Delivered in the voice of an Emeritus Faculty Member, College of Sciences
Date: June 12, 2025
Good afternoon, colleagues and students. It is a pleasure to return to the lectern once more and share with you the fruits of a computational reading — what one might call a distant reading — of a remarkable collection of texts assembled under the rubric of a study carrel named simply, and perhaps deceptively, "government."
Before I begin, let me say that the tools of digital text analysis — what some now call the "Distant Reader" — have afforded me the opportunity to examine nearly half a thousand documents comprising some fourteen million words. That is not a misprint. Fourteen million words. In my younger days, I might have spent a decade reading such a corpus by lamplight; today, a machine does the counting, and I do the interpreting. And interpret we must, for numbers without narrative are as sterile as narrative without evidence.
Let us begin with the bare facts of extent. The "government" carrel comprises 484 items and approximately 13,982,108 words. The overall Flesch Readability Score is 53, which places the corpus squarely at the level of a high-school senior or first-year college student — accessible, but not trivial. Some individual texts are far more demanding. The most readable items score in the mid-80s (narrative fiction, for instance), while the most impenetrable plunge into the 30s and below (densely theoretical political economy, psychoanalytic discourse, and the like).
What sorts of documents are we dealing with? The bibliographic metadata reveals a collection of extraordinary breadth. We have Emma Goldman's Anarchism and Other Essays and Kropotkin's The Conquest of Bread; we have the Federalist Papers and Madison's Journal of the Debates in the Convention; we have Tocqueville's Democracy in America and Locke's Second Treatise of Government; we have the Magna Carta, the Constitution of Japan (both 1889 and 1946), and the British North America Act. We have treatises on capitalism, articles on Islamic finance, essays on settler colonialism, and novels about millionaires. In short, the carrel is a cross-section of Western political thought — with significant excursions into Islamic, Southeast Asian, and Latin American intellectual traditions — spanning roughly four centuries.
The authors represented read like a roll-call of political philosophy and social criticism. Allow me to name but a few: Alexis de Tocqueville, John Locke, Karl Marx, Emma Goldman, Mikhail Bakunin, Peter Kropotkin, Alexander Hamilton, James Madison, John Stuart Mill, Walter Bagehot, Herbert Spencer, Vladimir Lenin, Mussolini (through the pen of William Bolitho), Anthony Trollope, Stephen Leacock, Robert Tressell, and Rosa Luxemburg. The list could continue, but I trust the point is made.
The dates of composition are instructive. Many items bear no explicit date, but among those that do, we find texts from 1941 (Egan on corporativism), 1936 (Michel on capitalism), the early twentieth century (the Mother Earth issues of 1906), and a great cluster of more recent scholarship from 2016 through 2022, addressing capitalism, climate change, COVID-19, and the Anthropocene.
Now we come to what I find most fascinating: the word frequencies. The unigram analysis tells us that the most common words in the carrel are, predictably, function words — "the," "of," "and," "to," "a," "in." But among the content words, the leaders are telling:
| Rank | Word | Frequency |
|---|---|---|
| 1 | people | 26,167 |
| 2 | state | 23,030 |
| 3 | government | 20,995 |
| 4 | power | 20,058 |
| 5 | time | 20,052 |
| 6 | man | 19,683 |
| 7 | men | 19,603 |
| 8 | states | 14,138 |
| 9 | life | 12,618 |
| 10 | law | 11,656 |
"People" leads the list, followed closely by "state" and "government." This is, I think, a fitting confirmation that the carrel's name is well-chosen. The discourse is fundamentally about the relationship between people and the structures that govern them. "Power" appears over twenty thousand times; "law" nearly twelve thousand. These are the vocabularies of political life.
The keywords extracted from the corpus reinforce this picture. The top keywords include "people" (207 occurrences in the keyword list), "time" (156), "power" (137), "world" (137), "government" (130), "man" (116), "life" (114), "state" (109), "good" (101), and "men" (90). One also notes the prominence of "capitalism" (62), "war" (67), "authority" (54), "king" (54), and "democracy" (29).
The bigram analysis is particularly illuminating. The most frequent two-word phrases include:
| Bigram | Frequency |
|---|---|
| united states | 5,502 |
| home rule | 2,235 |
| i think | 1,756 |
| i shall | 1,721 |
| we must | 1,644 |
| let us | 1,419 |
| you know | 1,379 |
| great britain | 1,345 |
| supreme court | 845 |
| higher education | 860 |
| federal government | 709 |
| civil war | 447 |
| social media | 410 |
| working class | 381 |
| surplus value | 559 |
Notice the tension here. "United States" dominates, but "home rule" — that great Irish question — is the second most common phrase. The carrel clearly contains a substantial sub-corpus on Irish constitutional politics, including texts by Dicey, Childers, Gladstone's speeches, and a host of Home Rule pamphlets. Meanwhile, "supreme court," "federal government," and "civil war" point to the American constitutional tradition, while "surplus value" and "working class" reveal the Marxist economic analysis threading through the collection.
The noun frequencies confirm the carrel's preoccupations. "People" (26,167), "state" (23,030), "government" (20,995), "power" (20,058), and "time" (20,052) lead the field. But one also finds "king" (10,739), "parliament" (7,372), "constitution" (9,851), "war" (7,933), "society" (7,689), "money" (7,243), "law" (11,656), "world" (11,107), and "democracy" (3,524). These nouns sketch the architecture of political life across centuries: monarchy, parliament, constitution, war, money, and the perennial question of democracy.
The verb frequencies are, if anything, more revealing. The most common verbs are "have" (46,028), "make" (33,704), "say" (29,327), "do" (23,405), "be" (22,506), "take" (20,304), "see" (18,473), "give" (18,303), and "know" (17,353). Among the more thematically significant lemmatized verbs, we find "go" (16,805), "come" (15,558), "think" (14,871), "find" (12,032), "become" (10,112), "make" (33,704), "call" (8,971), "seem" (7,842), and "work" (6,696). The verbs of governance — "govern" (2,565), "establish" (4,336), "provide" (4,575), "require" (4,416), "maintain" (2,844), "regulate" (1,257), "compel" (1,300), "elect" (2,254), and "appoint" (2,656) — are all well-represented. The corpus is not merely descriptive; it is prescriptive. It tells us what should be done, not merely what is.
The adjective list is a study in contrasts. "Other" (27,437) and "great" (25,278) lead, but one quickly encounters "social" (10,867), "political" (10,855), "first" (10,708), "public" (7,834), "national" (6,750), "necessary" (6,290), "irish" (6,197), "present" (6,146), "human" (5,868), "different" (5,769), "whole" (5,767), "free" (5,517), "economic" (5,288), "federal" (3,579), "democratic" (2,503), "capitalist" (2,537), and "revolutionary" (1,340). The adjectives paint a picture of a corpus preoccupied with the public, the national, the human, the free, and the economic — and with the perpetual tension between the democratic and the capitalist, the revolutionary and the constitutional.
The named-entity extraction reveals the human and geographic geography of the carrel. Among people, we find, in order of frequency: Marx (1,306), Owen (1,206), Madison (1,190), Gladstone (854), Hamilton (843), Shirley (798, a fictional character from Leacock's satire), Mussolini (323), Lenin (222), Stalin (101), and Hitler (36). The presence of fictional characters alongside historical figures is a reminder that the carrel includes novels as well as treatises — Trollope's The Way We Live Now, Leacock's Arcadian Adventures with the Idle Rich, and Tressell's The Ragged Trousered Philanthropists among them.
Among places, the geography of political thought unfolds: Ireland (6,783), the United States (5,822), England (4,542), America (3,441), France (3,065), New York (2,841), London (2,747), Germany (1,876), Rome (1,813), Italy (1,150), Paris (1,103), Russia (1,100), Great Britain (1,088), Spain (1,049), China (1,016), Washington (1,009), India (689), Japan (634), and Indonesia (504). The prominence of Ireland is striking and confirms the substantial Home Rule sub-corpus. The global reach — from Indonesia to Japan to South Africa — speaks to the carrel's ambition.
Among organizations, the leading entities are Congress (4,275), the Senate (3,552), the House of Commons (831), the Supreme Court (715), the House of Representatives (543), the Catholic Church (138), the Church of England (115), and the United Nations (75). More recently, we find Facebook (45), Google (12), Microsoft (43), and the European Union (57), reflecting the contemporary turn toward digital and transnational governance.
What I find most instructive about this carrel is not any single statistic but the texture of the whole. The corpus moves from the magisterial prose of the Federalist Papers — where Madison and Hamilton argue for ratification with the urgency of men who know that empires hang in the balance — to the urgent, sometimes breathless prose of contemporary scholars writing about climate change, surveillance capitalism, and the Anthropocene. It moves from the patient constitutionalism of Dicey and Bagehot to the revolutionary fervor of Goldman and Berkman. It contains both the Magna Carta and the Constitution of Japan, both Locke's Second Treatise and Mill's On Liberty, both Tocqueville's patient observation and Luxemburg's passionate analysis.
In a word, the carrel is a library of political imagination — or rather, of political imaginations, plural and sometimes contradictory. It asks, across four centuries and several continents, the same fundamental question: how shall we be governed, and by what right?
As an old man of science, I am perhaps more accustomed to quantitative reasoning than to literary exegesis. But I have learned, over a long career, that the most interesting questions are those that cannot be answered by computation alone. The "government" carrel gives us numbers — fourteen million words, four hundred and eighty-four items, a Flesch score of fifty-three — but the meaning of those numbers lies in the human voices they represent.
What does it mean that "people" is the most common noun? That "power" appears twenty thousand times? That "law" and "war" jostle each other in the rankings? It means, I think, that the central preoccupation of political thought has not changed since Locke wrote his Second Treatise or since the barons forced King John to sign at Runnymede. The question is always the same: who shall have power, and how shall it be constrained?
That the carrel contains both the Magna Carta and articles on surveillance capitalism, both the Federalist Papers and essays on the Anthropocene, is not a contradiction but a continuity. The technology changes; the question does not.
I commend this carrel to your attention. Read it closely where you can; read it distantly where you must. But read it. For in these fourteen million words, you will find the longest conversation our species knows: the conversation about how we shall live together.
Thank you.