Excerpt: Abraham Lincoln identified public sentiment as the deepest force in democratic politics — more powerful than statutes or court decisions, because it determines whether either can hold. America’s Plan puts sentiment first in the issue pipeline for exactly that reason. This article explains what that means in practice, and why affected-party knowledge is the foundation that makes sentiment durable.
“In this age, in this country, public sentiment is everything. With it, nothing can fail; against it, nothing can succeed. Whoever molds public sentiment goes deeper than he who enacts statutes, or pronounces judicial decisions.”
— Abraham Lincoln
The deeper lever
Lincoln said this in 1858, during the Lincoln-Douglas debates, and he was making a structural observation about democratic power — not a rhetorical flourish. His point was not that majorities vote, or that public approval is nice to have. It was that public sentiment determines what is politically possible, what law can actually be enforced, and what changes endure versus what gets reversed. Statutes passed against entrenched sentiment tend not to hold. Judicial decisions issued against it tend to be resisted, circumvented, or eventually reversed.
The historical record bears this out. Prohibition was enacted by constitutional amendment in 1919 — the highest legislative threshold in American law. It was repealed fourteen years later, not because the constitutional process failed, but because the sentiment that had supported it never ran as deep as the coalition that passed it believed. Brown v. Board of Education was a legally decisive ruling in 1954. It required decades more of organizing, litigation, and political pressure to produce meaningful desegregation, because the sentiment in much of the country was actively opposed and institutions found ways to absorb the ruling without complying with its intent. The law moved; the sentiment took longer. So did the change.
The labor movement of the 1930s understood Lincoln’s point from the other direction. The Wagner Act (1935) and the legal infrastructure of collective bargaining did not create the labor movement’s power — they ratified it. By the time those statutes passed, organized workers had already built the sense, across millions of people and dozens of industries, that workers’ rights were a legitimate civic claim. The legislation followed the sentiment shift; it did not produce it. The NAACP’s decades of documentation work — counting lynchings, building the evidentiary record, training lawyers — did the same thing in a different register: it shaped what the American public could see, and therefore what could be argued, and therefore what became politically possible.
This is what America’s Plan is trying to do. Not advocate for specific policies from the top, but do the deeper work Lincoln identified: build the structured civic process through which public understanding of civic problems gets formed. That is why Sentiment is the first stage of the issue pipeline. Not as a data-gathering step that precedes the “real” work. As the foundation that everything else rests on.
What “public sentiment” means here — and what it does not
The word sentiment covers a lot of ground in current usage, and most of what it covers is not what Lincoln meant and not what this platform is after.
Survey polling draws a sample, asks fixed-choice questions in a standardized way, and reports aggregate percentages. It measures what people prefer when presented with a specific framing. It is fast and quantifiable. It is also thin: it captures preference, not knowledge, and it is entirely dependent on the quality of the questions asked — which means it is dependent on who wrote the questions and what they chose to measure.
Social media sentiment analysis aggregates what people post publicly, often processed by algorithmic tools to identify emotional register and topic frequency. It is thinner still. Social media skews toward people who post publicly, toward strong emotions, toward topics that generate reaction rather than reflection. It measures what is loud.
Affected-party testimony is what people who live with a problem know from that experience. It is specific, situated, and knowledge-dense. It includes the exact mechanism by which a broken system produces harm, the workarounds people have developed, the things that have been tried locally and failed, the trade-offs that are invisible from outside. It is not easily quantifiable, but it is far more useful as the foundation for plan design — and it is far more useful as the raw material for building the kind of public understanding that actually lasts.
The Sentiment stage of this platform is after the third. The first two are not useless, but neither can substitute for it.
Polling does not mold sentiment — it samples what has already been shaped
Lincoln’s phrase “whoever molds public sentiment” points to something that polling actively obscures: the question of where sentiment comes from, and who shaped it before the poll was run.
Polling doesn’t mold sentiment. It samples it. And what it samples is often sentiment that has already been shaped by whoever controlled the question — which institution commissioned the survey, which framing was chosen for the question, which response options were offered. A poll asking “do you support expanding rural healthcare access?” will return majority support. It will not reveal that the phrase “rural healthcare access” means something very different to the person who drives two hours to a specialist than to the person who saw a news story about it. The question collapses that difference before the data is collected.
This is the polling problem in structural terms: polls can only surface responses to questions written in advance, by someone who had a particular framing in mind. If the question frames the issue incorrectly, the data confirms the wrong frame. Affected parties often know the framing is wrong before the poll is run. But polling has no mechanism for surfacing that knowledge.
There are four specific limitations that matter for civic use.
The question defines the answer. If the question is wrong, the data is wrong — and wrong in a way that looks authoritative.
It aggregates across people with different levels of proximity. A national poll gives equal statistical weight to someone navigating a broken system every day and someone who encountered it once in a news story. Aggregation hides this difference. The resulting “majority position” may primarily reflect the views of people with the least direct experience of the problem.
It measures preferences, not knowledge. Polling tells you what people say they want when asked. It does not tell you what they know, what they have tried, what has failed, or what the actual costs of different approaches look like. Policy requires the latter.
It cannot surface what people don’t know they know. Much of the knowledge held by affected parties is tacit — it exists in experience and practice rather than in positions that can be retrieved by a multiple-choice question. It emerges from extended conversation. A survey cannot ask “what have you noticed that researchers have missed?” and get a useful answer. A facilitated deliberative discussion can.
When policy is designed from polling alone
The pattern is consistent. Policy designed primarily from polling tends to address what people say they want — in a standardized, pre-framed way — rather than what would actually improve the situation.
Drug policy over the past four decades is an example with a long enough record to evaluate. Throughout the 1980s and 1990s, polling consistently showed majority support for harsh penalties for drug offenses. That polling data was cited as evidence of public will. What the polls could not surface: what people living in heavily policed communities actually experienced, what the collateral effects of mass incarceration looked like in practice, whether the enforcement approach was producing the outcomes it claimed, and what alternatives had been tried at the local level. The knowledge that would have complicated the policy — or ruled out certain approaches before they were tried — lived with affected parties. It was not in the polling data.
The same majorities who supported those policies in the 1980s shifted against them by the 2000s, after the consequences became visible enough to enter mainstream coverage. The policy failed in ways that were predictable from affected-party knowledge and invisible from polling. That is not an argument that majority opinion does not matter. It is Lincoln’s argument: sentiment that is not grounded in accurate understanding of the actual problem will not hold, because the problem will keep producing evidence against it.
What the Sentiment stage is actually doing
The Sentiment stage of the issue pipeline is the mechanism by which affected parties exercise the power Lincoln identified — the power to shape how a problem is understood — rather than having that power exercised on their behalf by institutions with different interests.
This is what the platform’s Core Ideas mean when they say affected parties should have the central role in defining a problem: not that their preferences should be polled, but that their knowledge should be the foundation from which the public’s understanding of the problem is built. That is a different claim, with different implications for how the work gets done.
What a well-run Sentiment stage produces that polling cannot:
Named mechanisms. Not just “the system is failing” but a specific account of how it fails — the sequence of events, the decision points, the places where the process breaks down. This is the raw material that the Analysis stage needs to identify causes.
Identified variation. The same problem often manifests differently in different communities, economic circumstances, and geographic contexts. A national poll collapses that variation. Affected-party testimony preserves it, which matters for designing interventions that can work in specific conditions.
Known failure modes. What has been tried, what failed, and some account of why. Affected parties have often already run the experiment that policymakers are proposing. That knowledge is worth having before the proposal is finalized.
A foundation for analysis and plan-building. You cannot do meaningful causal analysis from polling percentages. You can from a documented body of affected-party accounts, because those accounts contain the causal language — the “because” and “which caused” and “and then” — that percentage-point data strips out.
This is also where deliberation does its work: the Sentiment stage is not a stack of individual testimonies placed next to each other. It is a structured process in which accounts are compared, questioned, and refined until the group has a clearer picture than any single participant brought in. The output is not a percentage. It is a documented understanding — specific enough to build a plan from, accurate enough to build durable public support around.
What this approach does not claim
Affected-party testimony is not representative in the statistical sense. It reflects who participates, which is a self-selected group. The platform does not claim to produce a scientifically representative sample of public opinion, and it should not be mistaken for one.
What it produces is a knowledge base — specific, situated, and correctable. It is more useful as the foundation for plan design than polling data can be. And it is more likely to produce sentiment that holds — in Lincoln’s sense — because it is built from the accurate understanding of the problem that people who live with it carry.
The Accountability Stage eventually asks whether the plan produced what it promised. Getting that question right depends on having gotten the Sentiment stage right — on having built the plan from an accurate picture of what the problem actually was, not from what a standardized question said people preferred when asked.
This article was researched and drafted with AI assistance under human review. See our full AI and editorial practices.