From “great men” to “we White men”: British enfranchisement 1800–1911 CE and causal theory of democracy

1 Introduction

Linguistics, together with its companion discipline psychology, offers to remedy the twin fallacies widespread across recent theories of democracy. The fallacies combine an influential misconception of causality with a knowing neglect of the collective action problem that deters people from voting when electorates are even modestly large. Since democracy arises when people vote and rulers agree to introduce democracy in the expectation that people will, neglect of the collective action problem prevents political scientists from recognizing democracy’s cause.

Both fallacies can be remedied. The remedy deploys recent findings in linguistics together with the findings of psychological experiments originally designed to investigate prejudice, as well as the findings of a very wide range of other experiments in psychology and related fields. These experiments reveal the effect of language use on human estimation of the effort that outweighs gains from voting and other collective actions when potential participants are too numerous or dispersed to watch each other. The experiments reveal what to look for in the history of enfranchisement—the spread of the voting rights that distinguish democracy from its opposite, dictatorship—among people whose decease has put them beyond the reach of experimentation. Enfranchisement in Britain is not just an individual case. Because Britain’s global empire brings it into contact with populations everywhere, political discourse in Britain powerfully influences all other enfranchisements.

2 Twin theoretical fallacies

2.1 Confusion about causality

Geddes (2009, n.p.) documents the “inability of careful researchers to arrive at a consensus about the causes of democratization.” Dissensus arises because democracy’s “causes differ systematically” and also (Geddes 2011, 595) because “the phenomenon we label democratization actually includes several different causal processes.” These assertions echo her opinion 15 years earlier: “a compelling explanation for the transition from authoritarianism to democracy” is “probably impossible” (Geddes, 1994, 552). She is hardly alone (Haggard and Kaufman, 2016, 30; McAdam et al., 2001, 27; Huntington, 1991, 38). In historical institutionalism, “causality…is linked to contingent decisions made during critical junctures leading to new configurations of institutions characterized by path dependence” (Bernhard, 2015, 978).

These assertions about causality are quite puzzling—really, remarkably so. They claim that causality in political science differs from causality in any other science. For it is basic to the concept of causality that a cause must be uniform across instances of the same effect. Admittedly, Brady (2011) proposes a taxonomy of four concepts of causality: (1) neo-Humean, (2) counterfactual, (3) manipulation, and (4) mechanism. Yet these are not four concepts but one. To David Hume, one event may be said to cause another when two events regularly occur together and one event precedes the other. Neo-Humean concepts of causation arise because Hume (1739, 44) himself notices, “One object can be contiguous and prior to another without being thought to be its cause.” Encountering objections that reasoning might somehow identify causes, Hume (1902 [1748], 30) issues a challenge: “If you insist that the inference is made by a chain of reasoning, I desire you to produce that reasoning.” Manipulations, mechanisms, and counterfactuals rise to Hume’s challenge while retaining his requirements of regularity and precedence. One chain of reasoning distinguishes a causal relationship by manipulation, using experiments to investigate whether the removal of the cause can suppress its effect. But when experiments are infeasible, a mechanism may substitute. It being infeasible to switch off gravity, mathematics can be used to check whether already known orbits can be derived from novel laws of motion (Hyman, 1992, 147). Counterfactuals are claims of the form “if event c had not occurred, event e would not have occurred.” Although philosophers continue to ponder inadequacies and complexities encountered in construing this conception of causality, they generally seem to concur that mechanisms and manipulations—“interventions”—are claims of this form (Menzies and Beebee, 2025). One counterfactual is famous in the theory of democracy: “No bourgeois, no democracy” (Moore, 1966, 418). It is not only famous but also false. Having previously eliminated all private employment—the defining activity of the “bourgeois”—various post-communist states have become democracies. The possibility of falsifying the claim allows for the determination of whether it expresses a causal relationship.

Of course, it is possible, in the abstract, for democracy to lack any neo-Humean cause. But perhaps Geddes and others find no regular cause or set of causes because they have unduly narrowed the range of possible causes that they are willing to contemplate.

2.2 Evasiveness about collective action

Much extant theory limits the range of possible causes by evading the problem of collective action. Many theorists imagine the choice between dictatorship and democracy as a strategic interaction over the distribution of income among individuals assumed to worry about future incomes. Enfranchisement empowers more individuals who are poorer now to vote. They compose a majority. Enfranchisement carries the risk that this majority might elect candidates who promise to tax the rich and redistribute the proceeds to the poor. If so, the minority composed of those wealthier now stands to gain by resisting enfranchisement. Enfranchisement can only happen when the anxiety of the wealthy about increased taxation attenuates because the distribution of wealth becomes more even or taxable assets become more mobile (Acemoglu and Robinson, 2006; Boix, 2003).

Yet the assumption that all individuals worry about gain defined as income also contradicts these arguments’ expectations that the poor may vote to redistribute and that the wealthy will resist enfranchisement. For voting in any mass election is famous for its incompatibility with attending to gain. “Pivotal voter theories of turnout predict vanishingly small turnout rates in large electorates” (Cox, 2015, 50). Without further heroic assumptions, seldom, if ever, made explicit, vanishingly small turnout rates are actually zero. Turnout is claimed to exceed zero on the grounds that “if no one else votes, an individual citizen will find it worthwhile to vote: at the small cost of [voting], she can decide the entire country’s government” (Medina Sierra, 2018, 11). But the same theorist also says, “no single vote is responsible for absolute victory or defeat” because “human and machine mistakes” make the outcome of any close election indeterminate. The indeterminacy of close elections neglects an even weightier issue: whether any official elected by only one person can possibly expect everyone else to comply passively with the allocation of resources that everyone wants.

To these theorists, the paradox of expecting both voting and no voting does not matter. Knowing full well that voting presents a collective action problem, they resolve it by assumption: “as the least well-off overcome their collective action problems” (Boix, 2003, 13, also 26, 45); “if there are no collective action problems, or if they can solve any that do exist” (Acemoglu and Robinson, 2006, 20, 126–128). And in these passages, they also pretend that collective action is problematic only for the poor. They do not consider whether the wealthy need to act together to prevent the poor from voting, for example, by hiring a few of the poor to police the rest, or whether some of the wealthy might protect their own wealth from the demands of the poor by allying with the poor to confiscate from others of the wealthy.

The combination of confining the collective action problem to the poor and either dismissing it by assumption or ignoring it entirely appears in much other literature. An influential “historical institutionalist” agrees that “elite actors and their supporters…often” compete over the distribution of wealth (Bernhard, 2015, 980–982) but never even mentions collective action. The original formulation of selectorate theory mentions collective action only once in a long book (Bueno de Mesquita et al., 2003, 308). A later extension proposes that dictatorships succumb to democracy when a challenger promises to shift disbursements of state revenue from the provision of private goods among a dictator’s small coalition to the provision of collective goods among a large rebel coalition (Bueno de Mesquita and Smith, 2009, 174). No explanation is given as to why the promise of collective goods, available to each without effort, induces anyone to contribute effort to rebelling.

A theory distinguishes between enfranchisement from below achieved by popular protest and enfranchisement from above granted by dictators who concede the franchise without protests (Haggard and Kaufman, 2016). But this distinction ignores the strategic interaction between dictators and protesters. Since protest requires effort, protesters prefer to gain the franchise by concession from the dictators without protest. Dictators who order the suppression of protests sometimes suffer death or imprisonment if suppression fails. Dictators fearing that suppression could fail may prefer to concede before protests begin. Either way, any individual who suppresses or who might protest must be willing to contribute effort that is unnecessary if enough others contribute.

Other theories evade the collective action problem by keeping agents few. This evasion is particularly prominent in theories of dictatorship. Three specialists are invited to review “important contributions” of “formal models of non-democratic politics.” They explore such a model’s implications for societies with two, three, or four members. “This framework can be used to examine a variety of questions about authoritarian politics” (Gehlbach et al., 2016, 568-569). It can indeed, for polities whose inhabitants number in the single digits. Another formal model camouflages the same limitation of applicability to polities too small to qualify as polities. “Imagine a world where the population is made up of many small groups…Some individuals specialize in violence…The violence specialists may provide protection to a small group of clients, but the biggest threat facing the specialists is one another.” While “it is an equilibrium outcome for both specialists…to continue fighting,” it is also possible “for one specialist to stop fighting” only if that specialist sees “that it is in the other’s interest not to fight, a perception that both specialists must share” (North et al., 2009, 19-20). Without notice, what starts as “many” becomes “some” and then “both,” as if these quantifiers were synonymous. The authors themselves conclude: “We do not provide a coherent and well-integrated theory of the state” (North et al., 2009, 271).

The most persuasive arguments about the collective action problem in democracy claim that politicians solve it by mobilizing their followers (Cox, 2015, 65). Yet the theorist recognizes: “these models…do not explain…why their followers respond by bearing the costs of participation” (Cox, 2015, 50). Cox (2015, 54) offers two reasons for them to bear costs. First, “[social] network participants are sometimes willing to impose penalties on one another for nonvoting, even if the act of penalizing is itself costly.” Second, some politicians use opportunities to make voters compete for public services by giving priority to precincts or districts returning the highest favorable vote or denying services to those returning the most unfavorable vote. But if only some politicians make voters compete for public services and only some social networks penalize non-voting, neither is a general explanation of why people vote.

A recent theory uses mobilizational effort to connect the retention of wealth with decisions to enfranchise. The theorist argues that conservative political parties agree to enfranchisement if they can recruit poorer voters (Ziblatt, 2017). The English Conservative Party organizes a Primrose League with a listed membership increasing from 857 in 1884 to 1,556,639 in 1901 (Ziblatt, 2017, 109). Significantly, 1884 is also the year of the “most democratizing of Britain’s three major nineteenth-century suffrage reforms” (Ziblatt, 2017, 123-124). The argument that “authoritarian-led democratization” develops when “incumbent party strength” suffices to limit risk has since been generalized beyond Europe (Riedl et al., 2020, 18.2). But despite mentioning collective action often, this argument also reduces it to the actions of the few. The Primrose League is attributed to a few “wirepullers” conducting mobilization and a few aristocratic party leaders approving their conduct. Yet members of the Primrose League are said to provide “active services…required at the election as voluntary canvassers” (Ziblatt, 2017, 107). If so, joiners must exert uncompensated effort. This effort increases fastest among men benefiting least from the Conservative Party’s efforts to protect wealth. Wealthier males joining the Primrose League between 1884 and 1901 increase by a factor nearly equaling a hundred. Less wealthy males joining during the same years increase by a factor exceeding twenty-four thousand.

“Contacting” is by no means irrelevant to explaining why people vote despite the disincentive presented by the collective action problem. Indeed, good experimental evidence has been assembled that “contacting” or “canvassing” increases turnout (Gerber and Green, 2017). This evidence will require further consideration once a causal variable explaining democracy has been proposed.

In short, it is very hard to disagree with Geddes’s long-standing opinion that this literature can supply no valid theory of democracy or, for that matter, of dictatorship either. But maybe a different causal variable can. In shorter publications more recent than the mainly book-length arguments cited above, several theorists turn from the conflict between rich and poor to the question of how voters can overcome the incentive to abstain presented by the collective action problem. One theorist uses the metaphor “cogs” to propose that a sense of interchangeability with other participants in a collective machine potentially able to win elections can overcome reluctance to vote or, among fewer people in whom this sense is stronger, to protest even when protest is dangerous (Goodin, 2023). A pair of theorists show that moral values understood as shared with other voters reverse decisions to abstain motivated by considering only private gain confined to the individual (Alger and Laslier, 2021). This proposal revives a shift proposed earlier, and cited in the new analysis, from individual comparison of effort and gain to consideration of “moral values” shared by some group (Feddersen et al., 2009). Besides these proposals, one recent study finds that Americans underestimate, quite vastly, the effort that voting requires—“friction”—relative to the importance of political values—“beliefs”—that they share with others (Mazar et al., 2022). These proposals combine to raise two questions: how do people assign themselves to groups, and how does grouping themselves affect their estimates of effort?

3 Ending evasiveness: how talk enables collective action

No one can avoid the problem of collective action simply by discarding the assumption responsible for it that humans decide whether to act by comparing effort to expected gain over time. That assumption is a specialized form of the proposition that all mammals, at least, are optimal foragers. To discard the assumption underlying the collective action problem would imply either that mammals do not survive by optimal foraging, sometimes replacing migration to new locations by niche construction in a location already occupied (Zeder, 2012), or that humans are not mammals. The assumption has been validated in an enormous literature on the “Iowa card task” (Bechara et al., 1998). But while one may not abandon comparison of effort to gain, one may still consider its scope.

Humans are mammals that talk. That elementary observation makes possible a theory of how people can reevaluate the excess of effort relative to discounted gain that otherwise deters collective action in politics. Every democracy is a replacement for a preceding dictatorship that has ruled earlier inhabitants of the same territory or of some other territory from which ancestors have migrated. Every democracy does replace dictatorship when an earlier confinement of collective action to some minority of adults gives way to an expansion of collective action, gradual or sudden, that makes all adults eligible to join, at least in principle, although not all ever do. In the course of democratization, the nature of collective action changes. In a dictatorship, collective action consists of oppression by which some numerical minority imposes the ruler’s or the ruling council’s will on all other inhabitants of the territory under the control of the state. In a democracy, collective action consists at least of voting that chooses all those exercising unsupervised authority. Collective action that imposes dictatorship takes more effort from each individual participant—effort in the form of surveillance and violence—than voting in a democracy, which can take very little effort. But in both cases, the individual considering whether to exert any effort by joining in collective action confronts a choice between the effort to oppress or to vote and any expected net gain that might be incurred by joining. Both the potential oppressor and the potential voter must heavily discount the net gain from collective action by some very small probability that joining in oppression or voting affects whether the individual incurs the net gain. For the oppressor, the discount arises because the ruler allocating gain cannot monitor all the oppressors, paying only those who contribute or contribute fully, but any oppressor’s incremental contribution might still keep the ruler in power to continue providing the gain. That gain might be an emolument or maintenance of a property right from which the oppressor draws income. For the voter, the discount arises from the infinitesimal probability that the voter can affect who wins the election or the very small probability that the voter’s own ballot affects the description of the margin of victory. The explanans must answer three questions: how effort decreases relative to discounted net gain in both dictatorship and democracy; how an absolute decrease occurs in effort, in net gain, and probably in the probability by which the gain must be discounted as coercion in dictatorship succumbs to voting in democracy; and how eligibility to join in collective action spreads. If talk can explain those, it explains the replacement of dictatorship by democracy.

Even scholars often share the common supposition that language, the capacity to talk, exists for communicating with others. But Noam Chomsky (2002, 76-77) writes: “Language is not properly regarded as a system of communication. It is a system for expressing thought, something quite different…language use is largely to oneself: ‘inner speech’ for adults, monolog for children.” I understand Chomsky to intend that talk, most often silent, is one type of thinking. Some thought is not linguistic. The human visual system imagines a three-dimensional external environment by converting solely two-dimensional information that each retina supplies. The person acting in that external environment is unaware that any conversion has occurred, how the brain does it, how long it takes, or that the visual environment to which the person reacts is already about fifty milliseconds old when the person acts in it (Wolfe et al., 2006). This imaginative exercise is thought, and the visual system reports the results of its thought to Chomsky’s language system, where visual two-dimensionality informs the set of concepts presented by a language for spatial description. English speakers commonly know “vertical” and “horizontal” (Jackendoff, 1999, 15), but how many know the learned coinage for the third dimension forward? “Inner speech,” in turn, lets the thinker draw together not only information from the visual system but also from other brain activities, especially a brain activity keeping track of available effort and a separate activity evaluating potential gain or loss. Βut integration and coordination are not all that language does. Chomsky (2012, 71) says: “Language doesn’t just bring together information; it’s no recording device. We know this by introspection—you can think about what the world would be like if you chose this course and not that course.”

The capacity of talk to enable consideration of options, weighted not only by the brain’s assessment of effort and gain, makes possible collective action. When people talk about effort and gain, whether to themselves or to others, they systematically either underestimate or overestimate the effort required to accomplish a task. Very abundant experimental evidence collected using five separate experimental paradigms shows people accurately perceiving the effort needed to accomplish a very wide variety of tasks. But the same experiments also show the same people inaccurately describing the same effort needed to accomplish the same tasks (Anderson, 2018, 40–44). If people decide what to do by describing themselves to themselves as potentially needing to exert effort, inaccurate descriptions will displace accurate perceptions in shaping decisions about whether to exert effort.

In these experiments, some self-descriptions induce people to overestimate the required effort. For the purpose of explaining collective action, overestimating the required effort makes no difference. When others may produce a gain that the individual cannot produce alone, accurate estimates of effort already suffice to motivate abstention. Overestimating effort does not alter the choice to abstain. But underestimating effort might.

In the experiments, individuals underestimate effort when they describe themselves as occupying, or more able to cross, a greater extent in space—that is, as more sizable. While any such description can be expressed by multitudinous synonyms and metonyms, descriptors of extended spatial occupancy are found across many, very disparate languages and potentially in all (Wierzbicka, 2021, 326). Size descriptors may express physical or metaphorical extent, but if people understand metaphors by recalling other contexts in which the metaphorical term has previously been encountered (Davidson, 1978), then either physical or metaphorical usage of such descriptors will cue an underestimate of expected effort.

Self-description as sizable exerts a dual effect on estimates of expected effort. This duality of effects owes to the common property of any linguistic descriptor that all of them describe an individual item, human or not, by assigning the item to a category that includes other items, real or imaginary. Self-description as sizable, therefore, reduces the individual’s estimate of expected effort both by (a) assigning more size to the individual and (b) categorizing the individual as a member of a group, real or imagined, that collectively occupies more space than any of its individual members. The experimental literature teases apart these two separate effects of self-description by revealing that estimating effort together with a cooperative associate, or even while thinking about a cooperative associate, or even while thinking about a divinity, independently exerts the same or similar effects on underestimating effort as describing the self as sizable. Together with a cooperative associate, the individual occupies more space than the individual alone. One reported experiment within this paradigm surprises its own authors when an instruction to perform a task in cooperation with another person, rather than diminishing estimates of required effort as expected, increases the estimated effort instead. However, the authors design a task that is doable for each of the young experimental subjects acting alone but awkward when done as a pair. Since the awkwardness of a task is known to increase estimated effort, the authors substitute a task that is awkward for an individual acting alone but more convenient for a pair. That substitution eliminates the supposed costs of cooperation (Meagher and Marsh, 2014). Cooperation can, of course, make a convenient task more difficult, but it does so when one or both agents are uncooperative.

Self-description as sizable not only propitiates collective action but also restricts the proportion of individuals who will join collective action. Limitation to a minority owes to a semantic property common to expressions of size. Any categorization as sizable remains coherent only if it divides a population of items into a minority eligible for the categorization and an ineligible majority. Eligibility for the categorization belongs not to those exceeding the mean on the scale of relative size mentioned by the descriptor but only to a minority exceeding the mean to a degree that makes the eligible minority “stand out” relative to others in the relevant population (Kennedy, 2007). All whales qualify as large mammals because whales compose a tiny minority of mammals, but only a minority of whales can qualify as large whales.

Confinement of collective action to a small minority establishes dictatorship and activates its defining feature, oppression. Experiments present subjects with contrasting cues, linguistic or extralinguistic, that divide samples into two groups assigning themselves either cue. These experiments regularly result in decisions to award additional resources to persons marked with the same cue as the decider and to award such persons praise denied to individuals marked with the contrasting cue. In a comment particularly telling for enfranchisement, Marilynn Brewer (2012, 12) comments: “competition over material resources…has as much to do with the identity meaning of those resources as it does with actual group survival.” Since individual organisms invariably compete for resources and praise is a resource for humans that stops being praise if distributed equally, outside an experiment, no small minority can redistribute resources or praise in favor of its own members except by coercing the other group. Coercion succeeds because the majority, not describing its members as sizable, remains deterred from organized resistance by the excess of expected effort over discounted gain attendant on collective action in general. Acquisition, retention, or redistribution of wealth is not the goal of authoritarian coercion or the objective of enfranchisement; they are by-products of the discursive division of a population into minority and majority.

The excess of effort over expected gain need not prevent all resistance by the oppressed, since “oppressed” is itself a category capable of motivating collective action. Still, the categorical motivation of the oppressed to join collective resistance is counteracted by self-description with the many synonyms or metonyms expressing lack of size, all of which cause overestimation of the effort required for resistance. Because self-description as lacking size counteracts the encouraging effects of categorization as oppressed, resistance to oppression tends to take individual rather than collective forms: sabotage, property damage, or desultoriness at the margin from the labor of cultivating, extracting, refining, or crafting the resources that oppressors redistribute to themselves. Consequently, as long as minorities continue to assign themselves a descriptor expressing sizableness, oppression tends to persist. Oppression motivated by spatial self-description, in turn, motivates personification of the oppressor group in one or a few individuals. Size descriptors are scalar and polar (Kennedy, 2007): they rank individuals along a scale toward a positive pole. Scalarity and polarity accordingly imply a description of someone as most sizable and of a sub-minority among the oppressors as more sizable. The oppressor described as most sizable becomes the dictator, and those implied to be more sizable become the dictator’s council.

Of course, this entire discussion is no more than extended theoretical speculation, but it offers a coherent account of how collective action can establish either dictatorship or democracy. Categorizations retaining coherence only if they restrict to some minority of a human population motivate the oppression that installs and sustains dictatorship. Dictatorship erodes if those categorizations fade over time, rapidly or gradually. Democracy emerges as categorizations appear that can describe all or most members of some population. Finding and assigning such potentially uniformly pertinent descriptors can be quite tricky, a difficulty that accounts for constant infighting in democracies that must be continually resolved for them to remain democratic. Still, discursive conditions for democracy have become plain. The argument presents a simple hypothesis that can be tested for any state: if a dictatorship, then self-description of oppressors as sizable; if a democracy, then self-description with categories capable of embracing a whole population.

Now to British enfranchisement.

4 Enfranchisement in Great Britain

4.1 Possession or discourse: the collective action problem

A discursive theory that linguistic descriptors can activate a collective self re-reads textual evidence about the origins of enfranchisement. To show how linguistic descriptors affect enfranchisement, I may re-examine enfranchisement in Great Britain, which remains a paradigmatic case for democratic theorists, has exerted a powerful influence worldwide, and has been carefully examined in a recent book by one theorist discussed in the opening section.

As evidence for his assumption that individual wealth maximization animates contention over British enfranchisement, Ziblatt (2017, 68-69) offers a quotation from a landowner and conservative politician who later becomes Prime Minister—Robert Cecil, later Viscount Cranborne, eventually Marquess of Salisbury (1830–1903). In 1860, Cecil writes that the controversy between “the English constitution” and “democratic forces…laboring to subvert it” expresses “a struggle between those who have and those who have not” (The full excerpt is numbered “1” in Supplementary material). Cecil’s statement can, of course, be read with Ziblatt either as an accurate statement about 19th-century English politics or at least as a frank expression of Cecil’s understanding of English politics in his era. But like any utterance (Austin, 1962), Cecil’s utterance also performs at least one additional speech act. Here, the second speech act done by a first act performing description is an act performing division. It divides a population into two categories, “those who have” and “those who have not.”

The divisive speech act resolves the puzzle concealed by the descriptive speech act. If Cecil is correct that all participants in English politics are either trying “to keep what they have got” or “to get what they have not,” why is anyone struggling? Struggle takes effort. As individuals, every individual is known to confront an opportunity acknowledged by Ziblatt when he repeatedly mentions the collective action problem that he alleges the Conservative Party has solved. Each individual can either keep what the individual has or get what the individual has not by allowing others to keep it or get it for the individual as a consequence of keeping or getting it for themselves while each individual devotes his or her own effort to get something else, too. But if individuals categorize themselves as members of groups, they reduce projected effort, and the reduction causes at least many individuals to join in collective action either to keep what they have or to get what they have not. Since either group tries to prevent the other group from succeeding, a struggle happens and is plainly evident in the record.

4.2 Presentation and methodology

Many other addresses to Parliament in the 19th century repeat Cecil’s division by describing those “who have” as occupying an extent in space that “stands out,” in the sense used by Kennedy (2007), and denying it to “those who have not.” To find those descriptors and to examine whether they fade and are replaced by descriptors that do not inherently restrict to a minority, I have conducted repeated keyword searches of the digitized text of Hansard¹, the record of Parliament. For the 19th century, when enfranchisement spreads, Hansard does not provide a verbatim record. Its record instead consists of detailed paraphrases. Still, the speakers themselves are frequently reported as claiming to have consulted Hansard’s earlier records. If politicians rely on Hansard to decide what to say, its paraphrases count as political discourse.

A keyword search necessarily undercounts descriptors that assign persons to categories. In English, such descriptors are often attributives or predicatives. In either attribution or predication, intervening lexical items or punctuation may separate such a descriptor from the noun it modifies that mentions a human entity, whether singular, plural, or collective. Additionally, English features a suppletive plural for a singular individual, “people,” which is also a singular noun referring to a group. A description of the English or the British with the collective singular “great people” is obviously not comparable to the use of the suppletive plural “great people” to describe the minority of the English or British who belong to aristocratic families, but there is no keyword search that can separate these two usages. Furthermore, figurative language cannot be anticipated, and I have searched only for the most common spatial descriptors, not for all their multifarious synonyms. More sophisticated computer programs for text analysis and search might alleviate some of these shortcomings. My searches examine debates for the period from January 1, 1800, when Hansard’s summaries became more comprehensive, to December 31, 1911, when the House of Lords’ loss of its veto over legislation passed by the House of Commons registered the triumph of Cecil’s “democratic forces…laboring to subvert” “the English constitution.” Below, I comment on what I have found. Each time I comment, I provide a number in parentheses, sometimes with a modifying alphabetic character, that refers, with one exception, to a condensed excerpt from Hansard. The condensed excerpt can be read online in Supplementary material. Item (1) is quoted from Ziblatt (2017); all others can be found in Hansard on the date shown. The condensations use ellipses to replace extraneous matter that I have removed as diverting attention from the categories that are my concern here. Necessarily, condensation sometimes distorts clausal meaning, which should be sought in Hansard’s own unedited texts.

Keyword searching might be deprecated as unsystematic “cherry picking” relative to modern techniques for processing enormous quantities of text, such as Hansard. A prominent technique is Latent Dirichlet Allocation (LDA). Relying on the discipline of linguistics as the inspiration for a research agenda, I have grave reservations about LDA and its output, the word cloud. LDA explicitly treats a text as a “bag-of-words” in which it uses frequency to assign probability of occurrence and proximity to assign association. However, while linguistics uses frequency for certain, very restricted purposes, it views text as resulting when words, assembled from a much larger lexicon consisting of morphemes and word components, are combined according to syntax. Having counted words, LDA identifies “topics” by assuming that the most frequent words encountered in proximity to the most other words change as “topics” change. This assumption uses a common-language understanding of “topic” as mentioning what is discussed, whereas the same word has a very specific, quite different meaning in at least some branches of linguistics. What is being discussed can vary enormously even when the words remain the same—as in the meaning of “great people” above.

My reservations about LDA, however, are not why I eschew it here. Even if I accepted its claim to identify topics, the theory that I propose to test concerns how speakers categorize themselves in contrast to others, not any “topic” they discuss. I expect these categories to prevail regardless of the topic, either in the general sense of LDA or the technical sense of linguistics. Concentration on self-categorization leads to very different conclusions than attention to LDA topics. To see the difference, I urge the reader to compare what I have to say below about the spread of primary schools in 19th-century England with the use of LDA to study the same development (Smith, 2023). If elementary schools were “an effective policy instrument of greater cultural rationalization,” why did the opponents of enfranchisement struggle so desperately, and so effectively until self-categorization changed, to retard the spread of schooling?

4.3 Enfranchisement vs. democracy

Even within the British Isles, at the end of 1911, enfranchisement remains incomplete, but it takes the outbreak of World War I in 1914 to delay the enfranchisement of men and some women until 1918 and requires a partly victorious Irish revolt for independence to delay universal adult suffrage, confined within the British Isles without most of Ireland, until 1928. Enfranchisement within the British Isles passes through three stages. The so-called Great Reform Act of 1832, which enfranchises many voters in England and Wales, is accompanied by disqualifications of others that result in net disfranchisement. Although the estimated electorate for the House of Commons in England and Wales increases from 14.4 percent of adult males in 1831 to 18.4 percent in 1832 (O’Gorman, 1989, 179), since 1801, Irishmen have voted for Parliament too, and enforcement of the Act in Ireland produces a net reduction of the total electorate, especially compared to the period before 1829, when Irish disfranchisement begins (Hoppen, 1985). The first net enfranchisement is the Second Reform Act of 1867, introduced, unexpectedly at the time, by a Conservative Government. Much more comprehensive enfranchisement was secured by a Third Reform Act in 1884, this time by a Liberal Government. Up to 1911, unelected Lords can veto whatever bill any majority elected to Commons enacts. The conversion of their veto into a delay suspending for 2 years Acts not declared budgetary enormously increases the importance of the franchise.

Enfranchisement within Britain does not achieve democracy. Both Lords and Commons insist, as in (2b) below, that the English Parliament is an Imperial Parliament exercising a right to intervene throughout an empire ultimately ruling over about a quarter of the world’s inhabitants. The distinction between partly enfranchised Britain and disfranchised Empire is plain to contemporaries. Four years after the Second Reform Act’s limited enfranchisement, the Secretary of State for India writes about combining “an Imperial and a Democratic policy; in one part of our dominion…the widest popular suffrage; in another…a powerful administrative despotism supported by military force” (Mowat, 1973, 111). For the remnant of the United Kingdom to become a democracy with nearly everyone eligible for enfranchisement, five decades would be required from the 1928 enfranchisement of women on equal terms with men to Parliament’s abandonment of its sovereignty over the remaining disfranchised Imperial populations. Even then, only in the next millennium would hereditary Lords finally be silenced.

4.4 Spatial descriptors restrict the franchise

Of course, no single statement by Cecil or anyone else can suffice by itself to divide England’s population into active and inactive groups. But Hansard presents an abundance of restrictive spatial descriptors falling into at least five categories.

The first category describes certain individuals as great. In today’s English, “great” seldom occurs in spatial contexts, almost always acting as an emphatic or intensifier. But Hansard is reporting English of the 19th century, when “great” had yet to disappear from spatial contexts. In fact, that very disappearance from spatial contexts may owe, in part, to the processes I describe here. In Hansard, “great” occurs as a spatial descriptor of two kinds of political figures, exemplified in the contrast between (2a) and (3a). The spatial reference of “great men” is definite in a contrast, rare at the time, with “little men” in (2b). Synonymy with the aristocracy is evident in allusion to the Duke of York, holder of the highest aristocratic rank. “Great men” in (2a-b) are also defined by the subordinate clause in (2c) as sitting in Parliament not by the will of any electorate but by “personal right.” By contrast, the great man in (3a) is defined in (3b) by combining “humble origin” with “talents” and “virtues.” The two categories of great men define a coalition that protects against the menace of collective action in the form of “combinations of journeymen”—at the time, workers paid by the day—whom “agricultural laborers” might join. Passage (2a) is also specific that “great men” can constitute only a minority, “the few.” The great men in (2a-c), to the exclusion of those in (3a), are also identified by (4a-b), qualifying by belonging to a great family for appointment to high office.

The categories in examples (5a–e) are “large landowner” and “large proprietor.” These exemplify “syntactic modification” (Berg, 2011, 726). The first element of the compound “landowner” mentions what is large, the land, but the categorizing effect of the adjective “large” skips over the adjacent noun initiating the compound to modify the second noun “owner,” which itself decomposes into a verb that the adjective cannot modify, transformed into a noun by the agentive suffix “–er.” Similarly, in the non-compound “proprietor,” the element “property” is large, but the categorizing effect skips to the agentive suffix “–tor.” Of particular importance for England in the 19th century, only to someone educated in Latin word formation is the latter effect of syntactic modification visible.

In (6a), the speaker replaces a size descriptor with a formal term derived from the Latin for “equal” (OED, “peer”), capitalized to equate it with the “great men” of (2a). In (6a), this speaker assigns the Peers no size descriptor, reserving “large” for the expanse of land that they control. In the immediately following rebuttal (6b), the first speaker’s opponent uses the term for the possessor of land, “Peer,” interchangeably with a description of the possessor as “large.” Passages (6a-b) occur in 1846, 15 years later than the last of those in (2–5), a timing that will deserve further consideration in the next section.

The fourth category is a metonym for physical size, “high.” This adjective is metonymic of terms for extent in space if, in common experience, things that are visibly similar in width are also larger if they are noticeably higher. Accordingly, “high birth” in (7a-b) acts as a synonym for “great family” in (4).

A fifth category consists of descriptions of men as “big.” In contrast to the other four categories of spatial descriptors, neither “big man” nor “big men” occurs in a political context until 1866 (although “big man” does appear once in 1820 in an apolitical mention of physical size). Passages (8a-b), the first and third occurrences, are notable for the absence of the invidious distinctions so plain in (2a), (4), (5a-b), and (7a), while in (8c) the disparity between “big salary” and “do the work” reverses the invidious distinctions drawn in (2, 4, 5, 7). This pattern of either absence or reversal of invidious distinctions persists in almost all cases of the use of “big man/men” during the four and a half decades from the first occurrence to 1911, even when the speaker paraphrased in (8d) is Lord Robert Cecil, son of the then recently deceased Marquess of Salisbury, objecting to a proposal by a Liberal Government.

As late as 1845, an organ of the emerging Liberals, the Westminster Gazette, summarizes the effect of the first four categories of spatial descriptors—by 1845, “big man” is yet to join the list.

“The government of this country resides in a minority; the owners of land, a small proportion of the community, constitute, almost exclusively, one House of Parliament and form the large majority of the other (Middleton, 1970, 167)”.

4.5 Recontextualization promotes temporizing enfranchisement

How can this one-sided struggle produce enfranchisement if self-description as great, large, or high promotes collective action among oppressors while the opposite self-descriptions demobilize the disfranchised? For enfranchisement to occur, descriptions must change. If English political discourse continues to divide possessors from the dispossessed, any protesters demanding the vote should remain scattered and ineffectual and continue to lose any struggle. But already by 1860, when Cecil publishes excerpt (1), a new element is beginning to penetrate English political discourse. This new discourse both promotes a slight enfranchisement in 1867 and sets the stage for further discursive change that motivates the larger one in 1884.

Between the Great Reform Act of 1832 and the Conservative Reform Act of 1867, the spatial descriptors that have motivated collective action among oppressors change in two respects. First, the contexts in which these descriptors occur become ambivalent. Instead of mention followed by praise, as in (2, 3, 4, 7) above, positive spatial descriptors begin to presage qualification, reservation, objection, or even criticism. Decisive evidence of the new tone is a comparison of (9a-b), in which each “great man” mentions the same actual man, the Duke of Wellington. Praised in 1832 for withstanding “the democratic torrent” and still the “greatest hero of his age” 11 years later, in 1843, his skill at warfare no longer entitles his views to such “overpowering authority” regarding “internal affairs.” By 1863, “great men” have accomplished “scarcely anything” that “may not be caviled at”—and the speaker is the Conservative Benjamin Disraeli (9c).

Second, although discourse preserves the old spatial descriptors organizing oppressors against disorganized oppressed, those descriptors are now subsumed in an existing label respecified to encompass both. The Crimean War motivates Viscount Palmerston, head of a Liberal Government, to speak in defense of English great men. To defend them in (10), he denies any difference between peer and trooper, between those highest and those lowest! Others imitate him in (11a-b). “For the first time,” the Queen herself, who acts only according to guidance from her ministers whom Palmerston leads, honors equally “men of high birth” and “the humblest soldier” (11a). The sight of “evident superiority” without specification of social rank beckons even a South Asian native child to learn English and travel to England (11b). Fourteen months later, Palmerston shifts his emphasis to lauding the troopers while relegating praise for their officers to an afterthought (12). “Englishmen” now share “every quality which gives dignity.” Palmerston’s usage intrudes into prolonged debate in 1858 responding to South Asians’ uprising for independence (13a-f). In (13a), Palmerston ascribes “great and manly qualities”—certainly reminiscent, at least, of “great men”—not to aristocrats but to “our race” and “the people of this country.” In (13b), Disraeli echoes Palmerston’s own 1856 usage with “Englishmen…whether officers or privates” before launching into a segue blaming Palmerston’s appointee as Governor General, Lord Canning, for having been taken by surprise when an insurrection erupts. Variants of this equation of ranks are found in (13c–e).

To this transformed discourse, the onset of “big” in 1860 (8a) might tentatively be assigned. “Big” now appears in a context in which “great” has appeared previously—e.g., opposite “little” in (2b). The OED reads 19th-century occurrences of “great” as denoting “size, bulk, or extent that is considerably above average,” but occurrences of “big” are said to denote merely “relative size…more or less.” Such an interpretation of the onset of “big” is reinforced by the tendency of its usage, evident in (8a), to equalize “big” with “little,” which, when done to “great and little” in (2b), is a contested “duty.” Substitution of “big” for “great” seems to reduce relative size. However, there are not enough occurrences and too many alternative definitions for this assignment to be confident.

When both Conservative and Liberal leaders subsume “high” and “low” within “Englishmen,” shared discourse enables a new enfranchisement in the Second Reform Act of 1867. Coalition politics delays enactment. In (14a-b), a Reformer in 1861 and a Conservative in 1865 agree that successive proposals to enfranchise have foundered on factional strife among Liberal supporters. But factionalism that paralyzes Liberals energizes Conservatives. If contrasting categorizations incite contestation, so do contrasting subcategories within each category. Archetypes of the “great men” in (2) and (3) emerge as rivals to occupy the positive pole on the scale of Conservatives. Cecil/Cranborne is the archetype of “men of great family and much parliamentary influence” in (4b). Disraeli no less personifies the man in (3b) “humble in origin” but rising “by the exertion of his talents and the display of his virtues…to the highest offices of the state.” To remain head of Government when Conservatives compose only a minority of Commons, Disraeli proposes a Reform Bill. It should attract the “regular Liberals” of (14b) but also gains assent from John Bright, the Radical alluded to in (14b) as desiring “the institutions which prevail in America.”

Cecil/Cranborne protests by abruptly resigning in March 1867. A Radical asks why (15a). Cecil waits 4 months to answer (15b). But two conditions doom his opposition. First, in (15b), the new discourse of “Englishmen” demands from him a concession: the new “enormous majority” composed of “the working classes” will usually do “no harm,” as it will normally agree “with all other sober and reasonable Englishmen.” Second is a British version of that quintessentially American device, the gerrymander. A Liberal objecting to the Bill describes it in (16). Not only will districting concentrate sixty thousand voters favoring Liberals in one district in a large town while limiting small-town districts favoring Conservatives to seven or eight hundred voters each, but moreover, the latter voters will reject collective action (casting unrewarded votes that express their own preferences) in favor of private action (selling their votes for bribes paid by wealthy Conservatives).

4.6 Unrestrictive categorization promotes continuing enfranchisement

Further enfranchisement will require further changes in discourse. Both the Liberal’s and then the Conservative’s “Englishman” remain imprecise. It includes both high and low, but how low does it reach? Does the franchise chosen in 1867 coincide with the downward reach of “Englishmen”? Three more changes in discourse will move the boundary to encompass more of the United Kingdom’s population before another enfranchisement in 1884. These changes concern whom to describe as white, what to do about dialect, and whether to persist in restrictive categorization. The intersection of schooling with the description of more Englishmen as white adds to the importance of enfranchisement with the elimination of the Lords’ veto in 1911.

4.6.1 From whites abroad to white Englishmen

“White man/men,” “white person/people,” “white race/s,” or just “whites” occur frequently in Hansard from early in the 19th century, but until 1850 these terms mention no one in Britain—with three exceptions. Even (17), for its date a very unusual comparison that likens the flogging of British soldiers to punishments inflicted on black slaves, mentions “whites” only as settlers endangered by slave uprisings in the Caribbean. Other early uses in (18a–e) also restrict whiteness to persons overseas. The categorization of whites residing in the Caribbean in (18a) projects onto them the distinction between high and low in England. It also projects the anxiety, evident in (2a), about allowing the low to obtain political influence, as well as the customary praise, explicit or implicit, attributing benevolence to the high in England. In these regards, (18a) is more a restatement of discourse in England than any comment about conditions in Barbados. Excerpts (18b-9) parallel these assignments of superiority and benevolence, although over time, attributions of superiority and benevolence that originally separate whites into two categories begin to separate whites from blacks, still overseas.

Although the examples in (18a–e) are selected from many occurrences, I have found only three instances before 1850 in which “white” assigns color to any resident of England (19a–c). Passage (19a) compares the legal status of “white” child laborers in England unfavorably to that of adult slaves in the West Indies. Passage (19b) compares the opportunities for whites “born” to labor in England unfavorably to those available to blacks in Jamaica. Passage (19c) compares proposed rules for feeding whites in England, confined to workhouses for the unemployed, to the hypothetical subjection of blacks overseas to the same mistreatment. All three find anyone white in England only in comparison to blacks abroad.

In connection with the 1867 Reform Act, in (20a–f) the reservation of white color to Englishmen abroad progressively blurs. Progress toward the Reform Act, delayed throughout Palmerston’s ascendancy, advances after Disraeli’s concurrence on 8 February 1858 with Palmerston’s 1856 egalitarian description of Englishmen. Four days earlier, the South Asians’ uprising for independence that offers Disraeli an opening also introduces a new category not previously encountered: “British-born white subjects” in (20a). Even so, they are still to be recruited from the colonies. But by 1870, “We were a white people” appears in (20b). It is impossible to tell from Hansard’s paraphrase whether “were” is the speaker’s original subjunctive or the reporter’s consistent past tense. Either way, “white people” who impose a tax on gin in England cannot morally reject a tax on opium in India. In 1876, even after sepoys rebelled in 1857, Cecil/Cranborne issues the warning in (20c). Not “resistance from the subject races” imperils “English rule in India,” but splits “among men of the white race…the White rulers” who comprise both “mercantile communities in India” and “manufacturing communities in England.” In (20d), the speaker compares “untutored savages” with “we White men” in adherence to “a great and noble virtue”—love of country—but attributes only to “we White people” the expectation that contact with “Black savages” would end in the disappearance of the latter. By 1883, in (20e), “they” changes referent without notice. The first “they” mentions members of Commons in London; the second “they” mentions Englishmen in South Africa “fewer in number than either” Boers or Blacks but partaking of “the blood and treasure of the unfortunate English nation.” The two instances of “they” merge not only by unnoticeable change of reference but also by the parallelism of nearly synonymous predicates “make themselves responsible” and “undertake,” both introduced by the same future progressive. In (20f), in the same year, a Bill in Parliament has “stimulated race antagonism to an extent without parallel since the evil days of the Mutiny…We in England can barely imagine the strength of the indignation…We won India by the superior qualities of the English race…indomitable…unyielding…unabating resolve” to become and remain “the White rulers.”

4.6.2 From dialect to schooling

As Englishmen at home gradually whiten, more children attend school. The Liberal whose statement in (16) has opposed Disraeli’s advocacy of enfranchisement to maneuver against Cecil has also seized the opportunity to switch the topic to education with (21) from the same speech. In 1870, Parliament passes a new Elementary Education Act. Excerpts (22a–c) describe it. It is (22a) “only a compromise.” It lets communities of “working people” elect school boards to operate “Board schools” funded by fees, local taxes, and Parliamentary subsidies (22b), but only in locations where “voluntary education,” funded by fees, charitable endowments, and the same subsidies, “was found to be deficient” (22c). Almost all of the voluntary schools are said to be organized by clergymen of the established Anglican church, and, as in the case of an unnamed “Duke in the neighborhood” in (22b) who personifies the clergymen’s allies among “great men,” the Conservatives try to obstruct the opening of new Board schools. After only five years, the Conservative in (22c) asserts that the Act already “has been most successful…has really produced such great results.” To the Liberal in (22b), the results are no more than “slow growth,” assuring that “most of them would never see the fruits of their labors.”

The compromise of 1870 expresses the influence on enfranchisement of circumstances described by Joseph Fisher (1986, 2):

“From north to south and east to west, each village can understand its neighbor, although at the extremes the speech is mutually unintelligible…The fixing of sharp linguistic boundaries in Europe began with…the self-conscious creation of official written languages. As these written standards emerged, the dialects continued to be spoken”.

Although Fisher’s contrast between an official written language and multiple dialects remains muted in Parliament’s discourse, both Houses exhibit awareness of that contrast in Britain. Each village speaks its own dialect that allows communication with nearby villages. As Britain urbanizes, varying contact languages develop in each town or city as migrants from surrounding villages attempt to converse with each other. Migrants from adjacent villages can converse easily, but their dialects are unintelligible to migrants entering the town or city from villages in other quadrants surrounding the urban area. As early as 1832, the commitment of Commons to written standard English is contrasted with a sign of dialect in the three-way exchange (23a–d). In (23a), one member mocks another’s non-standard pronunciation, “Iddication.” The target of mockery replies with (23b), in which he feigns modesty about his mastery of the written standard “the King’s English,” but only as preparation to reproach his critic with “vulgarity of a pauper and insolence of a demagogue.” In (23c), the Speaker of the House intervenes with a politely indirect insistence on an apology. But it is not demanded for the original mockery. Instead, a “gross violation of order” has been committed by using lexicon from the King’s English that custom permits no Member to direct at another. Since the offending member’s compulsory apology in (23d) repeats “vulgar,” apparently only “insolence of a demagogue” is out of order.

Examples (24a–f), selected from a very few more, further demonstrate awareness of the gap between the written standard used in Parliament and the dialects or urban contact languages used by most inhabitants of England. In (24a), the speaker is the Earl of Ripon, a former head of Government who has since occupied other senior posts. In standard English, the Earl even claims familiarity with “the language of the peasantry” in Yorkshire, where his estates are concentrated. At the same time, he denies that being “ignorant of the Yorkshire dialect” should ever disqualify any “gentleman” from appointment to participate in administration there. He uses “language of the peasantry” and “dialect” as synonyms. “Dialect” is also used in (24b) for what today would be described as written language, in a generally accurate description of the origins of the written standard that differentiate it from both dialects and urban contact languages less influenced by Gallo-Romance. In (24c–e), dialect is reduced to “pronunciation.” Dialects do, of course, vary in their phonology, but they also vary from a standard language and each other in multiple other respects. While reductive, these passages also reveal awareness of the north–south and east–west clines that divide maps of English dialects (Britain, 2009, 126). Differences in dialect, reduced to “pronunciation” in (24c-d), hamper learning to read. In (24f), the standard language, found in the Book of Common Prayer that is used “in reading” to “the people of his parish,” leaves those people “unable to understand his arguments” because of “their slowness of apprehension” and “scanty vocabulary.” Reduced processing speed and lexical non-recognition are earmarks of any encounter between one interlocutor who speaks standard and another who speaks dialect when neither speaks both. So is the disdain exhibited by the speaker of the standard language toward the dialect speakers to whom the speaker of standard English imputes “want of general mental culture.”

When the Elementary Education Act would close this gap separating political discourse from dialect or urban contact language, Conservatives gain by frustrating educational progress to keep the gap as wide as possible. The passages collected in (25) document the conflict. In (25a-b), two Liberals describe the situation in 1862 and 1867, before Disraeli enfranchises. They attest to what even the Conservative speaker paraphrased in (22a) will later call a “great educational crisis.” Children of the “laboring poor” do not stay in school long; a third attend no school at all, and the speakers estimate that a “good education” reaches as few as one-ninth or no more than one-third. While the 1870 Act does open more schools, progress in education remains painfully slow. Not only dialects and urban contact languages hamper learning to read.

Seventeen years after 1870, a Liberal identifies one persistent obstacle as classroom instruction. In (25c), he objects to the teaching of English as spelling and grammar. For speakers of dialect and urban contact languages, arbitrary spellings that do not even correspond regularly to the pronunciation of the written language complicate the task of learning to read it even more. Dialect speakers are accustomed to pronouncing similar words with quite different phonology and to expressing themselves with a lexicon that, not found in the written language at all, lacks any spelling. Instruction in grammar adds an even more daunting impediment. In 19th-century English, “grammar” is encountered in contexts pertaining to Latin and Greek or even Hebrew, as in (25d). “Grammar” includes Latin word composition (25e), “concord between verb and nominative case” (25f) that is all but non-existent in English, and rules learned by rote (25f) because they are inapplicable to English, especially “accidence” (25h) that in 1858 still mentions conjugation and inflection, which are elaborate in Latin and Greek but very sparse in English. Grammar derived from ancient languages “is useless for speaking; it is equally useless for writing or composition” (25c). Possibly adverting to dialect, the speaker reduces grammar to teaching no more than to use “‘I go,’ instead of ‘I went,’ and the like.” To break the deadlock, the speaker of (25c) recommends the transfer of authority over the curriculum to “localities” where elected Boards can ease the task of learning to read for at least the pupils in the schools they administer. Nevertheless, objections to instruction in reading that make it a “torture to children” and to the teaching of “grammar” recur in 1888, 1890, and 1898 (25i–k).

Liberals also allege overt Conservative resistance. Even if Board schools open in more localities and Conservatives defend against their intrusion by opening more voluntary schools (22c), pupils must attend to learn. In 1882, on the day when an inspector visits a school, the average number of pupils attending is 72 percent of those enrolled—an all-time high (26a) achieved by a law in 1880 making attendance compulsory. By 1897, educational subsidies introduced 5 years earlier have “raised the average attendance…to 81½ per cent,” but “especially in the rural districts, the compulsory attendance clauses were scarcely enforced at all…a good many magistrates had apparently made up their minds not to enforce them” (26b). Refusal to enforce compulsory attendance at school is attributable to who the magistrates are. In 1893, “Magistrates…were only landowners, and of the landowners only those of one Party and one Church” (26c). In 1894, even legal training does not qualify men for “appointment as Magistrates” in preference to “a large landowner, or the son of a large landowner” (26d).

In 1883, it was already expected that even if full attendance could be achieved, educating everyone would still take two or three more decades (27a). In 1887, a Conservative government introduced legislation predicted to “surrender the schools of the country districts into the hands of the parson” at whose behest “the children were taught to fall down and worship the great god of the Clerical party—the landowner.” Church schools left illiteracy widespread in villages 17 years after the 1870 Act (27b). In 1894, magistrates used parental concerns as an excuse not to enforce attendance. School inspectors insisted that many pupils could not read adequately. Continuing neglect would prevent education ever being achieved (27c). In contrast to a Liberal MP’s own rural district in 1896, in other rural districts “a large proportion of the electors could not even read the names of the candidates” (27d). In 1897, rural children were “miserably educated,” sometimes “not able to read or write” (27e). The requirement for the Education Department to approve all curriculum changes, mentioned in (25i) as its “duty,” implies that officials appointed by the Conservative ministries in power after 1886, with one short exception, encountered abundant opportunities to delay the removal of impediments to learning.

The Third Reform Act in 1884 ended Liberal dominance in the Commons. Enfranchisement spread but fell short of doubling the proportion of adult males qualifying to vote in England and Wales, a proportion encompassing more than 60% but less than 70% in 1885 (Smith, 2004, 171). The Act did establish a uniform franchise throughout Britain and Ireland. Presumably unbeknownst to the Liberals who enacted it, uniform enfranchisement handed a decisive electoral advantage to the Conservatives.

The Conservatives lost the first election under the new rules held in late 1885 because they could not organize fast enough to exploit their advantage or did not recognize it soon enough. Though the Liberals won, they needed Irish nationalists for a coalition capable of securing a majority. Through his son, Gladstone floated an indirect endorsement of the Irish deputies’ demand for a separate Parliament to pass laws for Ireland. His endorsement caused many Liberals to defect. Without a majority, Gladstone had to call a new election in July 1886. With tacit support from the defectors, who called themselves Liberal Unionists, Cecil/Salisbury formed a minority government. The Conservatives ruled with only one interruption of less than 3 years for two more decades.

Already during the 1885 elections, Conservatives used to winning rural districts wondered what to do with those agricultural workers whose occupancy of separate cottages now qualified them to vote. Very few records survive from the 1885 elections. One study of those records inserts “[sic]” into advice given to a landowning nobleman. He is urged to hire “a paid laborers’ agent who can take talk [sic] to the laboring class in their own language” (Blaxill, 2011, 103). But if the agricultural laborers’ “own language” is unwritten dialect, “take talk” is neither a misprint nor a tautology. “Talk” is what men who cannot read need to hear, and someone must “take” it to them if they are to hear it. Just before the 1886 election, The Fortnightly Review generalizes this advice, urging that Conservatives “call together all the employers of labor…the masters and the higher order of farm servants” for the purpose of communicating “information and arguments,” originating with the party’s parliamentary leadership, “individually among the laborers” (Lynch, 2003, 54). If this last-minute publication recapitulates advice that has been circulating privately since the 1885 defeat, Conservatives can take advantage of it at least much more often than Liberals.

Here is where the experimental literature on “canvassing” in the United States at the turn of the millennium (Gerber and Green, 2017) pertains to late 19th-century England. “Canvassing” is not what influences people to decide to vote. It is a metaphor that compares canvassers spreading out through a district to canvas spreading out over a surface. What influences potential voters is not the spreading mentioned in the metaphor but the conversation that results when spreading brings the canvasser close enough to talk with the potential voter. To converse, the canvasser and voter must share a language. In a United States with universal elementary education, canvassers are likely to find people with whom they do share a language that allows them to converse about politics. In 19th-century England, the “masters and higher order of farm servants” whom Conservatives are urged to recruit can talk to agricultural workers contemplating whether to vote. Liberals much more often cannot. The masters employing agricultural labor are ordinarily “farmers” who are also “tenants” paying rent to “large landowners” or “large proprietors,” mainly aristocrats, who, if they farm their own lands at all, assign farm management to “agents.” Farmers or agents speak standard English to their aristocratic landlords, but they, or the “higher order of farm servants,” to whom farmers or agents may speak standard English, must speak dialect to the farmworkers. Although some landlords are Liberals, most are Conservatives. Because English land tenure law then allows landlords to vacate tenancies for any reason, landlords can retaliate against any tenants for failing to cooperate, in politics or otherwise, and they can evict any cottager even suspected of failing to follow tenants’ or agents’ guidance. Consequently, Conservatives win in rural areas in 1886 and for many years thereafter.

But no electoral advantage is forever. Language or dialect categorizes any speaker together with others speaking the same. Before the Education Act of 1870, the cleavage dividing great men by birth or talent and large proprietors from the remainder of the population also separates speakers of written standard English from speakers of dialect, an urban contact language, or one of three or more forms of Gaelic still spoken. Over three decades, compliance with the Act—even marginally enforced and even hindered by unnecessarily complex instruction in spelling and grammar—starts to blur that divide. Even if rural pupils were to compose all of the 18.5 percent of pupils counted absent—which they do not—still every year, 63 percent would be learning to read, however “miserably educated.” As schoolchildren mature into agricultural workers, the villages become populated with more literate people able to read political circulars and at least to understand, if not to repeat, campaign speeches in standard English. They can use dialect to explain to other villagers still unable to read. Villagers trust each other far more than they do the tenant farmers who employ them, and especially as the agrarian depression at the end of the 19th century worsens, the trust gap widens. As villagers converse with each other about politics instead of listening to tenant farmers, the Conservatives’ rural advantage steadily erodes.

4.6.3 From great, high, and large to large alone: disappearance and recontextualization

While the ending of a restriction to men overseas collects men within the British Isles into a white race, and the effect over time of schooling blurs the boundary between dialects and written standard English, the old descriptors attributing size restricted to a minority of men also fade. In 1886, a descriptor formerly obligatory for a holder of the highest aristocratic title is denied to him by name and title: “not a great man” (28a). Other “great men” are now “great” only relative to other members of the House or to MPs not occupying the “Government Bench” (28b–d), not as before relative to the far larger set of all men, and (28b) likens “great men” to other men rather than differentiating some minority. “Great family” recurs as late as 1910 (28e), but in context, not with praise but with the pejoratives “hatred and contempt.” “High birth” last occurs in 1886 but with its positive qualities said to be found in “any cottage in the land” and with negative consequences encountering “censure” even from Conservatives (28f). The next collocation of “high” with “birth” appears in “high birth rates”—fertility in the whole population. Multiple instances of “large landowner/s” and a few cases of doubly Latinate “large proprietor/s” persist, but adverse contexts are easy to find, like those in (28g–k).

4.6.4 Abolishing the lords’ veto: intersections

The spread of written standard English, the decline and recontextualization of restrictive spatial descriptors, and the emergence of whiteness as a non-restrictive descriptor all intersect to change categorization as the 19th century ends. In 1892, a prominent member of Commons makes the seemingly superfluous observation that learning to read requires something to read (29). Teaching children to read teaches them not only to impute meaning to graphemes but also what meaning they impute to the graphemes they see. Already in a series published as early as 1883–1886 and reaching students over the next decades, a standard textbook tells them how to categorize themselves: “The greater part of the people in England and Scotland…have generally fair hair and blue eyes, and are an independent, conquering race…masters of a very large portion of the earth’s surface” (Heathorn, 1995, 403). By 1904, an MP complains about “men who at home were content with honest labor” whom education has recategorized. “For political purposes it had been dinned into the ears of these men that they were boys of the bull-dog breed” (30).

The Empire that in a schoolbook has recast the downtrodden as “masters” also provides context for the spread over time of the category “white labor.” Gradually, instead of the rarity encountered before 1850, it becomes usual for workingmen even in the British Isles. In 1890, the context of emigration to West Australia resurrects “white subjects of Her Majesty from these British Islands” (31a). “White labor” is desired for Natal and Queensland (31b). “White stokers” outlast Lascars—seamen from India (31c). “British-born white subjects” populate Transvaal (31d). Lascars take jobs on ships from “white labor” or “white men” (31e-h). In 1901, “our white subjects” of “Dutch blood” or “English blood” feel “sympathies of race we all share” (31i). In 1902, “white men” feel a “prejudice against working with black men” that “exists wherever black labor is employed” (31j). In 1902, a Conservative uses “white labor” interchangeably with “so many English laboring men” (31k). Two years later, work that “the African native” refuses is “insulting to white workmen” (31l). By 1904, “Englishmen, Scotchmen, and Irishmen” constitute “tens of thousands of honest white laboring men” (31m). A Conservative large landowner, no less, talks about “the injury done to white labor in this country” by tolerating immigration from Europe rather than addressing the lot of “the white slaves in our great cities” (31n). In many cases, the “white working man” in South Africa is “born north of the Tweed” (31o).

In February 1904, the Leader of the Liberal Opposition, twenty months from becoming Prime Minister, declares “the necessity of having equal rights for all white men” (32a). In March 1904, the Conservative Prime Minister, the late Lord Salisbury’s nephew Balfour, agrees. He says about South Africa: “I plead for the great white population…a great community of whites of all classes…as it were, an aristocracy over a proletariat,” which consists of “thousands and millions of the black population.” To Balfour, “you cannot get the white man and the black man to work together on equal terms…Men are not born equal. They cannot be made equal…we are agreed…that this unbridgeable abyss separates the two races” (32b). When Balfour’s racial “abyss” replaces his uncle’s divide between “those who have” and “those who have not,” white labor has now entered the charmed circle of superiority—“an aristocracy over”—from which restrictive descriptors of size have formerly excluded it. Soon to become a second Liberal Prime Minister, Asquith agrees, declaring “the white population” in South Africa to be “the dominant class” relative to “the black population, which constitutes the great bulk of the indigenous race” as well as of “everybody…under the protection of the Crown” (32c). Five days earlier, an opponent has named Asquith’s three populations: “three colors…white, black, and yellow” (32d). The Liberal Unionist Joseph Chamberlain, whose coalition with Balfour provides the majority in Commons needed for the Conservatives to govern, agrees with Balfour and Asquith: “white labor will not work side by side on equal terms with colored labor. I think white labor is right. After all, we do hold our position by virtue of being the dominant race.” The line between superior and inferior that has run in 1887 between British landowners and “what they were pleased to call the lower orders” (27b) separates whites, including laborers, from populations assigned other colors throughout the Empire. Other speakers confirm the assignment of superiority to white labor (32f–h).

At the end of 1905, the spread of standard English by schooling has converged with the assignment of superiority to a color category embracing all but a very few of the British Isles’ inhabitants and with the disappearance, diminution, or adverse recontextualization of size descriptors. That intersection of discursive changes collapses Conservative control of the Commons that has lasted two decades, with a single interruption lasting less than 3 years. The Liberals win a colossal victory in the 1906 election. Far from having “solved” the collective action problem of attracting voters, the Conservatives resort to the same stronghold that has allowed them to dominate Imperial politics throughout the century. In 1895, the House of Lords has been described as both “an influence thwarting and hindering the progress of useful Radical legislation” and the sole issue in “the next election,” but the discursive changes have not yet advanced far enough to overcome “the fact that you yourselves are not united on this question of the House of Lords” or to spur any “great wave of popular enthusiasm” (33). Winning in 1906, the Liberals and their allies propose a budget rejected by “the large landowners of this country…the 5,000 men who own half of the land in this country…in the House of Lords,” who plead again like Cecil in 1860, or like the speaker in (2a) in 1807, that “they would be brought to absolute ruin” (34). But this time, another election in 1910, even though it reduces the majority backing the Liberals, does grant them “authority to put an end…to the constitutional condition which enables a non-representative and irresponsible authority to thwart the purposes and mutilate the handiwork of the chosen exponents of the peoples’ will” (35). In August 1911, unelected lords agree to surrender entirely their veto over funding and see it reduced to delaying other acts. When they do, a franchise that has spread more widely also acquires hugely expanded powers.

5 Discussion

I set out three explananda. First, collective action among persons sufficiently numerous and dispersed requires underestimation of effort relative to discounted expected gain, as everyone accurately estimating or overestimating effort prefers to abstain, almost certainly sparing effort in the expectation of sharing the gain if others make an effort and incurring no gain if they do not. Second, the oppression that establishes a dictatorship requires much more effort, in the form of surveillance and coercion, than a democracy, which requires only the minimal effort of voting. Third, the scale of collective action increases vastly from a minority deciding whether to join in deterring others’ collective action by surveillance and coercion to everyone becoming eligible to decide whether to cast a vote.

In Britain between 1800 and 1911, talk about politics produces all three effects. Talk categorizes people. The known effect of categorization is the underestimation of effort. The known effect of size categories is the intensification of the underestimate of effort. The known effect of categories expressing diminished size is to counterbalance the effect of categorization in reducing estimates of effort. Another known effect of size categories is to restrict the proportion of any population that the categories can coherently describe. When oppression restricts the number of voters to a small minority between 1800 and 1867 and maintains unelected landowners in control of both Houses of Parliament, talk uses the descriptors of size “great,” “large,” and “high” to describe the minority controlling Parliament. As two of the “great men,” Palmerston and Disraeli, converge in describing the “low” as sharing positive qualities of the “high,” the franchise expands. As the older descriptions persist together with the emergence of a color label in “we white men,” the franchise expands again. As the newly enfranchised receive schooling that enables them to interpret and disseminate the gist of political messages spoken or written in standard English, they too become white, in the expression “white labor,” and when they do, elections unable to reduce the powers of landowners in 1895 accomplish a sharp reduction of their power in 1911. Talk that produces all three effects causes the enfranchisement definitive of democracy. Talk is a single cause, and talk replaces evasiveness about the collective action problem with a simple and compelling solution to what is the decisive, but commonly neglected, problem for understanding either dictatorship or democracy.

Now, of course, Britain is a single case, but in 1911, Parliament is governing about a quarter of the globe’s land area, and the British example is influential everywhere else. If anyone doubts whether an explanation for Britain extends to other cases, I invite them to study talk in those cases for the purpose of examining whether these categories, or other forms of expression with parallel effects, are found there and then, too. I may add that what I find in Britain here resembles very closely the changes in discourse that I have documented for the advent of the 1993 election in Russia and for the other colonial powers (Anderson, 2014, 2018), although every other variable except widespread literacy distinguishes Russia at the end of the 20th century from Britain at the beginning. Of course, also, Britain, even after 1911, remains far from democracy. A minor matter is the limited enfranchisement of 1884 still in effect, as Ziblatt (2017, 128-129) remarks. But that limitation does not even come close to the major disqualification prevailing for Britain in 1911, or for that matter, even in 1928 when women acquire the vote on equal terms. That major disqualification is Parliament’s rule over an Empire with a population almost entirely disfranchised. Britain in 1911, and indeed for decades after 1945, combines democracy in the home islands with an administrative despotism relying on military force elsewhere. And the population deprived of the vote for Commons outside the home islands massively exceeds the whole population inside: “millions…at home…tens of millions…abroad” in 1831; in British India alone, “300,000,000” composing “three out of four” of all Crown subjects in 1907; “our Empire consists of something like five to one colored men” in 1909; “The King…has far more black subjects than white” who “are not represented here” in 1911 (36a–d). The disproportion remains blatant at least until 1947, when Britain retreats from South Asia, and the proportion remains close for decades afterward. The evidence I assemble here is incomplete temporally as well as spatially. But again, the wars that deprive Britain of its colonies and liberate it to become a democracy are exogenous to this explanation of democracy only if the collective action necessary for those wars to occur is somehow independent of the discourse produced by colonialism. And neglected though collective action remains in the study of warfare, its discursive roots need not remain ignored.

Also evident here is the overwhelming impact of the Empire on enfranchisement in Britain. The non-restrictive category “white,” which displaces the restrictive spatial categories, begins as a metaphor for humans residing in the Imperial territories, understood to include the disaffected colonies that rebelled and established the United States. From the colonial encounter with persons labeled “black,” “yellow,” or “red,” the label “white” spreads to Britain, where speakers of English can use it to describe themselves, and without that impetus, it is not obvious how the spatial categories motivating oppression would have been displaced. Nor can downtrodden English laborers attain mastery and racial superiority without “blacks,” “yellows,” and “reds” to oppress and, often, exterminate. As the empire disappears, democracy in the home islands, and not only there, confronts challenges from whites disaffected by democracy’s requirement that they share their rights with persons residing among them assigned to other color categories.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpos.2026.1801692/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Roula Nezi, University of Surrey, United Kingdom

Reviewed by: Agus Dedi Putrawan, Universitas Islam Negeri Mataram, Indonesia

X. San Liang, Fudan University, China

Footnotes

https://hansard.parliament.uk