Source-linked AI summary
Individuals, Institutions, and Innovation in the Debates of the French Revolution
Alexander T. J. Barron, Jenny Huang, Rebecca L. Spang, Simon DeDeo
TL;DR
The paper asks how ideas and speech patterns entered, persisted in, or disappeared from the French Revolution’s first parliament, and how individuals and institutions shaped that process. It analyzes reconstructed speeches using information theory and finds that left-wing speakers often innovated while right-wing figures maintained existing patterns. Committees also became sources of new patterns and extra-rhetorical power.
Problem
The study examines how patterns of speech circulated in revolutionary debates, complementing analyses that focus on the evolution of particular ideas and arguments.
Source: Discussion · p. 6Method
The authors quantify language-pattern transmission in 44,913 reconstructed Assembly speeches, using topic modeling and information-theoretic measures.
Source: Results · pp. 2–3, Materials and Methods · p. 7Results
Right-wing speakers often maintained established patterns while left-wing speakers produced a wider spectrum of innovations, only some of which persisted.
Source: Discussion · p. 6Takeaways & Limitations
Both individuals and institutions shaped parliamentary development, with committees becoming sources of new speech patterns and extra-rhetorical power.
Source: Discussion · pp. 6–7Individual strategy and reception could depart from system-level innovation bias, including when prestige or social power helped preserve a line of argument.
Source: Organizational roles and individual… · p. 3
Abstract
from arXiv · showhide
The French Revolution brought principles of "liberty, equality, and brotherhood" to bear on the day-to-day challenges of governing what was then the largest country in Europe. Its experiments provided a model for future revolutions and democracies across the globe, but this first modern revolution had no model to follow. Using reconstructed transcripts of debates held in the Revolution's first parliament, we present a quantitative analysis of how this system managed innovation. We use information theory to track the creation, transmission, and destruction of patterns of word-use across over 40,000 speeches and more than one thousand speakers. The parliament as a whole was biased toward the adoption of new patterns, but speakers' individual qualities could break these overall trends. Speakers on the left innovated at higher rates while speakers on the right acted, often successfully, to preserve prior patterns. Key players such as Robespierre (on the left) and Abb\'e Maury (on the right) played information-processing roles emblematic of their politics. Newly-created organizational functions---such as the Assembly's President and committee chairs---had significant effects on debate outcomes, and a distinct transition appears mid-way through the parliament when committees, external to the debate process, gain new powers to "propose and dispose" to the body as a whole. Taken together, these quantitative results align with existing qualitative interpretations but also reveal crucial information-processing dynamics that have hitherto been overlooked. Great orators had the public's attention, but deputies (mostly on the political left) who mastered the committee system gained new powers to shape revolutionary legislation.
Results
The paper treats speech patterns as cultural evolution: patterns circulate, compete, and are copied between speakers. It uses Kullback-Leibler divergence to measure how speech patterns surprise an observer and to track their persistence over time.
Results
- Cultural evolution here is the circulation, selection, and differential propagation of speech patterns across a group.This focuses on shared language patterns rather than the ideas of a single individual.Source: Results · pp. 2–3
- Kullback-Leibler divergence measures the surprise produced when a pattern violates expectations formed from another pattern.The paper uses this information-theoretic measure to study rhetorical influence and attention over time.Source: Results · p. 3
- Novelty measures how unexpected a speech’s patterns are given the past, while transience measures how little those patterns persist into the future.The same surprise framework is applied to both past and future speech patterns.Source: Results · p. 3
Innovation Bias
Novelty is associated with influence in the Assembly: novel speeches are more resonant than an equilibrium system would predict. This innovation bias persists for about a day, but fades at longer timescales.
Innovation Bias
- At w = 7, novelty and transience often align, but many speeches depart sharply from that pattern, including top-quartile-novelty speeches with bottom-quartile transience.Thus, “what is new is quickly forgotten” describes the average, not every speech.Source: Innovation Bias · p. 3
- From w = 1 to w ≈ 100, positive Γ indicates that resonance increases with novelty, an innovation bias lasting at least a day before fading at longer timescales.Γ is the slope of the novelty–resonance relationship.Source: Innovation Bias · p. 3
Organizational roles and individual strategies
Individual speakers and institutional roles could depart from the Assembly’s overall innovation bias. Committees and speakers on the left were associated with novel patterns, while some prominent conservatives exerted influence by preserving existing patterns.
Organizational roles and individual strategies
- Speakers could choose whether to introduce or copy patterns, but their political, social, and rhetorical traits also shaped how audiences received those patterns.Prestige or social power could help a speaker maintain a line of argument with low novelty and high resonance.Source: Organizational roles and individual… · p. 3
- The Assembly assigned speakers institutional functions: the President enforced the daily agenda, while committees deliberated privately and presented content for public review.Committee proxies’ new-item speeches introduced official content, typically articles and draft decrees, to the Assembly floor.Source: Organizational roles and individual… · p. 3
- The analysis compared novelty, resonance, and resonance beyond that predicted by novelty across orators, Presidents, and committee proxies at multiple timescales.At w = 36, Δz(R) measures mean resonance relative to the OLS-predicted mean given novelty; the figure also depicts novelty against transience and resonance against novelty.Source: Organizational roles and individual… · p. 4, Figure 1 · p. 4, Table 1 · p. 5
- Committees had above-average novelty and resonance but lower resonance than expected, while the President had below-average resonance and little lasting influence on later speech patterns.Committee speech could introduce new items to the floor or engage with delegates after an item was introduced.Source: Organizational roles and individual… · pp. 3–4
- Thirty of the top forty speakers significantly departed from aggregate patterns, and high novelty was overwhelmingly associated with the left and bourgeoisie while right-wing speakers were low-novelty.Twenty-three speakers showed deviations at p < 10^-3.Source: Organizational roles and individual… · p. 4
- Robespierre and Pétion had the Assembly’s highest average resonance, while Camus and Vernier combined high novelty with low resonance and Maury and Cazalès combined low novelty with high resonance.The latter pattern was characteristic of the right wing, which achieved system influence despite anomalously low novelty.Source: Organizational roles and individual… · pp. 4–5
The Emergence and Evolution of the Committee
The committee system’s function shifted around late 1790, from privileged guidance of debate to a role where highly novel reports could pass with little discussion.
The Emergence and Evolution of the Committee
- 31 October 1790 was the modal best-fit change-point, and the two-epoch model outperformed single-epoch and linear-shift alternatives under AIC.The model allowed committee-function parameters to vary between two epochs.Source: The Emergence and Evolution of the… · p. 5
- Before the shift, new-item committee speeches matched system-wide resonance, while committee members gained a resonance bonus when speaking in debate.The two speech types thus differed in their initial ability to guide subsequent debate.Source: The Emergence and Evolution of the… · p. 5
- Afterward, highly novel new-item speeches had anomalously low resonance, associated with committee reports increasingly accepted with minimal discussion.The figure identifies this high-novelty, low-resonance pattern as the “undebated tail.”Source: The Emergence and Evolution of the… · p. 5, Figure 2 · p. 6
- For hand-annotated “passed without debate” cases, mean z(N) was 1.57±0.16 and mean z(R) was 0.14±0.13, matching the high-novelty, low-resonance pattern.The authors interpret this as further validation of the “undebated tail” as a signal of emergent committee power.Source: The Emergence and Evolution of the… · p. 6
- Committee members retained a privileged role in fixing and propagating patterns when committee matters were substantively debated.In-debate committee speech continued to have this role even as new-item speech acquired the undebated-tail pattern.Source: The Emergence and Evolution of the… · p. 6
Discussion
The discussion distinguishes rhetorical influence from institutional power and describes how both shaped the transmission of patterns and the management of debate in the NCA.
Discussion
- The authors focus on transmission of language patterns rather than the differential propagation of ideas, offering a new view alongside analyses of particular arguments.They frame this as a complementary window into existing qualitative analyses.Source: Discussion · p. 6
- Conservatives maintained established conversational patterns, while left-wing speakers generated broader innovation; Robespierre was high-novelty and high-resonance.These roles emerge from patterns of language use without first considering the ideas in the speeches.Source: Discussion · p. 6
- Rhetorical skill gave orators such as Robespierre, Pétion, and Maury influence, while committees became workhorses contributing specialized knowledge and extra-rhetorical power.The discussion identifies these as distinct routes through which individuals and institutions shaped the parliament.Source: Discussion · pp. 6–7
- Committees regulated debate and maintained continuity of principles, but their private work contrasted with the public spectacle of democratic debate.The authors connect committees’ growing influence to information management in bodies too large for direct participation.Source: Discussion · p. 7
- By 1793, committees such as the Committee of Public Safety had effectively replaced the republican government and executed thousands of people labeled enemies of the revolution.The authors describe this as a later expansion of committee power beyond the NCA’s earlier information-processing functions.Source: Discussion · p. 7
Conclusion
The conclusion frames cultural change as involving not only ideas, but also the mechanisms and roles that transmit them. In the NCA, quantifying pattern flow reveals political struggles over conversational direction and collective information management.
Conclusion
- Human culture involves emerging information-processing mechanisms and roles that create and selectively propagate ideas through time.The conclusion calls for attention to shifting transmission mechanisms as well as to the characteristics selected.Source: Conclusion · p. 7
- Quantifying pattern flow between speakers frames revolutionary debate as a struggle over the direction of conversation, not only a battle of ideas.Source: Conclusion · p. 7
- NCA political players adopted different roles in pattern propagation, and together they invented mechanisms for collective information management.The conclusion describes their interaction as both cooperative and agonistic.Source: Conclusion · p. 7
Materials and Methods
The study analyzes a reconstructed corpus of NCA speeches, representing each speech through topic modeling and recording speaker and institutional roles.
Materials and Methods
- The corpus contains 44,913 speeches and 4,765,773 words, reconstructed from primary-source records and covering the NCA from 9 July 1789 to 30 September 1791.After stop-word removal, speeches were represented as count vectors over the 10,000 most common words.Source: Materials and Methods · p. 7
- Each speech was matched to a named speaker and could be tagged as presidential or as committee speech introducing a report or commenting on it.All delegates and speakers in the data were male.Source: Materials and Methods · p. 7
- LDA modeled speeches as probability distributions over K = 100 topics.Source: Materials and Methods · p. 7
Novelty, Transience, Resonance
The section defines novelty and transience as surprise measured against past and future speech patterns, respectively. Their difference, resonance, indicates whether a novel pattern influences subsequent discourse.
Novelty, Transience, Resonance
- Novelty averages a speech’s surprise against earlier speeches across a window, extending analysis beyond a single exchange.At the smallest timescale, novelty is measured by the KLD of the current speech pattern against the immediately preceding speech.Source: Novelty, Transience, Resonance · p. 7
- A speech can be surprising relative to its past yet transient if later speakers do not retain its new patterns.An ignored interjection illustrates this case: it is surprising both before and after, producing symmetric surprise under time reversal.Source: Novelty, Transience, Resonance · p. 7
- Resonance is the positive asymmetry in which novelty exceeds transience, indicating a speech that breaks from the past and influences the future.The paper defines resonance as novelty minus transience, with transience measured as novelty under time reversal.Source: Novelty, Transience, Resonance · p. 8
- Novelty’s effectiveness, Γ, is the rate at which resonance increases with novelty, approximated using a linear model.For a system in equilibrium, expected novelty and transience are equal; significant deviations can indicate a non-equilibrium system.Source: Novelty, Transience, Resonance · p. 8