Caillebotte's silences

A current exhibition of the paintings of Gustave Caillebotte (1848-1894) at the Art Institute in Chicago is quite remarkable. It demonstrates the eye, the hand, and the sensibility of this great late-Impressionist painter. But the exhibition is remarkable in another way as well: there is almost no evidence in the paintings on exhibit, or the curatorial texts that support the exhibition, that conveys the intense and prolonged social, political, and military conflict of the period from the late 1840s through the defeat of the Paris Commune (1871).

Caillebotte himself served in the French military during the siege of Paris by the Prussian Army (1870-1871). In the text describing the exhibition at the Musée d’Orsay in Paris a curator writes briefly of his service: “But during the Franco-Prussian War (1870-1871), [Caillebotte] was drafted into the 7th Battalion of the Garde Nationale Mobile de la Seine and assigned to the defense of Paris” (link). The detail about his service in the 7th Battalion of the Garde Nationale Mobile is especially telling. Michael Howard describes the Battle of Buzenval in these terms:

The battle of Buzenval, as it was to be called, settled the fate of Paris. More, it destroyed once for all the belief that a People in Arms could overwhelm a trained enemy by sheer numbers and burning zeal. It was the action for which the clubs had for so long yearned—the sortie en masse. Nearly 90,000 men were involved, of whom about half came from the Garde Nationale, and at dawn on 19th January they debouched from Mont Valérien, and advanced against the German defences between Bougival and St Cloud along a four-mile front. (Howard, The Franco-Prussian War, p. 373).

This battle was yet another disaster for the French military. It essentially sealed the fate of the besieged city and forced surrender of the last part of France still resisting German control. If Caillebotte was an active participant in this battle, he would have had traumatic and transformative experiences of war; and if he was held in reserve in the city during this final break-out attempt, he would have had personal knowledge of the significance and suffering created by the siege of Paris and the several unsuccessful efforts to break the siege through counter-attacks. Surely this is an important element in his development as an observant human being and a painter. And yet there is no evidence in his paintings of the impression the trauma of Paris may have had upon him. It would seem that this is an important contextual feature that should play a highlighted role in the curatorial presentation of the exhibition; but it does not.

Stéphane Guégan, a scientific advisor to the Musée d'Orsay in Paris and author of Caillebotte: Peintre des extrêmes, considers the “military presences” in Caillebotte’s paintings. Referring to the first exhibition of the Impressionist group in 1874, he notes that “the [first Impressionist] exhibition did contain a few resounding echoes of the catastrophic situation from which the country was barely emerging: the crushing defeat of the Franco-Prussian War and the ensuing upheaval of the Paris Commune” (“The Shared War”, Gustave Caillebotte: Painting Men, p. 42). But virtually no such references occur in the exhibition that has travelled from Paris to Los Angeles to Chicago. Guégan notes several exceptions: the painting of a soldier in uniform and an almost imperceptible representation of a uniformed soldier in Le Pont de l’Europe.

But neither image gives any sense of the true military catastrophe of the Franco-Prussian War or its aftermath. Instead, the curators have chosen to organize their ideas about Caillebotte’s paintings around the form of masculinity represented by his work.

Contrast Caillebotte’s silence in his painting with that of Jean-Louis Ernest Meissonier, a generation senior to Caillebotte. Here is Meissonier’s 1849 depiction of a massacre he apparently witnessed following the defeat of the workers’ uprising in June 1848.

And here is Meissonier’s 1884 painting representing a scene of death and destruction during the siege of Paris in 1870:

Finally, here is a daguerrotype of a poignant scene following the massacre of Communards following the fall of the Commune in 1871:

The traumas represented in these images were part of the experience and memory of the people of Paris during those decades, including both Meissonier (who was 33 at the time of the 1848 workers’ uprising) and Caillebotte (who was born in 1848 and was a serving member of the Garde Nationale Mobile in 1870-1871 during the final months of the siege of Paris and the suppression of the Commune itself (by French forces). How could either of these painters not have been deeply affected by these traumatic events of contemporary French history? Meissonier’s paintings take note of these fundamental facts, but Caillebotte’s do not. And yet the exhibition gives no historical context at all that would highlight these important and surely formative events.

And what about Meissonier? His depiction of the massacre of workers in 1848 might suggest that he was offering sympathy and homage to the working class men and women who rose up in June, 1848. Alexis de Tocqueville and Alexander Herzen, both observers of the fighting in Paris in June 1848, offered sympathy and sorrow for the violence that overwhelmed the workers’ uprising. Here are comments offered by Herzen:

I listened to the thunder and the tocsin and gazed avidly at this panorama of Paris; it was as though I was taking my leave of it. At that moment I loved Paris passionately. It was my last tribute to the great town; after the June days it grew hateful to me. On the other side of the river barricades were being raised in all the streets and alleys. I can still see the gloomy faces of the men dragging stones; women and children were helping them. A young student from the Polytechnic climbed up on to an apparently completed barricade, planted the banner and started singing the Marseillaise in a soft, sad, solemn voice; all the workers joined in and the chorus of this great song, resounding from behind the stones of the barricades, gripped one’s soul. . . . The tocsin was still tolling. Meanwhile, the artillery clattered across the bridge and General Bedeau standing there raised his field-glasses to inspect the enemy positions. . . . (From the Other Shore, After the Storm, 46)

But this is not the current interpretation of Meissonier’s work. Rather, critics have suggested that the 1849 painting of the massacre at the barricade conveys a middle-class view of the insurrection, and serves as a caution for the future: “insurrection leads to massacre and death”, while the 1888 painting conveys a sense of patriotism and heroism.

An element of historical change that is entirely evident in Caillebotte’s paintings is the transformation of Paris by Baron Haussmann at the direction of Napoleon III. The “Paris Street: Rainy Day” painting above reflects the Haussmannization of Paris — the broad avenues, the “modernization” of life in the city, and the destruction of working class residential areas. This is a central theme in T.J. Clark’s interpretation of mid-century depictions of Paris in The Painting of Modern Life:

It seems that only when the city has been systematically occupied by the bourgeoisie, and made quite ruthlessly to represent that class’s rule, can it be taken by painters to be an appropriate and purely visual subject for their art….. For the House knew well that Haussmann’s modernity had been built by evicting the working class of Paris from the centre of the city, and putting it down on the hill of Belleville or the plains of La Villette, where the moon was still most often the only street light available. And what did painters do except join in the cynical laughter and propagate the myth of modernity? (The Painting of Modern Life, p. 51)

According to Clark, there was an overriding theme of class conflict and a fear of insurrection that drove both Napoleon III and Baron Haussmann in this urban project. They were concerned to “modernize” Paris in a way that would make working class rebellion (and the barricades through which previous uprisings had proceeded) impossible; troops and cannon would be able easily to clear the avenues of insurrection.

There was no disputing that part of Haussmann’s modernity was his wish to put an end to insurrection. He stated as much himself: it was a good argument to lean on when pleading for funds from the Conseil Municipal. Years after the event, he was still musing in his Mémoires over the hidden benefits of the Boulevard Sébastopol: “It meant the disembowelling of the old Paris, the quartier of uprisings and barricades, by a wide central street piercing through and through this almost impossible maze, and provided with communicating side streets, whose continuation would be bound to complete the work thus begun. The subsequent completion of the Rue de Turbigo made the Rue Transnonain [symbolic capital of the barricades] disappear from the map of Paris!” Nor was this merely a matter of hindsight on Haussmann’s part. The details of counterrevolution weighed heavily on the planners’ minds at the time: Napoleon intervened directly in 1857 to prevent the encirclement of the Faubourg Saint-Antoine from being spoiled by a mere architect’s whim: “the construction of arcades on the Boulevard Mazas,” he wrote, “would seriously damage the strategic system of Paris.” The arcades were quietly dropped from the designs. (75-76)

Here again the current exhibition’s curators have seemingly ignored the social and political context of the Haussmannization of Paris. They emphasize the “new modernity”, the dress of the mostly bourgeois men and women passing across the boulevards, and the relaxed scenes of conversation and amusement among Caillebotte’s male friends. But there is no curatorial mention at all of the political fears and imperatives that appear to have driven Napoleon III and Baron Haussmann in this wholesale restructuring of the urban environment of Paris. And there is no suggestion in the paintings on display in the current Caillebotte exhibition of a sensibility on the part of the painter to the underlying conflicts between working class Parisians and the bourgeoisie. There are notes of awareness and sympathy for working men and women in his corpus — for example, in the painting “The Floor Scrapers”. But there is no suggestion of his own awareness of the concrete circumstances of injustice, exploitation, or unnecessary misery in his paintings.

So the social conditions of class and war seem to be almost entirely absent in Caillebotte’s work. There is no sense of “social critique” or self-awareness of upper-middle-class position in these paintings. Caillebotte is silent about the most momentous events that transformed France and Paris during his adulthood, and this is surprising. But it is also surprising that the curators have not taken more notice of this absence.

Real multicultural democracies

Chicago is a highly diverse city, and it is a good example of life in a multicultural democracy. The image above is a photo of the crowd on Navy Pier on a recent Saturday summer evening. According to local estimates, as many as 120,000 people visit Navy Pier on a Saturday night, and it is a good practical example of the benefits of multicultural democracy. The crowd is highly diverse, with adults and children from all racial groups and many ethnicities and language groups. And there is a substantial degree of social class mixing as well, from young professionals from the North side to working class families from the South and West sides of the city. Turn your head in different directions and you will hear a dozen different languages. The atmosphere is comfortable, fun, accepting, and interactive, with a Latino music performance going on in the open-air music venue, families enjoying a meal from the food court, and a beautiful view of the Chicago waterfront and skyline. It's a fun outing for all the residents of the city. (Chicago's population is about 2.8 million, so a typical Saturday night on Navy Pier in the summer draws almost 5% of the city's residents.)

What are the facts of Chicago's diversity? Chicago's population is now about 2.75 million, of whom 21% are foreign born. According to the US Census Bureau 2020 Decennial Census (link), the largest racial/ethnic communities in the city include White (36%), Hispanic/Latino (30%), Black/African-American (29%), and Asian (7%). 11% of respondents reported "two or more races" in the Census questionnaire. (It will be noted that these population groups add up to more than 100%. The Census Bureau provided some information about changes in methodology in 2020 which may account for this discrepancy; link.)

Segregation in Chicago region

So Chicago is highly diverse. However, the city remains significantly segregated by neighborhood, and these patterns of segregation produce a continuing legacy of disadvantage in terms of important measures of social wellbeing (health, economic opportunity, educational outcomes). The Metropolitan Planning Council and the Urban League have studied these trends carefully, and their "Shared Future" report (link, link) serves both to detail the facts of segregation in Chicago today and to outline some strategies for reversing these trends.

So Chicago's problems of achieving racial equality persist. And yet on a warm August evening in the center of the Chicago Loop, it is possible to see how this city is creating a climate of mutual respect and civic equality. Multiple community-based organizations do the work of striving for racial justice and establishing an inclusive community for all Chicagoans through ongoing efforts, programs, and community alliances. The city's political leadership recognizes that "unity within diversity" must be the beginnings of Chicago's public values and urban politics. And academic institutions like the University of Illinois Chicago's Institute for Research on Race and Public Policy (link) have a continuing commitment to documenting the facts about racial and ethnic equality in Chicago, and identifying policy initiatives that can lead to meaningful progress. It is possible for our society to become more just and more harmonious through our own patient collective efforts.

If we look carefully at the photo of the Navy Pier crowd, we will see something surprising looming over the horizon of this vibrant mass of multicultural humanity. We see in the distance the luxury hotel and tower developed by the president of the United States, located a half-mile up the Chicago River. The contrast could not be more striking, between the glittering symbol of the political movement that is demonizing diversity in our country, versus the social bonds and community spirit of mutual acceptance that constitute the reality of our multicultural democracy.

Langston Hughes caught much of the paradox of race in America when he wrote these lines in 1935:

Let America be America again.
Let it be the dream it used to be.
Let it be the pioneer on the plain
Seeking a home where he himself is free.

(America never was America to me.)

Let America be the dream the dreamers dreamed—
Let it be that great strong land of love
Where never kings connive nor tyrants scheme
That any man be crushed by one above.

(It never was America to me.)

And Walt Whitman was right too when he wrote, "I hear America singing, the varied carols I hear" as his own celebration of the breadth of experience of American society. America sings on Navy Pier and the many other places where citizens do better than politicians at facing the challenge of creating durable multicultural democracy. Being there reinforces one's confidence that the community and diversity of our country will prove stronger than the forces of xenophobia, mistrust, and antagonism that are being mobilized against us.

Ethnography of high-energy physics

Science proceeds through research communities whose participants share important and often distinctive features of thought and method. This is one of the key insights of the “historical turn” in the philosophy of science initiated in the 1970s (link, link), and it underlies much work within the interdisciplinary field of Science and Technology Studies. But what more specifically goes into the “denkkollectiv” (Ludwik Fleck), “research programme” (Imre Lakatos), or “disciplinary matrix” (Thomas Kuhn) of a specific scientific field? One way of gaining knowledge about those features of thinking and experimenting in specific research communities is through immersive study by ethnographers and micro-sociologists. Paul Rabinow offered an especially fruitful example of this kind of investigation in Making PCR (link). Rabinow was specifically interested in discovering the mental and material worlds of biotechnology researchers.

This book focuses on the emergence of biotechnology, circa 1980, as a distinctive configuration of scientific, technical, cultural, social, economic, political, and legal elements, each of which had its own separate trajectory over the preceding decades. It examines the “style of life” or form of “life regulation” fashioned by the young scientists who chose to work in this new industry rather than pursue promising careers in the university world…. In sum, it shows how a contingently assembled practice emerged, composed of distinctive subjects, the site in which they worked, and the object they invented. (Making PCR, 2)

And what about the most esoteric of contemporary scientific research, high-energy particle physics (link)? How does this extended network of researchers think and work as this community seeks out further features of fundamental physics? Peter Galison’s Image and Logic: A Material Culture of Microphysics is a brilliant, clear, and extensive exposition of the interface between theory and experiment in physics. Galison thinks of contemporary physics as an overlapping set of three kinds of activity: experimentation, instrumentation, and theorizing. In this book he looks at instrumentation and the machines of physical investigation as a realm that requires its own careful study — from a historical-sociological point of view as well as from an epistemic one.

These machines have a past. To walk through the laboratories of the twentieth century is to peruse an expanse of history in which physics has played many parts. Over here, film for atomic physics, X-ray film out of boxes destined for medicine; over there, a converted television camera rewired as part of a spark chamber. In this corner a piece of preparatory apparatus for a hydrogen bomb, in that a cannibalized bit of computer. Around you in the 1950s the structure of mutable, industrial-style laboratories introduced to physics in the wartime scramble to ready nuclear weapons and radar. Shaped by the exigencies of industry and war, but also shaping the practices of both, the machines of physics are part of a wider technological material culture—neither below it, nor above it. (xviii)

And Galison emphasizes that the realm of “the practice of physics” encompasses many forms of activity: institutions, social networks, extended working groups, peer-reviewed journals, and specialized forms of knowledge developed in industrial, military, and corporate spaces.

Even this penciled sketch is but a partial presentation of the multitude of worlds within physics; there were other worlds beyond. Left out are the different university and national groups participating in large experiments, not to speak of the theorists, phenomenologists, administrators, and industrialists; there are computer programmers simulating runs and figuring out how to acquire, store, and sort the data; there are postdocs running shifts. Somehow, out of it all, comes an argument. This picture of science fits badly into the narrowly construed rationality of the algorithmic, and equally badly into the image of an unreasoned struggle by opposing forces to divvy up the territory of knowledge. Physics as a whole is always in a state of incomplete coordination between extraordinarily diverse pieces of its culture: work, machines, evidence, and argument. That these messy pieces come together as much as they do reveals the presence, not of a constricted calculus of rationality, but of an expanded sense of reason. (xxii)

Moreover, Galison suggests that laboratory machines have “meaning”, in a fairly specific sense: they have been designed and adapted by intentional agents with specific explanatory goals in mind.

I will argue that laboratory machines can command our attention if they are understood as dense with meaning, not only laden with their direct functions, but also embodying strategies of demonstration, work relationships in the laboratory, and material and symbolic connections to the outside cultures in which these machines have roots. (2)

This point amounts to a denial of technological determinism — the idea that technologies (machines) have a specific and inherent logic of development. Against this view, Galison puts forward an “agentic” view of the group processes of instrumentation and experimentation. Individuals and teams make informed guesses about what kinds of probes and instruments will illuminate particular problems, and they design instruments to carry out those investigations. And we can also look at this as a “social embeddedness” conception of the physics laboratory: the physicist (theorist, experimenter, instrument designer) brings with him or her assumptions and mental frameworks drawn from the broader society in which they emerge.

Another important insight Galison offers has to do with the “logic of experimentation” itself. In the empiricist tradition there is the idea that experiments are the means through which observation enforces the constraints of evidence on theory. But Galison emphasizes throughout the book that the nature of “experiment” and “experimenter” has changed dramatically over the past two centuries — perhaps most radically in the past fifty years. “Big science” at CERN or the Fermi Laboratory necessarily involves the extended and collaborative work of thousands of experts and technicians; so who is the experimenter there? Rather, it is necessary to interpret and reinterpret the results of the data collected after high-energy collisions, and these data do not speak univocally for themselves. “It is amid these intimate bits of machines, data, and interpretations that the categories of experiment and experimenter are embodied: defined, dismantled, and reassembled” (7).

Galison offers a novel approach to the problem of “scientific incommensurability”. Introduced by Thomas Kuhn as “incommensurable paradigms” guiding related research communities, the idea has proven elusive. Galison approaches the problem from the point of view of small differences in language and vocabulary across closely related laboratory communities; he uses the anthropologist’s ideas of creoles and pidgins to capture the differences in meaning that he identifies (48). He writes:

Because the picture of physics sketched here is one of distinct but coordinated subcultures, the notion of an interlanguage is a useful decentered metaphor. In different forms the same kind of question arises; How should we think about the relation of theorists to theorists, of theorists to experimenters, of physicists to engineers, of chemists to physicists, of image instrument makers to logic instrument makers, and of the myriad of detector subgroups within a hybrid experiment one to the other? To homogenize these various groups artificially is to miss their distinct ways of going about their craft; to represent them as participating in isolated conceptual schemes “translating” back and forth is to shut our eyes to the productive, awkward, local coordination by which communities, machines, and knowledge get built. Consider three aspects of the interlanguage. (49)

Through these “interlanguages”, Galison suggests, the separate subcultures are able to communicate about the terms and procedures of their collaborations. And this suggests a practical response to W.V.O. Quine’s hypothetical worries about the “indeterminacy of translation” that he believes confronts all inter-linguistic encounters. This is an interesting and clearly formulated framework for seeking to understand the micro-level transactions across research communities in a large research project like the activities conducted at CERN or the Fermi Laboratory. Galison writes:

In many different ways this book is a working out of the following observation: pieces of devices, fragments of theories, and bits of language connect disparate groups of practitioners even when these practitioners disagree about their global significance. Experimenters like to call their extractive moves “cannibalizing” a device. (54)

There is a further point to emphasize in Galison’s approach: his consistent avoidance of the idea that “the experimental method” exists as a general and uniform exercise in empirical science. Against this idea, he emphasizes the contingency and capacity for change that historical studies of scientific episodes display — if we are alert to the fallacy of over-generalization. For this reason he explicitly denies that the episodes he considers in this book point to a common model of “experimentation” that might be incorporated into the philosophy of science or general statements about scientific method:

The chapters of this book, like the Medieval and Renaissance histories I have cited, are grounded in the local. But I resist the designation “case study” because I do not believe that there is a set of defining precepts that can be abstracted from these or other studies to “experiment in general” (or, for that matter, “theory in general” or “instruments in general”). (62)

Rather:

My question is not how different scientific communities pass like ships in the night. It is rather how, given the extraordinary diversity of the participants in physics—cryogenic engineers, radio chemists, algebraic topologists, prototype tinkerers, computer wizards, quantum field theorists—they speak to each other at all. And the picture (to the extent one simplifies and flattens it) is one of different areas changing over time with complex border zones that sometimes vanish, coalesce, and even burgeon into quasi-autonomous regions in their own right. (63)

This is history of science at its best: attuned to the contingency and heterogeneity of various scientific research practices, sensitive to the powerful influence of context (political, ideological, economic, military) on the conduct of science, and respectful as well of the quality and rigor of scientific work when it is done well.

Anthropologist Arpita Roy took up some of these questions through an extended period of field work at the European Organization for Nuclear Research (CERN) beginning in 2007, during which she interacted intellectually and practically with dozens of physicists as they performed their scientific work. The primary result of Roy’s ethnography is her recent book, Unfinished Nature: Particle Physics at CERN. The book is most interesting when the author reports and discusses specific conversations and topics that came up with a range of specialists during her field work (theorists, experimentalists, instrumentalists, engineers, computer analysts). These conversations offer the reader a basis for reaching his or her own conclusions about the micro-culture of the CERN technical environment. Also useful is her discussion of the explosion that occurred in the accelerator tunnel in September 2008 and that interrupted work for about fourteen months. And the stated goal of the book is valuable as well:

In that vein, it is not my intention to offer an exhaustive description of a science nor a prescription for a better science but to look closely at some of the presuppositions that serve in an interesting way to connect the technical procedures of a laboratory with wider principles of intellectual classification…. By presuppositions, I mean the class of beliefs that is collectively and unconsciously held by participants and of which they are unaware but that informs every aspect of scientific thinking and activity. (5)

The book is less convincing when the author turns to reflections drawn from Marx, post-modern thinkers, and other areas of philosophy. It is unclear, for example, how Marx’s conception of the division of labor is genuinely illuminating when it comes to understanding the workings of a large laboratory complex. There is a division of labor in this institution, of course; but Marx’s delineation seems to shed little light on this fact (any more than Durkheim’s discussion might have done).

Detailed inquiries into the concrete practices and mentalities found in “big science” laboratories and research institutes are important contributions to both the sociology of science and eventually to our understanding of the epistemic standing of physics. Realist philosophers of science are confident in one of the dualities criticized by Arpita Roy — the distinction between the knower and the properties of the physical world, or the distinction between subject and object — but the cognitive and social practices involved in the scientific enterprise are deeply interesting in their own right, and ethnographic studies of the ways in which scientists and engineers go about their work are deeply interesting. Ludwik Fleck attempted such studies in the 1930s, and this tradition of investigation of “science in the making” has proven to be profoundly insightful (link, link). And emphasis on extra-scientific features of “context”, including gender, race, business interests, and national security pressures is plainly relevant to the conduct of big science — the military-industrial complex described by President Eisenhower almost 75 years ago.

Arrtificial intelligence tools for historians

Historical research may seem to be a field in which AI tools will be especially useful. Historians are often confronted with very large unstructured digital collections of documents, letters, images, treaties, legal settlements, contracts, and diplomatic exchanges that far exceed the ability of a single human researcher to sift and analyze for valuable historical insights. Can emerging tools in the AI revolution help to make systematic use of such historical collections?

Earlier applications of new quantitative methods of analysis of historical data

Several earlier periods of innovation in twentieth-century historical research suggest that historians can often borrow fruitfully from new methods and analytical tools developed in other sciences. The cliometric revolution of the 1980s (Fogel and Elton 1984; Rawski 1996; Wright 2015) brought tools of econometrics, demography, and statistics more fully into play in historical inquiry. Historians have made extensive and productive use of quantitative methods borrowed from the social sciences to investigate questions concerning the health status of various historical populations and the standard of living in different cities and regions (Crafts 1980; Lee and Feng 1999; Allen 2000; Allen, Bengtsson, and Dribe 2005). These tools usually depend upon the availability of structured databases of comparable data over time—for example, census data, birth, marriage, and death records, military records of recruits, and price data for representative goods (wheat, rice, salt). There are issues of comparability, reliability, and validity that arise in these applications of large historical datasets, but these issues are no more difficult for historians than for sociologists or political scientists. Another major area of innovation was the geospatial revolution of the 1990s (Presner and Shepard 2016; Skinner, Henderson, and Yuan 2000; Thill 2020). Efforts to place historical data and events into spatial order have been very productive in suggesting new historical patterns and causal influences not visible in purely narrative accounts. G. William Skinner’s pathbreaking work on the economic regionalization of China is an outstanding example (Skinner 1977), and Peter Bol and colleagues have collaborated in the establishment of a major historical GIS database for China (Bol 2006; Bol 2007). So it is quite understandable that some contemporary historians are interested in the potential value of emerging tools of digital humanities, semantic search, and big-data analytics in their efforts to make sense of very large archives of digitized text and image materials.

However, archival collections of interest to historians present special obstacles to digital research. They are usually unstructured, consisting of collections of heterogeneous text documents, contracts, local regulations, trial documents, imperial decrees, personal letters, and artifacts and images. Moreover, the meaning of legal, political, and religious vocabulary is sometimes unclear from a modern perspective, so translation and interpretation are problematic. The written language of the documents itself is problematic. Often handwritten, interspersed with references and asides in other languages, and often using vocabulary that has no exact modern equivalent, the task of interpreting each historical document itself is challenging for the historian and for the software system. Are there tools that allow the historian to sift, summarize, categorize, and highlight the texts, sentences, and paragraphs that are included in a large archival collection? Major new capabilities have emerged in each of these areas that have substantially enhanced the ability of historians to classify and analyze very large unstructured text databases and archives. These capabilities involve advances in machine learning, large language models, semantic search tools, and big-data analytics. Like any innovation in methods of inquiry and inference, it is crucial for researchers to carefully evaluate the epistemic reliability of the tools they utilize.

Digital humanities

In the past several decades scholars in the humanities, including comparative literature, art history, and various national literatures, have explored applications of computational tools for the analysis of digital texts that permit a breadth and depth of analysis not previously available. These research efforts are now described as digital humanities. Several handbooks and overviews on digital humanities have appeared (Schreibman, Siemens, Unsworth 2004; Schreibman, Siemens, Unsworth 2016; Eve 2022). The goals of research within the field are varied, but in almost all cases the research involves computational analysis of large databases of text, image, and video documents, with the general goal of discovering large patterns that may be undetectable through traditional tools of literary or art-history analysis. Franco Moretti’s Graphs, Maps, Trees: Abstract Models for a Literary History (2005) and Distant Reading (2013) offer excellent examples. Moretti wishes to explore “world literature”; and the field of documents included in this rubric is too large for any single critic or research team to read closely all the available works in the field. Moretti writes, “A larger literary history requires other skills: sampling; statistics; work with series, titles, concordances, incipits—and perhaps also the ‘trees’ that I discuss in this essay” (2013: 67). In place of the insights of close reading, Moretti emphasizes the value of “distant reading” and the effort to discover broad and long patterns across national literatures and centuries. This requires using analytical tools of computational social science to classify texts, identify word patterns, create spatial networks, and (perhaps) to algorithmically assign markers to topics and styles in the texts subject to analysis. Martin Paul Eve writes, “Under such a model, the idea is that computational detection of style, theme, content, named entities, geographic place names, etc. could be discerned at scale and aggregated into a broader and continuous literary history that would not suffer from the same defects as a model that required one to read everything” (Eve 2022: 130).

Efforts in the digital humanities have evident relevance to the problems presented by large text and image datasets available in many areas of historical research. One promising area of application involves using big data tools of text analysis—for example, machine learning, content extraction, and semantic search—to systematically survey and classify all the documents in a collection. The impetus and initiatives of the field of “digital or computational history” are described in Siebold and Valleriani 2022 and Graham, Milligan, Weingart, and Martin 2022. The methods currently creating a great deal of interest among historians are based on joining machine learning methods, big-data analytics, and large language models (LLMs) in order to permit analysis and codification of the semantic content of documents. To what extent can emerging computational tools designed for management and analysis of large unstructured text and image databases be adapted to assist the historian in the task of assimilating, interpreting, and analyzing very large databases of historical documents and artifacts?

Pre-processing and information extraction

An avenue of research in computer science that supports analysis of large unstructured datasets containing texts and images is the field of information extraction (Adnan and Akbar 2019). Information extraction technology consists of algorithms developed to analyze patterns in text (and images or videos) to apply labels or tags to segments of the data. These are generally “big data” tools using machine learning to identify patterns in target documents or images. Adnan and Akbar put the goal of information extraction tools in these terms: “It takes collection of documents as input and generates different representations of relevant information satisfying different criteria. IE techniques efficiently analyze the text in free form by extracting most valuable and relevant information in a structured format” (Adnan and Akbar 2019: 6). In general terms, information extraction tools are expected to provide a structured basis for answers to questions like these: What is the document about? What persons or things are mentioned? What relationships are specified within the document? What events are named? The tools are often based on natural-language models that require training on large text datasets and sometimes make use of machine learning based on neural networks (Rithani et al. 2023). “The concept is to automatically extract characteristics from massive artificial neural networks and then use these features to inform choices” (Rithani et al. 2023: 14766).

A useful tool developed within the field of information extraction that has clear relevance for historians attempting to analyze large unstructured databases is named entity recognition and classification (Goyal, Gupta, and Kumar 2018). This is a group of text-analysis algorithms designed to identify meaningful information contained in a given document —for example, “person, organization, location, date/time, quantities, numbers” (Goyal et al. 2018: 22). The named entities may be specialized to a particular content area; for example, public health historians may wish to include disease and symptom names. These tools are used as a basis for pre-processing of a set of documents. The tool creates a meta-file for each document including the named entities and classes that it contains, along with other contextual information. For example, historians interested in the role that agriculture played in large periods of time may be interested in quickly identifying a selection of documents that refer to hunger, famine, or starvation. Goyal, Gupta, and Kumar carefully review the methods currently in use to identify named entities in a body of texts, including rule-based identification of named entities and machine-learning identification, with or without supervision. They emphasize that none of these methods is error-free, and false positives and false negatives continue to arise after training. This means that some lexical items in a document are either missed as referring to a named entity, or are incorrectly associated with a given named entity. Nonetheless, a historian can certainly use named-entity recognition and classification to provide a basis for important exploration and discovery in a large unstructured text database.

Keller, Shiu, and Yan (2024) provide a recent example of a machine-learning approach to automatic text analysis based on the most common large language model technique (“bidirectional encoder representations from transformers” (BERT)). They use GUWEN-BERT, a BERT model pre-trained on classical Chinese characters. They evaluate the power and accuracy of this tool in analyzing the Veritable Records of the Qing Dynasty to identify events of social unrest. The document archive is vast, encompassing more than 1,200 volumes of records from the sixth century to the end of the Qing Dynasty. Their research task is to identify episodes of social unrest, and then to classify these episodes into three categories—peasant unrest, militia unrest, and secret-society unrest (Keller et al. 2024: 4). This process of event identification and classification then permitted the researchers to seek out correlates of unrest, including fluctuations in grain prices. A useful example applying the same technology is provided by Liu, Wang, and Bol (2023), demonstrating largescale extraction of biographical information from a large collection of local gazetteers. Machine recognition of hand-written Chinese literary characters and translation of sentences and phrases in classical Chinese have made great progress in the past twenty years; Liu, Jaeger, and Nakagawa 2004, Leung and Leung 2010. This capability represents a major step forward in the ability of Chinese and Asian-language historians to make extensive use of large databases of historical documents such as the Veritable Records archives.

RAG, GraphRAG, and vector-similarity search

An important tool that has been of interest to historians exploring digital tools is retrieval-augmented generation (RAG) as a complement to LLM text generation systems. This area of research attempts to provide a basis for joining LLM query engines to specialized databases so that responses to queries will be based on data contained in the associated database. RAG tools are sometimes celebrated as solving two persistent problems arising in the application of natural-language generative chat functions based on large language models: the lack of auditability and the generation of fictitious responses (hallucinations) by the generative chat program. Kim Martineau describes a RAG tool in these terms: “Retrieval-augmented generation (RAG) is an AI framework for improving the quality of LLM-generated responses by grounding the model on external sources of knowledge to supplement the LLM’s internal representation of information. RAG implementation in an LLM-based question-answer system has two main benefits: It ensures that the model has access to the most current, reliable facts, and that users have access to the model’s sources, ensuring that its claims can be checked for accuracy and ultimately trusted” (Martineau 2024). A RAG framework is intended to allow the introduction of real, documented data into a natural language query-and-response system, and it is designed to be auditable. RAG picks up where pre-processing tools discussed previously leave off. RAG tools permit the retriever tool to parse a given query into component questions, and then to retrieve relevant data from pre-existing databases of documents (Lewis et al. 2021; Zhao et al. 2024).

RAG tools have in turn been extended with two related innovations. Vector similarity search is a semantic search tool that represents a document as a vector of abstract terms (like those identified in the discussion above of named entity identification and classification) (Mohoney et al. 2023). This further simplifies the task of querying the database for documents that are “about” one or more entities or events. A second valuable analytical tool is GraphRAG, which permits the construction of a network graph of the links among the elements in a document collection. Introduced by research scientists at Microsoft in 2024, GraphRAG was designed to permit analysis of global features of a large unstructured data collection. (See Larson and Truitt 2024, Edge et al. 2024a, and Edge et al. 2024b for technical descriptions of GraphRAG capabilities.) GraphRAG combines the data provided by RAG tools and connects these to LLM generative response systems. GraphRAG thus integrates indexing, retrieval, and generation. The key output of GraphRAG analysis of a database of text documents is a knowledge graph showing relationships among the various documents based on the content vectors associated with each document. (Experienced historians who make use of RAG and GraphRAG tools note that scaling up from moderate to large databases is challenging and computationally demanding.)

Limitations of the tools for historians

These tools suggest research strategies for historians confronting very large digital collections of documents and images. They permit computational procedures that classify and index the materials in the data archive that permit the historian to quickly identify items that are relevant to particular research questions -- the occurrence of famine, civil strife, dynastic unrest, or the transmission of ideas. And they permit natural-language query of the target database that provides suggestive avenues of further investigation for the historian. Crucially, these tools provide the ability to "audit" the results of a query by returning to the specific documents on which a response is based. The problem of "hallucination" that is endemic to large-language models by themselves is substantially reduced by tying responses to specific items in the database. And the algorithms of vector search allow the AI agent to quickly pull together the documents and "chunks" of text that are most relevant to the query.

These applications present powerful new opportunities for historians to make extensive use of very large databases of texts, but they also pose novel questions for the philosophy of history. In particular, they require that historians and philosophers develop new standards and methods for validating the computational methods that are chosen for various research tasks presented by the availability of large text collections. This means that we need to examine the strengths and limitations of each of these methods of analysis. Crucially, the designers and researchers of these tools are quite explicit in acknowledging that the tools are subject to error: the problem of hallucination is not fully removed, the content database itself may be error-prone, there may be flaws and limitations inherent in the training database in use, and any errors created during the information-extraction stage will be carried forward into the results. It is therefore incumbent upon the historian who uses such tools to validate and evaluate the information provided by searches and natural language queries. Nothing in the design of these tools suggests that they are highly reliable; rather, they are best viewed as exploratory tools permitting the historian to look more deeply into the collection of documents than traditional methods would permit. It will be necessary for historians to think critically about the quality and limitations of the information they extract from these forms of big-data analysis of historical databases.

References

Adnan, Kiran, and Rehan Akbar, 2019. “An analytical study of information extraction from unstructured and multidimensional big data,” Journal of Big Data, 6(1): 91. doi10.1186/s40537-019-0254-8

Allen, Robert C., 2000. “Economic Structure and Agricultural Productivity in Europe, 1300–1800,” European Review of Economic History, 3: 1–25.

Allen, Robert C., Tommy Bengtsson, and Martin Dribe (eds.), 2005. Living standards in the past: New perspectives on well-being in Asia and Europe, Oxford; New York: Oxford University.

Bol, Peter, 2006. “Creating the China Historical Geographic Information System,” in History in the Digital Age Symposium, University of Nebraska-Lincoln, video lecture. Bol 2006 available online.

Bol, Peter. 2007. Creating the China Historical Geographic Information System (text and slides). (Digital History Project, University of Nebraska-Lincoln). http://digitalhistory.unl.edu/essays/bolessay.php.

Crafts, N.F.R., 1980. “National income estimates and the British standard of living debate: A reappraisal of 1801–1831,” Explorations in Economic History, 17: 176–88.

Edge, Darren, Ha Trinh, Newman Cheng, Joshua Bradley, Alex Chao, Apurva Mody, Steven Truitt, and Jonathan Larson, 2024a, “From Local to Global: A Graph RAG Approach to Query-Focused Summarization”, manuscript at arXiv.org.

Edge, Darren, Ha Trinh, Steven Truitt, and Jonathan Larson, 2024b, “GraphRAG: New Tool for Complex Data Discovery Now on GitHub”, blog post at Microsoft Research, 2 July 2024.

Eve, Martin Paul, 2022. The digital humanities and literary studies, first edition, Oxford: Oxford University Press.

Fogel, Robert William, and G. R. Elton, 1983. Which road to the past? Two views of history, New Haven: Yale University Press.

Goyal, Archana, Vishal Gupta, and Manish Kumar, 2018. “Recent named entity recognition and classification techniques: A systematic review,” Computer Science Review, 29: 21–43. Goyal, Gupta, & Kumar 2018 available online.

Graham, Shawn, Ian Milligan, Scott Weingart, and Kimberley Martin, 2022. Exploring big historical data: The historian’s macroscope, second edition, New Jersey: World Scientific.

Kamath, Uday, Kevin Keenan, Garrett Somers, and Sarah Sorenson. 2024. Large Language Models: A Deep Dive: Bridging Theory and Practice. 1st 2024. ed.: Springer Nature Switzerland: Imprint: Springer.

Keller, Wolfgang, Carol H. Shiue, and Sen Yan, 2024. “Mining Chinese historical sources at scale: A machine learning approach to Qing state capacity,” Working Paper 32982, National Bureau of Economic Research, Cambridge, MA. Keller, Shiue, & Yan 2024 available online.

Larson, Jonathan, and Steven Truitt, 2024, “GraphRAG: Unlocking LLM Discovery on Narrative Private Data”, blog post at Microsoft Research, 13 February 2024.

Lee, James Z., and Wang Feng, 1999. One quarter of humanity : Malthusian mythology and Chinese realities, 1700–2000. Cambridge, Mass.: Harvard University Press.

Leung, K. C., and C. H. Leung, 2010. “Recognition of handwritten Chinese characters by critical region analysis,” Pattern Recognition, 43(3): 949–961.

Lewis, Patrick, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, and Douwe Kiela, 2021. “Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks,” manuscript available at arXiv.org.

Liu, Zhou, Hongsu Wang, and Peter K. Bol, 2023. “Automatic biographical information extraction from local gazetteers with Bi-LSTM-CRF model and BERT,” International Journal of Digital Humanities, 4: 195–212.

Liu, Cheng-Lin, Stefan Jaeger, and Masaki Nakagawa, 2004. “Online recognition of Chinese characters: The state-of-the-art,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 26: 198–213.

Martineau, Kim. 2024. "What is retrieval-augmented generation?". IBM Research. Accessed 11/23/2024. https://research.ibm.com/blog/retrieval-augmented-generation-RAG.

Mitchell, Melanie, Alessandro B. Palmarini, and Arseny Moskvichev, 2023. “Comparing humans, GPT-4, and GPT-4V on abstraction and Reasoning tasks,” manuscript at arXiv.org.

Mohoney, Jason, Anil Pacaci, Shihabur Rahman Chowdhury, Ali Mousavi, Ihab F. Ilyas, Umar Farooq Minhas, Jeffrey Pound, and Theodoros Rekatsinas. 2023. "High-Throughput Vector Similarity Search in Knowledge Graphs." arxiv>cs>arXiv:2304.01926.

Moretti, Franco, 2005. Graphs, maps, trees: Abstract models for a literary history, London; New York: Verso.

–––, 2013. Distant reading, London; New York: Verso.

Presner, Todd, and David Shepard, 2016. “Mapping the geospatial turn,” in A new companion to digital humanities, edited by Susan Schreibman, Raymond George Siemens and John Unsworth. Wiley/Blackwell.

Rawski, Thomas G. (ed.), 1996. Economics and the historian, Berkeley: University of California Press.

Rithani, M., R Prasanna Kumar, and Srinath Doss, 2023. “A review on big data based on deep neural network approaches,” Artificial Intelligence Review, 56(12): 14765–14801

Schreibman, Susan, Raymond George Siemens, and John Unsworth (eds.), 2004. A companion to digital humanities (Blackwell Companions to Literature and Culture), Malden, MA: Blackwell Pubications.

Schreibman, Susan, Raymond George Siemens, and John Unsworth (eds.), 2016. A new companion to digital humanities, Chichester, West Sussex, UK: Wiley/Blackwell.

Siebold, Anna, and Matteo Valleriani, 2022. “Digital perspectives in history,” Histories, 2(2): 170–177.

Skinner, G. William. 1977. "Regional Urbanization in Nineteenth-Century China." In The City in Late Imperial China, edited by G. William Skinner and Hugh D. R. Baker. Stanford, CA: Stanford University Press.

Skinner, G. William, Mark Henderson, and Yuan Jianhua. 2000. "China's Fertility Transition through Regional Space: Using GIS and Census Data for a Spatial Analysis of Historical Demography." Social Science History 24 (3): 613-652.

Thill, Jean-Claude (ed.), 2020. Innovations in urban and regional systems: Contributions from GIS&T, spatial analysis and location modeling, 1st edition, Cham: Springer International Publishing.

Wang, Dongbo, Chang Liu, Zhixiao Zhao, Si Shen, Liu Liu, Bin Li, Haotian Hu, Mengcheng Wu, Litao Lin, Xue Zhao, and Xiyu Wang, 2023. “GujiBERT and GujiGPT: Construction of Intelligent Information Processing Foundation Language Models for Ancient Texts,” manuscript at arXiv.org.

Zhao, Penghao, Hailin Zhang, Qinhan Yu, Zhengren Wang, Yunteng Geng, Fangcheng Fu, Ling Yang, Wentao Zhang, Jie Jiang, and Bin Cui, 2024. "Retrieval-Augmented Generation for AI-Generated Content: A Survey," manuscript at arXiv.org.

Stock ownership as system-wide exploitation?

 

A prior post made an effort to gain greater analytical clarity concerning the unfairness involved in the separation between the “one percent” economy and the rest of us. In what ways is the wealth owned by the super-billionaires an “unfair” extraction from the rest of US society? How can we account for the very rapid accumulation of wealth in the hands of the richest 1 percent of US wealth holders since 1980? The answer seems to largely turn on the rapid expansion in wealth represented by the US stock market over that period, and the fact that a very small number of wealth holders captured the lion’s share of these gains. The following graph shows a five-fold increase in the value of the US equity market in part of that time, from about $12 trillion in 1998 to $52 trillion in 2024. The wealth owned by the top 1% of households increased at about the same rate, which implies that this class rode the wave to wealth right along with the stock market in those years. “Corporate equities and mutual fund shares” are the largest component by far of the wealth portfolios of the top .1% and 1%, as reflected in the second chart below, produced by the Federal Reserve.

Screenshot

It was shown in the earlier post that the growth of the super-billionaires’ share of the nation’s wealth cannot be explained in normal “business profit” terms. (For reference, the top twenty billionaires in the US own 2.8 trillion dollars of wealth; link.) Rather, the bulk of the wealth now held by individuals like Mark Zuckerberg, Elon Musk, and Jeff Bezos represents the rapid appreciation of value in capital markets of the companies in which they have large ownership stakes. The companies themselves do not generate billions of dollars in dividends; rather, their total stock value has witnessed billions of dollars in gains over very short periods of time.

So why should we think this is in any way unfair? How is it exploitative? Is it not more like the fortunate visitor to “Antiques Road Show” who finds that the forgotten painting in the closet is in fact an early Picasso and is worth millions on the art market? This is good fortune for the owners of the canvas, but surely these facts don’t suggest “exploitation” of anyone else. Perhaps not in the case of the Antiques Road Show guest; but the majority owner of Amazon, Tesla, or Meta is in a different set of circumstances. Rather, the existence and continuing success of these companies depends on background conditions to which all sectors and components of the US economy contribute: a stable system of law and regulation, a robust education and research sector, a skilled workforce, an infrastructure of roads, ports, rail lines, fiber optic cables, and electricity providers. The value of US companies is at least in part a system effect: it is facilitated and constituted by a vast network of private and public stakeholders, all of whom contribute ultimately to the success of the company and the value it finds within the equity market. So the value of the US company is inseparable from the large and heterogeneous economic and political system in which it operates, and the increase in value over time of the US company reflects the continuing contribution expected by the investing public from the functioning of that system.

It will be said, of course, that the companies and their executives themselves contribute to the value that investors attribute to them: innovative products, good management systems, efficient decision-making, appropriate personnel practices, “entrepreneurship” and risk-taking. This is true. But it is also true that these contributions represent only a portion of the increase in value that the company experiences over time. The system effects described here represent an independent and important component of that substantial increase in value. So we might say that “system-created increase in value” is the uncompensated part of wealth creation in today’s economy. Companies pay little or nothing to cover the cost of these system-level inputs on which they depend; these are the inverse of “externalities”, in that they are benefits taken without compensation from the public (rather than harms imposed without compensation on the public). And these system-created increments in value constitute a very important part of the increase in value that they experience over time.

We might therefore look at “system-created increase in value” as the counterpart to “unpaid labor time” in the classic theory of exploitation. It is the source of wealth (profit) that the owners of wealth derive simply in virtue of their position in the property system and in their opportunity to benefit from the economic system upon which they depend. But now it does not derive from the “surplus value” contributed to profits by each worker, but rather from the synergies created by the socio-economic system as a whole.

It should also be noted that the ability of private companies to “extract” value from system-level inputs without compensation depends on their ability collectively to influence government policy. Therefore owners of private companies and stock wealth have strong incentives to shape the decision-making of elected officials, government policy makers, the fiscal system, and the regulatory process. This reinforces the arguments made by Thomas Volscho and Nathan Kelly in “The Rise of the Super-Rich: Power Resources, Taxes, Financial Markets, and the Dynamics of the Top 1 Percent, 1949 to 2008” (link). It follows, then, that achieving powerful influence on public policy and economic rule-making is not just a hobby for the oligarchy; it is an existential necessity.

This analysis of “system-input exploitation” has important consequences for distributive justice. If the whole of society contributes to the creation of the system-level properties that generate a significant fraction of the new wealth created in the past forty years, then surely fairness requires that all participants should receive some part of the gains. It would seem logical for the non-wealth-holding stakeholders — workers, farmers, and uncompensated contributors to social reproduction — to demand economic reforms that direct a fair share of that new wealth to the benefit of the whole population.

The previous post suggested one possible mechanism that would do this. The post discusses a hypothetical “public investment fund” that “would be automatically vested with ownership shares of businesses and corporations as they are created and grow, and that would function as a ‘wealth reserve’ for all citizens”. This would constitute a large and growing asset to be used for the benefit of the whole of society. In that discussion a distribution of gains resulting in public ownership of 1/3 of all capital was considered. Such a division would reduce (though not eliminate) the most extreme inequalities of wealth that currently exist, and would provide a financial basis for a more genuine “free community of equals” through the secure establishment of a high level of the resources most needed — healthcare, education and training, environmental protection, and provisioning of basic human needs for children, the disabled, the elderly, and the unemployed.

This idea of a public investment fund corresponding to the “systemic value creation” of the economy might go a long way towards the securing political values embodied in John Rawls’s concept of a “property-owning democracy” (link). Rawls argues that “the equal worth of liberty” is incompatible with a society in which political influence is proportional to wealth and where wealth is extremely unequally distributed. Wealth inequality of this magnitude means that the oligarch’s liberty and worth are magnified many times relative to the ordinary citizen’s situation. The creation of a substantial public investment fund representing the value created by our social, economic, and political system of cooperation would reduce the total proportion of the total value of the economy that the multi-billionaire class is able to expropriate. It would create real property entitlements for the great majority of society, and it would redress the current horrendous inequality of political influence that exists between the super-rich and the ordinary citizen.

A new form of exploitation

 

Much thinking about economic justice for working people has been framed by the nineteenth-century concept of “capitalism”: owners of enterprises constitute a minority of the population; they hire workers who represent the majority of the population; wages and profits define the distribution of income throughout the whole population. This picture still works well enough for a range of economic activities in the advanced capitalist economies when it comes to manufacturing, agriculture, and service industries. According to recent tabulations by the US Bureau of Labor Statistics (link), there were 158 million workers in wage and salary employment in 2023. Manufacturing represented 8.2%, retail and wholesale trade 13.7%, information 1.9%, financial services 5.8%, leisure and hospitality 10.5%, and federal and state government 14.4%. This adds up to 54.5% of the US labor force, and these workers and firms can be thought of in roughly the framework offered by the traditional idea of “capitalism”. Many of these workplaces are amenable to union representation (though relatively few are in fact unionized). But improving access to union rights and workplace consultation would significantly improve the conditions of life for this segment of the US population.

Marx’s view of the unfairness of capitalism, then, comes down to workplace exploitation — the capture of “surplus value” by the firm’s owner from the workers whom he or she employs. Profits derive solely from surplus value, so wealth accumulation is fundamentally limited by the size of an enterprise.

However, current realities seem to suggest that this classical Marxist account is no longer sufficient. To see this point it is crucial to look at the details of the distribution of wealth and income in the U.S. Consider the graph of median US income by quintile above in constant 2018 dollars. Since 1989 only the top quintile of household income has demonstrated significant growth (in a timeframe of more than thirty years); and the top 5% of households shows the greatest increase of any group. 80% of US households are barely better off today than they were in 1967; whereas the top 5% of households have increased their incomes by almost 250% in real terms. The bottom 80% range in household income from “poor”, the bottom 20% at an average household income of about $14,000, to the second quintile (60%-80%) of about $102,000. But virtually all of these households — 80% of all households — earn their livings through wage and salary income, in “capitalist” workplaces.

Further, only a very small fraction of these households are in a position to accumulate significant savings or investments. As the second graph shows, the bottom 50% of households have only 2.6% of all U.S. wealth, and the 50%-90% segment owns only another 30.8%. The top 0.1% owns 13.9% of all wealth, and the remainder of the top 1% owns 16.6%. That amounts to 30.5% of all wealth, held by 1% of households — and almost incomprehensible figure.

These two graphs have a very clear, unmistakable implication: that working people, including service workers, industrial workers, and most professionals have received a declining share of the economic product of the nation over the past 40 years. (Amazon warehouse workers fall in the 2nd-lowest quintile (poorest 21-40%).) Further, the vast majority of U.S. residents have only a tiny share of all property in the U.S. According to the Federal Reserve 2022 Survey of Consumer Finances, median household net worth in 2022 was $192,700, including private savings, retirement savings, and personal property and home value (link). And, of course, this implies that the median household net worth of the bottom 80% of the U.S. population is significantly lower.

It seems apparent, then, that capitalist exploitation is no longer the primary mechanism through which wealth is accumulated by the top 10%, 1%, and .1% of wealth holders. The top group gains income at a rapid rate and increases its share of the national wealth comparably; whereas the bottom 80% have almost stagnant incomes and negligible wealth. And this accumulation occurs almost entirely through rising value of the stock issued on behalf of private companies. The national economy generates all of this wealth; but the vast preponderance of the fruits of this production flow to the top 10% and 1% of wealth holders. This is a different kind of exploitation: not exploitation of a specific group of workers (employees of General Motors, for example); but exploitation of the whole of the U.S. economy for the benefit of a tiny minority of wealth holders.

Essentially it seems fair to say that the contemporary U.S. system involves two economies — one that includes 60%-80% of all people, and who depend on wages and salaried income to earn their livings; and a second economy that is itself steeply stratified, involving only the top 10%-20% of households. This second economy includes highly paid professionals, executives, and individuals who derive a substantial income from investments, financial assets, and other capital assets. The distribution of income and wealth in this second economy depends on ownership of capital (including human capital) of increasing value in a “knowledge” economy.

It appears, then, that the gross advancement of wealth inequalities in the past three decades has little to do with traditional “exploitation” – an unfavorable wage relationship between owners and workers. Instead, the sudden explosion of tech-oligarchy in the US seems to have to do with financial markets, the stock value of private companies, and the environment of business and tax policy in which they operate. The super-wealthy class in the US came into multi-billionaire status through the rapid acceleration of market value of companies like Amazon, Tesla, and Facebook/Meta. And this process reflected a macro-level mechanism that we might describe as “exploitation of the US economy as a whole” rather than “exploitation of a specified group of workers employed by these companies.

Thomas Volscho and Nathan Kelly provide a careful analysis of the dynamics of income inequality in the US economy over time in “The Rise of the Super-Rich: Power Resources, Taxes, Financial Markets, and the Dynamics of the Top 1 Percent, 1949 to 2008” (link). They note that there was considerable variation in the share of income flowing to the top one percent between 1900 and 2020, with a rapid rise beginning in about 1980. And they attribute much of this variation to facts about political power, public policy, and fiscal legislation. (This bundle of hypotheses is referred to as “Power Resources Theory”.) And a key finding in this literature is that the relative levels of political power and influence held by economic elites versus working people have a very large effect on the degree and direction of change in inequality at the top.

Consider the short history of Facebook. Mark Zuckerberg’s wealth increased from 2008 from $1.5 billion to $236 billion in 2025. The employee count of Facebook/Meta increased comparably during that period, from 85 employees in 2008 to 76,800 employees in 2025. But Zuckerberg’s wealth does not reflect the “surplus value” created by these workers, but rather the perceived value of the company in the eyes of private and institutional investors. And critically, it is difficult to imagine institutional changes within Facebook/Meta that would lead to greater overall societal equity simply by providing the company’s workers more input into the management of the company. The median income for a Facebook/Meta worker is $257K – hardly an exploitative wage. It is the rest of society that is disadvantaged by Zuckerberg’s $236 billion, not the direct employees. The same seems to be true for Tesla and the wealth accumulated by Elon Musk and for Amazon and the wealth of Jeff Bezos. Amazon’s business operations have many of the same features of domination and exploitation identified by Engels in Manchester; but these operations do not constitute the fundament of Bezos’s wealth except perhaps for the “performative” of a company single-mindedly devoted to efficiency and speed of operations.

The experience of the reforms of the welfare state after WWII shows that capitalist exploitation can be reformed through measures that improve the public provision of some crucial services (education, healthcare, retirement income, unemployment insurance); improve the ability of workers to represent themselves effectively in the workplace (legislation ensuring unionization rights); and improve conditions of health and safety in the workplace (OSHA protection). These reforms are “redistributive” in the sense that they depend on taxation of income and profits of private individuals and corporations to fund public provisioning. But can reforms like these address the inequalities — economic and political — created by the two economies described here? Can the oligarchy economy be reined in? It would seem that the answer is “no”.

So we are forced to ask, what kinds of fiscal and tax reforms could effectively rein in the wealth inequalities created at the very top of the wealth distribution? The annual wealth taxes proposed by progressive Democrats extend to taxes in the range of 1%. But this would represent a negligible reduction in the oligarch’s portfolio, and does essentially nothing to reduce the steepness of the distribution of wealth in America. A “confiscatory” tax of 33% would have a measurable effect by increasing available public funds for expenditure; but even reducing Elon Musk’s wealth from $368 billion to $245 billion – still results in a staggering inequality relative to 99% of US workers. And this still leaves the wealth-holder with a million-fold advantage in his/her political and media influence relative to almost all other US persons. (As mentioned above, the median net worth of all Americans is currently about $192,000. It is of course striking that three of America’s largest tech-oligarchs privately own a media company: Zuckerberg (Facebook), Musk (X/Twitter), and Bezos (the Washington Post).)

It appears, then, that standard “New Deal” or “welfarist” approaches to greater economic equality have no prospect for success whatsoever when it comes to reducing the overwhelming inequalities of wealth that exist between the two US economies described here. A graduated income tax works to moderate income inequalities (when it works at all); but the rapid accumulation of wealth represented by the emergence of the “tech-oligarchy” and the graph of wealth distribution above do not derive from income inequalities. The richest 1% did not primarily gain their wealth through annual savings from their high salaries; rather, they gained their wealth through stock ownership in companies whose value appreciated exponentially during the time of their ownership. And taxing the holders of wealth on the income generated by their holdings does not materially affect the distribution of wealth across the population and across generations.

Suppose we viewed a national economy as an interconnected and highly complex form of “joint production”, in which the efforts of all parties are instrumental in the creation of the new wealth and prosperity of the economy. And suppose we believe that this system should be organized as a “fair system of cooperation” in which all parties benefit in a fair way. Can the workings of capital markets and financial systems be incorporated into our institutions in ways that would give the working public (the 80%) a fair share of the products of cooperation? Could we imagine a fiscal mechanism that would provide the public with a “fair share” of the U.S. economy as a whole, including the growth of the value of private companies (Caterpillar, General Motors, Krogers, Facebook/Meta, Microsoft, …)?

For example, would it be possible to imagine a public investment agency along the lines of CalPERS that would be automatically vested with ownership shares of businesses and corporations as they are created and grow, and that would function as a “wealth reserve” for all citizens? Suppose the hypothetical “public investment corporation” eventually possessed assets worth about 1/3 of the total value of the US stock market. (The value of stocks listed on the New York Stock Exchange is currently $28.3 trillion, so we are imagining a public wealth fund of about $10 trillion.) On this model, private owners and shareholders would own 2/3 of the capitalized economy, and the public would own 1/3. Would such a system be feasible? Could such a system redress the insupportable economic and material inequalities that exist in our country? Could it redress the gross inequalities of influence and power that exist between a tiny class of oligarchs and the vast majority of democratic citizens? Could the shareholder voting rights that correspond to the public shares created in this way serve to alter corporate priorities?

It seems clear that the photo below taken from Donald Trump’s inauguration represents a horrendous flaw in contemporary democracy. The “tech oligarchs” turned out in force for the new administration, and a group of wholly committed political partisans stand behind them to enact policies in the United States that serve their interests. If this is the best that our democracy can currently offer working people, then we need to work much harder at finding political and economic solutions that can elicit broad support from ordinary citizens, workers, farmers, and Uber drivers to push forward a better agenda for democratic equity.