as Table 5 shows, also altered the House balance of power in terms of another deeply contentious issue, especially during the Constitutional Convention. The percentage of members of the House held by the seven smaller states declined with the inclusion of free women and children compared to what it would have been without them.
Furthermore, because the inclusion of free women and children tilted the balance of power northwards in the House of Representatives, it also did so for the Electoral College. That made no difference for the 1792 presidential election, when every elector cast a ballot to return Washington to office. But 1796 was much closer. Massachusetts’s John Adams garnered seventy-one electoral votes, while Virginian favorite son Jefferson tallied sixty-eight. By increasing the relative number of northern votes, the enumeration of free women and children probably threw Adams over the top.
Despite its novelty and that it held important consequences for representation, the decision to enumerate free women and free children, as opposed to only free men, appears to have been met with silence. It occasioned no recorded comment during the Constitutional Convention. The three authors of this piece conducted a thorough search of the index and text searches of The Documentary History of the Ratification of the Constitution. We found no evidence of debate concerning this issue in any of the state ratification proceedings. A survey of the scholarly literature came up empty. We also emailed John Kaminski, longtime editor of the Documentary History of the Ratification of the Constitution project; Jan Lewis, the foremost authority on the representation of white women in the Constitution; and Rosemarie Zagarri, author of prominent books both on the issue of representation and on women’s rights during the American Revolutionary period. None could recall coming across late-eighteenth-century Americans discussing the political ramifications of enumerating white women and children in the Constitution. If the members of the wordy—sometimes prolix—Revolutionary generation discussed the issue, they left little or no trace, suggesting that they simply did not think of it.
This is not to say that no one considered the broader philosophical question of whom the Constitution represented. In her seminal article on this topic, historian Jan Lewis argued that the Framers considered white women as essential to representation in the Constitution. Lewis concluded that James Wilson, who first drafted the clause mentioning women, intentionally included them in the body politic because the purpose of government, in Wilson’s eyes, was to facilitate happiness for all of its citizens. Lewis’s claim was like a wildly post-modernist house situated on an otherwise staid city block. Her question of whether women were represented was a major departure in terms of what historians had asked. Lewis’s reasoning, though, fit squarely in the greater historiographic framework concerning the debates over representation, placing the representation of women as a logical consequence of late eighteenth-century republican and liberal thought. But Lewis’s articles on the subject did not consider the immediate political consequences of counting free women, much less the issue of doing so for free children. The political ramifications of including free women and free children for the purpose of apportionment have remained a lacuna in the literature concerning representation and the Constitution.
The above tables beg the question of how such a politically powerful innovation could slip by unremarked. During the debates within the states over the ratification of the Constitution, Americans hotly contested nearly every other consequential clause. Making claims concerning why something did not happen—in this case, controversy over a significant element of representation—requires more speculation than interpreting why events happened how they did. We suggest a combination of two factors, one having to do with the particular historical context, the other a broader human failing.
First, the context. Had the expansion of the expansion of enumeration beyond free men to include white women and children occurred on its own, perhaps there might have been comment, even if Americans had estimated that the proportionality of the increases rendered moot its political consequences. However, as we have noted, that did not happen independently. Rather, it occurred in tandem with the counting of three out of every five enslaved people. For white Americans, once enslaved people were being counted—and in such a strange and controversial way—to exclude white women and children would have been unthinkable, regardless of how it would have affected apportionment in Congress. The glare generated by the fractional inclusion of enslaved people perhaps blinded late-eighteenth-century Americans to the mathematical ramifications of including white women and children.
We suggest that the other, deeper reason for the inattention stems from what may be a timeless human shortfall when encountering this kind of quantitative problem. In one of their pathbreaking articles establishing the concept of implicit bias, psychologists Amos Tversky and Daniel Kahneman identified an interesting phenomenon: when simultaneously presented with gauging the probability of two events, they rate the chance of both happening more highly than the rate the possibility of one or the other happening independently. For example, Tversky and Kahneman gave a group of students a description of someone with certain liberal political proclivities. They asked the students how likely the person was to be a bank teller. The students were then given the opportunity to consider the probability that the person would be a feminist activist bank teller. The respondents actually rated the latter to be more likely than the former—notwithstanding what would seem the obvious conclusion that the chances of someone being a bank teller, regardless of their politics, should be higher than that of them being both a feminist activist and a bank teller. Tversky and Kahneman labeled this failing of human reasoning the “conjunctive fallacy in probability judgment.” They and subsequent researchers have reproduced this research in a variety of settings, and found that even people highly knowledgeable in statistics can fall victim to this misunderstanding of basic probability.
That may have been exactly what happened with the Framers and the inclusion of white women and children along with three fifths of enslaved people. The Framers likely assumed that, had only white people been counted, the probability that the inclusion of women and children would have left the comparative size of the various state delegations unaffected was very high. That guess turns out to be mostly correct. As the 1790 census indicates, adult males comprised no less than 30.66 percent and no more than 35.60 percent of any state’s free population. So far, so good. However, when presented with the familiar formula of white population plus three-fifths of enslaved people, the Founders, like the students in Tversky and Kahneman’s study, may have assumed the probability of including white women and children not affecting apportionment was even more likely than without the three-fifths clause.
The authors’ own experience reinforces our conjecture. Schocket had assigned Lewis’s article as reading for a class session in an undergraduate course on the American Revolution. Perhaps prompted by his own coursework in demographics, Spicer asked how the inclusion of women impacted apportionment, a question McLaren-Czerr followed up on. Crucially, neither were operating under the assumption that including free women would necessarily leave apportionment unaffected regardless of the three-fifths clause. They were thus less susceptible to the deductive fallacy under which Revolutionary-era Americans and previous historians, Schocket included, operated. Putting the 1790 Census numbers on the board during that class session revealed that the inclusion of free women and children in combination with three-fifths of enslaved people did indeed result in a surprising concatenation of affects.
The expansion of enumeration to include all white people in addition to three-fifths of enslaved people at the national level set an important precedent. As Lewis noted in her more recent article following up on her original one, it resulted in a nation in which all the nation’s residents, rather than only propertied white men, were nominally represented constituents of and subjects to the federal government. More immediately, the inclusion of free women and children resulted in higher representation of free states than otherwise would have been the case both in the House of Representatives and in the Electoral College. The failure of early Americans to notice the momentousness of that inclusion reinforces an important truth: in this as in every other way we have found to examine the Founders, rather than demigods, they were no less fallible than the rest of us.
Kinzey M. McLaren-Czerr is pursuing her Bachelor’s degree in History at Bowling Green State University. As an honors student, she is a member of the Historical Society along with the Honors College, Student Advisory Board, Phi Alpha Theta, and the National Society of Leadership and Success. Along with her co-writers, she is thrilled to share with the public domain her first piece of published work. In addition to her primary focus of History, she also studies Women’s, Gender and Sexuality Studies. Through this article, she hopes to broaden the conversation of gender involvement in regards to documentation in Revolutionary History.
Andrew M. Schocket is Professor of History and American Culture Studies at Bowling Green State University. He is author, most recently, of Fighting over the Founders: How We Remember the American Revolution (NYU Press, 2015) and an editor of the Cambridge History of the American Revolution (Cambridge University Press, in progress). His writing has appeared in the Washington Post, San Francisco Chronicle, Newsday, and Slate, and he is a co-creator of the Magazine of Early American Datasets.
Colin J. Spicer is an undergraduate student at Bowling Green State University. This is his first written work to be published. He was a student in Professor Schocket’s class and posed the question of how the inclusion of women would affect apportionment. While he loves studying history, he is a geography major that hopes to one day be a city planner that focuses on revitalizing struggling communities. He hopes he can have a long and successful career.
“An Ordinance for the Government of the Territory of the United States North West of the Ohio River” (1787), memory.loc.gov/cgi-bin/ampage?collId=lljc&fileName=032/lljc032.db&recNum=343.
John Frederick Martin, Profits in the Wilderness: Entrepreneurship and the Founding of New England Towns in the Seventeenth Century (Chapel Hill: Published for the Institute of Early American History and Culture, Williamsburg, Virginia, by the University of North Carolina Press, 1991), 217–28.
Francis Newton Thorpe, The Federal and State Constitutions, Colonial Charters, and Other Organic Laws of the State, Territories, and Colonies Now or Heretofore Forming the United States of America (Washington : G.P.O., 1909), 6: 3755; and 3: 1898, archive.org/details/cu31924019912074. Although the document says “ratable,” it apparently means “rateable,” that is, subject to taxation.
Rhode Island and Connecticut kept their colonial charters in force until well into the nineteenth century. For a longer discussion of the principles involved at the state level, see Marc W. Kruman, Between Authority & Liberty: State Constitution Making in Revolutionary America (Chapel Hill: University of North Carolina Press, 1997), 65–76.
United States Department of State, Documentary History of the Constitution of the United States of America, 1786-1870., vol. 3 (Washington: Department of State, 1900), 296, catalog.hathitrust.org/Record/100772439; William Paterson, “Papers of William Paterson on the Federal Convention, 1787,” American Historical Review9, no. 2 (January 1904): 332–324.
James Madison, The Debates in the Federal Convention of 1787: Which Framed the Constitution of the United States of America, ed. James Brown Scott and Gaillard Hunt, Journal of the Federal Convention (New York: Oxford University Press, 1920), 88, catalog.hathitrust.org/Record/010446630.
Thomas Jefferson, “Opinion on the Apportionment Bill, 4 April 1792” (n.d.), founders.archives.gov/documents/Jefferson/01-23-02-0324, accessed May 10, 2019.
M. L. Balinski and H. P. Young, “The Jefferson Method of Apportionment,” SIAM Review20, no. 2 (April 1978): 278–84, doi.org/10.1137/1020040.
United States Department of State, Return of the Whole Number of Persons within the Several Districts of the United States: According to “An Act Providing for the Enumeration of the Inhabitants of the United States,” Passed March the First, One Thousand Seven Hundred and Ninety-One (Philadelphia: Childs and Swaine, 1791), 4.
National Archives and Records Administration, “U. S. Electoral College Historical Election Results: Electoral Votes for President and Vice President 1789-1821,” U.S. Electoral College, www.archives.gov/federal-register/electoral-college/votes/1789_1821.html#1796, accessed May 24, 2019.
Merrill Jensen, John P. Kaminski, and Gaspare J. Saladino, eds., The Documentary History of the Ratification of the Constitution, 34 vols. (Madison: Wisconsin Historical Society Press, 1976), digital.library.wisc.edu/1711.dl/History.Constitution.
John Kaminski to Andrew M. Schocket, “Re: Query about Representation & White Women and Children,” November 28, 2018; Jan Lewis to Andrew M. Schocket, “RE: Question from SHEARite Concerning ‘“of Every Age, Sex, and Condition,”’” April 21, 2018; Rosemarie Zagarri to Andrew M. Schocket, “Women as Part of the Counted Inhabitants in Constitution,” July 23, 2018.
J. R. Pole, Political Representation in England and the Origins of the American Republic (New York: Macmillan; St. Martin’s Press, 1966); Rosemarie Zagarri, The Politics of Size: Representation in the United States, 1776-1850 (Ithaca: Cornell University Press, 1987); Gordon S. Wood, The Creation of the American Republic, 1776-1787 (Chapel Hill: Published for the Institute of Early American History and Culture at Williamsburg, Va., by the University of North Carolina Press, 1969).