Abstract: (Major release draft, 31 Oct 2026). This reality-first account asks what a person is epistemically entitled to believe. It defends the idea that belief is not possession of truth. Reality is determinate; truth is alignment with reality. Human belief is different; belief needs justification, credence, and recalibration. Empirical truth requires successful testing against reality. Rational truth also must have valid empirical application. Justification of a belief is the accumulation and evaluation of evidence. Potential empirical ideas attempt to describe the material world directly and therefore must be empirically testable: falsifiable. Potential rational ideas attempt to describe the material world indirectly and therefore must be capable of coherent application to it. Justification is not the same as credence basis. Calibration needs to be a recurring process. The position developed here originated primarily in Chapters 10, 18, 20, 22, and 27 of 30 Philosophers, published in 2024.
Introduction: Belief
We believe things all day long, but why?
We believe the weather report. We believe what we read. We believe what our friends tell us. We believe our memories.
And every belief shapes what we do next. Beliefs guide our votes. Our investments. Our parenting. Our relationships. Our fears. And bad beliefs don’t just sit quietly in the mind. They cause great harm. They can cost money. They strain families. They ruin trust. They start wars.
So the real question is:
What should we believe?
And even deeper:
What makes a belief justified? How strongly should we believe it? When and how do we adjust?
If we are going to let belief guide our lives, we need disciplined answers.
Metaphysics, Truth, and Belief
Before we jump into belief, let’s review metaphysics and truth. The Split establishes that reality and our ideas about reality are different things. The theory of truth establishes the success condition: an idea is true when it aligns with reality. And the empirical, rational, and irrational categories of the Idea of Ideas tell us what kind of idea we are evaluating and therefore what sort of support is appropriate.
As established in the TST metaphysical framework, because induction cannot logically guarantee the future (see Hume in paper 1 of this series), TST treats empirical regularity not as an absolute certainty, but as a provisional, revisable working assumption.
The material world exists whether we like it or not. It does not bend to opinion. It does not adjust to preference. It does not care what we hope is true. That’s nature. It is how things actually are, objective reality includes nature’s own absolute truths. Causes produce effects. Gravity pulls. Fire burns. Bodies age. But here’s the important distinction: those absolutes exist in the world, not in our minds. Nature exists whether what we experience is base-level reality or a simulation.
Much of epistemology evaluates beliefs already held: it asks whether they are justified, how strongly they are held, and whether that confidence should be revised in light of the evidence. These questions are often analyzed separately under topics such as justification, credence, and calibration. There is value in integrating them into a single practical process. In that way, this approach remains firmly within the great normative tradition of epistemology: not merely describing what people believe, but asking:
What people ought to believe, how strongly people ought to believe it, and when people ought to change their minds.
In my writing, applied philosophy is the practical expression of normative philosophy: taking questions about how we ought to think and act and putting them to work in everyday life.
Let’s start this theory of belief by explaining the relationship between truth and belief plainly. Our belief in something is not possession of truth. Truth is alignment with reality; our degree of belief that an idea is true exists in the mind and is called credence. And one’s credence has a basis, what we’ll call a credence basis here.
One’s actual belief has a simple architecture:
A basis informs a credence.
That is the descriptive side of doxastic formation.
Now, epistemically responsible belief is different. Responsible belief requires normative epistemic evaluation. It requires justification, appropriate credence, and recalibration.
The architecture is precise:
Evidence and reasoning provide justification, justification supports a degree of rational credence, and recalibration allows you to adjust.
The normative principle of this theory is that a belief should be held to the degree that the evidence and reasoning appropriately support the kind of idea being evaluated. In practice, a person’s actual credence often diverges from that rational standard, and this theory accommodates that reality. The accommodation is descriptive, not relativistic: it identifies the divergence without treating competing credences as equally warranted.
And one’s belief system can be audited and even strengthened through public truth systems, especially the Three Truth Hammers of science, law, and journalism. These systems serve public truth by preserving evidence, exposing claims to criticism, and testing them beyond personal bias. This approach can also help individual judgment contribute responsibly to public truth through the public truth systems. Public truth influences public belief but they are not the same thing. See the paper titled TST Doxastic Formation: Public Belief, Tribe, and Worldview.
This theory of belief concerns how strongly a mind should hold an idea as true or truthlike under a reality-first framework. In practice, this means identifying one’s actual credence and its basis, evaluating the available justification, identifying the rational credence that justification warrants, and then deliberately deciding whether one’s actual credence should be revised. The result is a fallibilist model of rational belief that resists dogmatism, relativism, pure rationalism, and skeptical paralysis. In plain terms:
Use your best evidence, stay open to being wrong, and embrace that some beliefs are better supported than others.
The three core topics of TST’s theory of belief are justification, credence, and recalibration.
First up, justification.
Epistemic Justification
Justification here is used in the traditional sense. Justification for belief concerns the normative quality of the evidence and reasoning supporting an idea. And because evidence and reasoning vary in strength, the strength of justification is graduated rather than simply binary. An idea itself can still be true or false; what varies is how justified we are in believing it. The stronger the evidence and reasoning, the more confidence we are justified in placing in the belief.
In practical terms, evaluating justification often begins by steelmanning an idea: constructing the strongest fair version of the argument using the best available evidence and reasoning. A strawman does the opposite: it weakens or distorts a position, making it easier to attack. In general, one supports critical thinking, one does not:
Steelmanning clarifies; strawmanning distorts.
A steelman argument is not necessarily true or justified, it’s a strong version that is clear enough to evaluate fairly.
When evaluating an idea, steelman its argument, do not strawman it. Your goal is to build the strongest fair version of a viewpoint before evaluating it.
When contrasting views, steelman both sides. OVM’s true believer and true skeptic viewpoints help you shift your mind. Using these two, of the three OVM viewpoints, allows you to construct opposing cases. Just create a steelman argument for and against. Take evolution, you can research and create a steelman argument for the true believer viewpoint, the justification for evolution. Then you research and create a steelman argument for the true skeptic viewpoint, the justification against it. Steelmanning both viewpoints does not imply that both are equally justified. The purpose is to give each its strongest fair case before evaluating the epistemic support each actually earns.
Now, it’s also true that we must impose a rule that keeps this exercise sane. Claims about the material world remain accountable to empirical evidence and open to future testing. This means you must give empirical implications empirical priority. The empirical precedes the rational in the development and testing of ideas, so whenever a rational claim reaches into the material world, empirical reality gets the final say. Rational ideas may be coherent, elegant, and internally strong, but they also need to survive valid testing. If their testable empirical implications fail appropriate testing and that failure survives scrutiny, the justification fails too, and the idea falls into the irrational category. The idea needs updating; otherwise, it must stay in the irrational category as speculative or disproven. If it has not yet been adequately tested, it remains speculative; if it has failed appropriate testing, it is disproven.
For example, let’s analyze the justification for three statements:
- “A ghost knocked the book off my shelf.”
- “Cigarette smoking substantially increases the risk of lung cancer.”
- And, “It will rain tomorrow.”
These examples make the distinction between truth and justification especially visible. The key is to evaluate how warranted a belief is using critical thinking skills.
Take this poorly justified statement:
“A ghost knocked the book off my shelf.”
The observation—that the book fell—is evidence, but it does not specifically support the ghost explanation over ordinary causes like gravity, vibration, or another person. The claim may conceivably be true, but the available evidence does not justify it. The conclusion outruns the evidence.
Now let’s examine the other side of the spectrum with this strongly justified one:
“Cigarette smoking substantially increases the risk of lung cancer.”
This belief is supported by decades of research including repeated observations across populations, biological mechanisms, and multiple independent lines of evidence. Individual exceptions do not refute the general causal claim. The evidence is abundant, convergent, independently tested, and durable, so the justification is very strong.
Beliefs that have strong justification based on empirical evidence and weak beliefs based on little or no justification are easier to judge. It’s the middle that gets muddy and is harder to judge. Take this moderately justified prediction:
“It will rain tomorrow.”
If several weather models predict rain, clouds are moving in, humidity is rising, and the forecast gives a 70% chance, justification rises. There is meaningful evidence supporting the belief, so believing rain is likely is justified. But the evidence is predictive rather than conclusive, and weather models can fail. The belief deserves real confidence, but not certainty.
Those three examples give us a useful justification ladder:
- An idea has low or no justification when it outruns the evidence.
- An idea has some justification when evidence meaningfully supports a claim.
- An idea has high justification when multiple independent lines strongly support and constrain a claim.
The middle case demonstrates well why justification is graduated. “It will rain tomorrow” does not have to be either justified or unjustified in the same simple sense as a binary switch. It can be reasonably but provisionally justified.
And none of the three examples makes justification identical to truth. The ghost claim could turn out to be true despite being badly justified, and tomorrow’s well-supported weather forecast could turn out false. That separation is central to this argument.
The primary epistemic role of justification is normative: determining whether, and to what degree, a belief is warranted. On the side of the mind, this requires coherent and appropriate reasoning; in its relation to reality, it requires responsiveness to relevant evidence and counterevidence.
We do not possess absolute truth. Our access to truth is fallible. We form descriptions of reality. We test them. We refine them. We sometimes get them wrong.
Absolutes belong to the material world; beliefs belong to minds.
And that means something crucial: justified beliefs about the material world must ultimately answer to the material world.
Next up is credence.
Epistemic Credence and Basis
Credence here is used in the traditional sense. Credence is the current degree of belief a mind holds in a specific idea, not the strength of the evidence for it. The actual reasons for a mind’s credence is the basis, or credence basis. Do not confuse basis with justification. Justification is the normative quality of the evidence and reasoning supporting an idea.
To start our exploration of credence, let’s take this claim:
“The Earth orbits the Sun.”
A person might hold extremely high credence in this belief—something approaching certainty. And the basis for that certainty can range from “I saw it on TV” to a formal justification. From a gut feeling to the wealth of well-established empirical and rational knowledge. The important point is that credence names the degree of confidence, not the basis nor the justification. From a practical philosophical perspective, even here, with something that appears so certain, a mind must avoid claiming absolute certainty, because a mind’s belief remains a representation held by a fallible mind. You can still call it a fact, just be open to new evidence.
Next up, let’s talk about predictions.
“My package will arrive today.”
This naturally produces intermediate credence. Tracking may show the package is out for delivery, perhaps making someone something like 90% confident, but delays remain possible. Heck, cancellations do too. This is a good reminder that nothing forces credence into rigid belief or disbelief,
nothing forces a mind into the believer or skeptic viewpoints.
Even for ideas that you are a believer in, your mind can, and in many cases should, reasonably hold back. You can hold an idea as a fact and reserve room for reevaluation. Middle-ground examples, ideas with some justification, demonstrate why credence comes in degrees, why it must be a dial rather than a switch.
Now let’s push this idea further. Take this claim:
“A ghost knocked the book off my shelf.”
This one is useful precisely because a person could hold very high credence in it even though, as discussed in the previous section, the justification is poor. Someone who strongly believes in something is aligned with the believer OVM viewpoint for it. In this case, a believer in ghosts might be 95% convinced a ghost knocked the book off the shelf. Their basis for that credence might rely on personal experience or family tradition. By way of clarification, the basis for credence in an idea is not the same thing as the justification for it. A mind’s basis for credence does not make the justification strong and certainly does not make the claim true.
To be clear, credence is not a binary yes or no. You hold it in degrees, a continuous spectrum of confidence. And you can give it a percentage. I believe this at 99%. I believe that about 55%. A credence percentage represents the subjective strength of a cognitive state. In formal epistemology, credence is commonly modeled as subjective probability.
Finally, a mind’s actual credence is not the same as “rational credence.” They can diverge, sometimes substantially. Credence is the degree you believe something. Rational credence is the degree you should believe it. And the reasons for a credence and a rational credence are two separate things. A mind’s basis for belief, its credence basis, is distinct from epistemic justification, even though the two may overlap. This distinction is one reason this epistemological account of belief takes a process-oriented approach, and why recalibration is so important to the thinking mind.
With justification, credence, credence basis, and rational credence in place, we can now turn to the process of recalibration.
TST Recalibration: A Process
Traditionally, calibration measures whether stated probabilities match observed frequencies over repeated judgments. For example, whether credence assigned 80% confidence turn out true about 80% of the time. The normative implication is that credence should be adjusted according to the justification, although that adjustment is not always the focus of calibration itself. TST makes the adjustment explicit and recurring. Recalibration is the recurring practice of auditing actual credence against the rational credence warranted by the available justification. It identifies whether the two are reasonably aligned and whether revision should be considered.
Calibration asks:
Is your confidence level reliably matching outcomes?
The normative use of calibration says:
You ought to have well-justified credences.
In simple direct applied philosophy terms:
You ought to let evidence set the strength of your belief.
In practice, you can recalibrate anytime, especially when new data becomes known or doubt rises. This approach strives to keep confidence proportionate to the available justification.
Epistemically, as relevant evidence changes, the justification changes and justification sets rational credence. Philosophically, that should have impact on a mind’s credence and alter the basis — your confidence should adjust accordingly. From a critical thinking perspective, recalibration is required whenever relevant new evidence materially changes the available justification. Also from a critical thinking perspective, overconfidence and underconfidence are failures of recalibration: thinking errors.
Calibrate-Recalibrate Examples
You can think of the recalibration process as a test of fitness. Suppose two people have exactly the same evidence. The evidence justifies moderate confidence. One says, “I’m about 60% confident.” The other says, “I know this with absolute certainty.” The objective justification is identical, but their credence differs. Good critical thinking confirms one and asks the other to adjust.
During recalibration, you audit a mind’s credence to the justification. Take this prediction:
“It will rain tomorrow.”
Suppose the available evidence supports roughly a 70% chance of rain, and the person holds the belief with about 70% confidence. A match. The mind’s credence is well aligned with the rational credence warranted by the justification. Let’s call that the weatherperson credence. Over time, the weatherperson’s forecasts can also be tested for calibration: forecasts given at 70% should prove correct about 70% of the time.
A viewer of this weatherperson, may hear this analyses and claim 99% certainty it will rain because his gut tells him such. Or he just feels it. Let’s call this the believer credence. Good thinking dictates that an unsupported gut feeling does not, by itself, warrant overriding stronger evidence.
Next, let’s construct a valid change to justification. Same scenario, but the viewer understands that 70% is an average for the larger area, and he has data showing it rains more frequently where he is. Under that analysis he might reasonably adjust the 70% forecast upward, perhaps substantially, to say 90%. Over time, we can test it too. We can track if it rains 9 out of 10 times where he is with that prediction.
This example supports the idea that the adjusting of credence to warrant requires good critical thinking skills. To think well, you need to pay attention to reality.
Next, let’s return to this example:
“A ghost knocked the book off my shelf.”
Suppose someone is 95% convinced because they already believe strongly in ghosts, even though the available evidence provides little support for that explanation over ordinary alternatives. Their credence greatly exceeds the justification. Recalibration identifies the mismatch and calls for reducing confidence—not necessarily declaring the idea impossible. And, I’ve met people who acknowledge this nuance. They tend to say things like,
“It sure felt like a ghost knocked that book off of my shelf, and I hope ghosts to exist, but I know the evidence is not conclusive. I just like to believe.”
I actually kind of admire that honesty, but epistemically the credence is still misaligned with the available justification. It’s not good critical thinking. It’s saying, I understand how to think well, I’m not ignoring evidence so much as I’m just enjoying my journey. This is an example showing that TST Recalibration preserves agency without becoming relativistic.
All ideas about ghosts are speculative because we don’t have a reliable way to test them. So, return to this statement:
“A ghost knocked the book off my shelf.”
That claim is not testable and falls into the speculation subtype of the Irrational category of the Ideas of Ideas. If we ever find a way to reliably test for ghosts, we’ll be able to construct empirically and rationally true ideas. If they exist, we will be able to tell ghost stories. Rational stories about an empirical past. We will be able to say things like this:
“A ghost knocked that book off of the bookshelf.”
And we will be able to say empirical things like that this:
“That ghost has the ability to knock a book off the bookshelf.”
However if we can never test an idea, it remains forever as speculation. Refer to the paper titled TST Philosophy of History for more information.
For our third recalibration example, let’s return to this statement:
“My package will arrive today.”
In the morning, tracking shows “out for delivery,” reasonably supporting high credence. Later, the carrier reports a delay until tomorrow. If the person continues to hold 90% confidence that it will arrive today, his credence is overconfident. The justification changed while their credence did not. Their mind was overly stubborn despite the evidence.
This third example captures the idea that recalibration is not merely checking whether confidence was appropriate when first formed. It is an ongoing audit.
Before we wrap this up, let’s frame belief within the Idea of Ideas and demonstrate a few complete applications.
The Idea of Ideas
Our minds categorize ideas all day long. New ideas are looked at differently than something we have already evaluated. The Idea of Ideas gives that categorizing a meta structure that sits above binary logic. It raises category placement to a conscious level. The Idea of Ideas has three meta-categories. The mind uses them at a top-level to sort understanding.
Before thinking about binary true-false logic, you are asked to categorize new ideas as either a potential empirical or potential rational idea. That classification determines what kind of justification is appropriate, after which confidence can be recalibrated to the strength of the available evidence and evaluation, and the idea placed within the appropriate category: empirical, rational, or irrational.
Empirical truths describe the material world directly. Rational truths indirectly, and irrational ideas do neither.
Empirical First: Direct Contact With Reality
Whether a simulation or not, reality exists independent of us and our strongest contact with it is empirical.
Empirical ideas come through observation. Observations that can be repeated are easier to verify than one-time historical events. Our strongest empirical claims survive repeated measurement and testing. They are not guesses. They are not preferences. They are claims that can be checked. Empirical ideas are the mind’s direct perceptual representations of material reality whether sensory or nonsensory.
In other words, empirical contact is not limited to unaided human senses. Some things are sensory to us; others are nonsensory and require tools. Microorganisms existed before microscopes. Radio waves existed before radios. Distant galaxies existed before telescopes. The tools did not create those things. They extended our access to a reality that was already there.
Radio waves are as real as rocks.
This matters because reality does not wait for reason to authorize it. The structures and patterns of nature exist first. Minds encounter them, form ideas about them, and then reason about those ideas. Rational ideas are constructions of the mind built from, about, or applied to empirical ideas. And rational ideas must remain coherent and have valid empirical application. This is the foundation beneath justification.
Philosophers have long debated justification. Essentially justification answers the question:
Why are you entitled to believe that?
In epistemology, justification is what makes holding a belief permissible. That is why justification, credence, and recalibration are organized under a single ordering principle:
Empirical precedes rational.
This does not mean reason is unimportant or that every rational idea requires direct empirical evidence. It means empirical implications give empirical priority. Rational ideas may precede testing in development, but whenever they make claims about the material world, empirical reality gets the final say. Meaning, all ideas are built from and constrained by a world that exists before them. And whenever an idea reaches back into that material world, reality has priority. Finally, and very important, this also means that every rational idea must have empirical application.
Once a rational idea has earned empirical justification, that justification need not be demonstrated every time the idea is used. However, because minds are fallible, even our most established rational ideas remain open to challenge, testing, and revision.
In TST, valid empirical application means more than a possible application that has not yet been tested. A rational idea earns its truth category through empirical justification of its application to nature, not through coherence alone. As with all empirical knowledge, that justification remains provisional and open to revision.
That principle means something specific and simple:
If an idea conflicts with observable evidence, the evidence wins.
This requirement distinguishes mathematical coherence from rational truth in TST. A mathematical system can be internally coherent, and its conclusions valid within that system, without establishing anything about the material world. Such conclusions are not necessarily wrong. They simply have not earned TST’s category of rational truth until valid empirical application is established. Internal coherence is necessary, but it is not sufficient.
Ideas, such as a rational construction, theory, model, or argument, must answer to reality. Rational truths matter. Logic matters. Inference matters. But when rational claims make empirical predictions, those predictions are not optional; they must be tested. Binary truth, epistemic classification, and credence operate at different levels. Reality determines whether a proposition is true or false. The mind then classifies the idea as empirical, rational, or irrational according to its epistemic status, while credence describes how strongly a mind believes it.
Science demonstrates the power of this ordering. A pattern may be observed a million times, giving us very strong justification for expecting it to continue. But that repetition does not make the belief untouchable.
Take this claim:
The Sun has risen my whole life, therefore it will always rise.
For thousands of years, we had no reason to doubt it. Yet science has proven that the Sun will not always rise. In around five billion years, the Sun will begin depleting its hydrogen energy source. That’s now known using updated evidence that changed the rational credence warranted by the publicly available justification.
As a more concrete example, one, and just one verified observation showing that under a particular condition something different happens can, when reproducable, force us to revise the idea. In Chapter 20 of 30 Philosophers, I used the boiling point of water to illustrate how empirical knowledge matures. Humans observed for thousands of years that water boils when heated. Later, Fahrenheit gave that observation precision: water boils at 212°F. But reality pushed back. At higher altitudes, where atmospheric pressure is lower, water boils at a lower temperature. The earlier observations were not worthless; they were incomplete. The idea was refined: under standard atmospheric pressure, water boils at 100°C, or 212°F. This is the strength of empirical knowledge. Repeated observation strengthens justification, but even one well-established exception can reveal a missing condition and force the idea to become more precise.
And here is the ordering rule:
Where empirical implications exist, they are decisive.
If a belief makes contact with the material world, that contact is the final judge. This does not make reason unimportant. It places reason in service to reality — not above it.
This also means that a mind should not begin by asking reason for permission to trust repeated experience. Stable patterns in reality provide provisional justification precisely because they are repeatedly observed. This is the Principle of the Uniformity of Nature. Reason helps us organize those patterns, identify conditions, expose contradictions, and build better explanations. But reason does not stand above reality as its judge.
Descartes helps here, but only after revision. His great contribution was to dramatize doubt. His weakness was that he both tore knowledge down and rebuilt it using irrational tools. We can keep the useful part, the disciplined doubt, and then rebuild differently. In TST, as in natural philosophy more broadly, the rebuilding begins with the material world.
At this point, Hume helps with his work on impressions. Hume showed that repeated observation does not logically guarantee that a pattern must continue forever.
Hume blocks certainty about induction. He does not block calibrated reliance on it.
On Hume in chapter 27 of 30 Philosophers, I write:
“He explains that the logic behind induction is inherently circular: we expect future events to be like past events because this has generally been true in the past. This justification itself is inductive, making the whole process circular and self-referential.”
The mind can accept that limitation, but because TST begins empirically rather than with logical certainty, that limitation does not paralyze knowledge. Repeated patterns in reality provide provisional and graduated justification until reality gives us reason to revise them.
In starker terms, reality is determinate. Our access is partial. Prediction is rational extension, not empirical possession of the future. Therefore induction warrants calibrated confidence, never certainty.
Repeated observations justify increased confidence in a particular regularity, while the extension of that regularity beyond observation remains a fallible inductive inference rather than a demonstrated truth.
Knowledge does not rise from pure reason alone, and it does not need a supernatural guarantee. It begins with impressions, perception, measurement, memory, pattern, and correction. In that sense, every baby reruns Descartes’ thought experiment from birth, and every family and civilization reruns it across time. We begin with contact in a world that pushes back, and from there we sort our ideas into empirical, rational, and irrational. For more on Rational Cartesian Skepticism and Hyperbolic Doubt, see Cartesian Skepticism in chapter 22 of 30 Philosophers.
The Rational Subtypes
Rational ideas divide into two subtypes: single-layered and multi-layered. Single-layered rational ideas remain close to observation. Counting rocks, measuring distance, comparing weights, and basic arithmetic are rational because they are not the observed objects themselves; they are structured thoughts about those objects. If I see four apples and think four, the apples are empirical, while the number is rational. The bridge back to reality is short and easy to trace.
Multi-layered rational ideas are built from additional layers of rational interpretation. Ideas such as justice, beauty, love, rights, meaning, and worldview are not one step removed from observation. They are constructed from other ideas that include things like categories, judgments, comparisons, memories, cultural frameworks, and often emotion. They can still be rational, but their justification is more complex because more layers must be examined.
This distinction matters for justification. A multi-layered rational idea is not weaker merely because it is more abstract. But each additional layer creates another place where assumptions, errors, or hidden premises can enter. The more layered an idea becomes, the more carefully its structure must be traced. Rational justification therefore depends not only on whether an idea is coherent as a whole, but also on whether the supporting layers hold together.
And it’s important to note that rational coherence alone is not enough. An idea can be perfectly consistent and still be wrong about reality. A theory may be elegant, precise, even beautiful, and yet fail when it encounters the material world. When rational claims imply observable consequences, those consequences must be tested.
The distinction between single-layered and multi-layered ideas helps here too. With a single-layered idea, the path back to empirical reality is often immediate. With a multi-layered idea, the path may be longer and more difficult to trace, but the same ordering principle still applies: when any part of the rational structure makes an empirical claim, reality is allowed to push back.
Rational structure is necessary for justified belief, but it is not sufficient on its own. It ensures our ideas are coherent. Reality determines whether their empirical claims survive contact with the material world.
The third category, irrational ideas, marks the boundary of justified truth claims. Irrational does not necessarily mean meaningless or absurd. Such ideas may be speculative, imaginative, symbolic, or personally important. But when an idea lacks adequate justification, it is not warranted as established truth. The irrational has two subtypes: speculative and disproven. Speculative ideas may remain open questions; disproven ideas should be relinquished as truth claims. The category, therefore, lets us keep an idea open for investigation without treating an unverified idea as if it already qualifies epistemically as knowledge.
Applying Rational ideas
Rational ideas describe the material world indirectly, but they must have empirical application. Simple rational ideas are easy to apply. More complex, multilayered ideas may require a more careful translation into a direct claim about the material world.
Consider the rational idea:
“People fall in love.”
That idea is complex, abstract. Which people? What do you mean by love? The move is to test it as many times as needed. Rational ideas can be exercised empirically by applying it to a specific case:
“Those two people are in love.”
Now the claim directly describes the material world and can be tested. To do that well, we first need a sufficiently precise definition of love; that defines the scope of what we are testing. And with love, almost any defined scope will likely invite endless debate. And that’s okay. Once scope is defined, we then examine relevant evidence from the two people: their statements, behavior, commitments, emotional responses, history, and other observable facts. The idea of love is complicated, but the process is the same.
Rational ideas have empirical application.
Whenever you struggle to calibrate your credence in a rational idea, try converting it to one or more empirical ideas to clarify the issue. Applying rational ideas is not foolproof, especially with complex, multilayered rational ideas, but it helps. Like the Socratic method, it clarifies.
The Testing Mechanism: Discipline and Error Correction
Justified belief is not a feeling. It is a process.
This is where discipline enters.
The TST Framework provides structure for that discipline. The Five Thought Tools help us think well. They process claims carefully. The Four Mind Traps warn us about illusion. About where we distort, rush, or deceive ourselves. The Three Truth Hammers remind us that public truth systems — science, law, and journalism — exist to test claims beyond individual bias.
This matters because private judgment has limits. No individual can personally reproduce every experiment, inspect every crime scene, interview every witness, or verify every reported event. Public truth systems extend our epistemic reach. At their best, they preserve evidence, expose claims to competing scrutiny, document correction, and allow ideas to be tested by people who do not share the same interests, assumptions, or biases.
These systems are not infallible, and their authority does not make a claim true. Their strength is procedural. They create structured opportunities for reality to push back. A scientific result can fail replication. A legal claim can collapse under contrary evidence. A reported fact can be corrected when better documentation emerges. The process matters because justification grows stronger when a claim survives serious attempts to test it.
Science provides perhaps the clearest example. A pattern can succeed a million times and earn extremely strong justification, yet one verified counterexample may expose a missing condition and force revision. That is not a weakness of empirical knowledge. It is one of its greatest strengths. Repeated success strengthens justification, while reality always retains the power to push back.
Beliefs must therefore be exposed to the kinds of pressure appropriate to the ideas supporting them. Rational ideas must survive examination of their premises, coherence, and inference. When an idea makes empirical claims, those claims must also survive observation and testing. Falsifiability matters where empirical implications exist. Revisability matters throughout.
If a claim cannot be questioned, its justification cannot be seriously examined. If its reasoning fails, it must be repaired or rejected. If its empirical implications fail, the idea must answer to that evidence. Justification is strengthened not by protecting ideas from criticism, but by exposing them to disciplined attempts at correction.
Belief is not a trophy we win and display. Justification must remain open to reassessment as evidence and reasoning change.
Justification is not possession of truth; it is disciplined alignment with reality.
From Private Justification to Public Truth
Justification begins with minds, but knowledge does not remain private. Evidence can be recorded, reasoning can be shared, experiments can be repeated, testimony can be challenged, and conclusions can be revised by others. In this way, privately justified belief can enter a public process of evaluation.
While public belief is often made true or false by agreement, public truth is different. Public truth is not made true by agreement. It is the accumulated body of ideas that have become publicly available for examination and have survived sufficient empirical, rational, and institutional scrutiny to warrant broad acceptance. Consensus can strengthen confidence in that process, but consensus itself does not determine truth. Reality retains that role.
The Three Truth Hammers help make this transition possible. Science, law, and journalism allow evidence and reasoning to move beyond the limits of one mind. Their value lies not in institutional authority alone, but in their capacity to preserve justification and rational claims. They preserve evidence, expose claims to criticism, correct errors, and allow reality to push back publicly.
Degrees of Justification: Justification Is Not Binary
Justification is not all-or-nothing.
A claim is not simply “justified” or “unjustified” in some dramatic final sense. Justification comes in degrees. Evidence accumulates. Arguments strengthen or weaken. Predictions succeed or fail.
This follows naturally from a fallibilist, reality-first account. Evidence is rarely complete, reasoning can be improved, and new observations can expose conditions we previously missed. A belief may therefore be weakly justified, strongly justified, or supported so well that doubting it would require substantial contrary evidence. None of this requires pretending that justification has become certainty.
As justification strengthens, confidence should grow. As justification weakens, confidence should shrink.
Justification concerns the quality of the case for belief; rational credence concerns how strongly a mind should hold that belief in light of the case. They are related but distinct. The graduated nature of rational credence is developed separately as evidence builds.
For the critical mind, absolute confidence in one’s representations is unwarranted. The material world contains absolute truths, but human beings can only hold descriptions and approximations. We adjust them as we learn more. Even if a mind possessed an exhaustive and perfectly accurate description of reality, it could not establish with certainty that the description was exhaustive and perfectly accurate. This protects us from dogmatism: the illusion that we have arrived. It also protects us from paralysis: the fear that we can know nothing at all. That doesn’t mean you can’t fully believe in certain facts. You can have extreme confidence in things like established logical truths, while remembering that even our strongest beliefs remain representations held by fallible minds.
We do not need certainty to act. We need justification strong enough for the belief and action under consideration, with confidence proportionate to that justification.
Truth and Justification
Truth concerns whether an idea aligns with reality. Justification asks a different question. It does not determine reality and does not make an idea true. It concerns the quality of the evidence and reasoning that warrant believing the idea to be true. A true idea may be poorly justified by the person who believes it, just as a false idea may sometimes appear well justified given incomplete or misleading evidence.
That distinction is essential. Truth belongs to the relation between an idea and reality. Justification belongs to our reasons for believing that the relation holds. Truth does not become stronger or weaker as evidence changes; our justification does.
Consider the statement,
“There is microbial life on Europa.”
Reality already determines whether that statement is true or false, but we do not currently know which. The idea therefore, remains speculative and irrationally false in a logical setting. There may be circumstantial evidence worth some credence, but the process of recalibration should keep that confidence below the threshold required to identify the idea as empirically true. And with every false claim, you can usually write a true version of it. The following is a true statement as I write this:
“We do not currently know if there is microbial life on Europa.”
If future investigation verifies microbial life on Europa, reality did not change. Our classification will. The speculative idea will earn empirical status. If we ever gather enough evidence and conclude that life has evolved on Europa, the previous statement will move to empirical truth and we can confidently say:
“There is microbial life on Europa.”
And if that day comes, that will be a special day.
What TST Is Not
TST is not relativism. Truth does not change because we vote on it. Widespread agreement may provide evidence about what people believe, but agreement does not determine what is true. A proposition about the material world remains accountable to reality whether one person believes it, everyone believes it, or no one has discovered it yet.
TST is also not blind empiricism. Empirical priority does not mean empirical evidence interprets itself. Observation requires rational structure: categories, measurement, comparison, inference, and explanation. Raw observations alone do not produce knowledge. Data without rational structure becomes noise. Empirical evidence grounds the inquiry; reason organizes what the evidence means.
Nor is TST pure rationalism. Coherence is necessary for rational justification, but coherence alone cannot establish that a rational construction accurately describes the material world directly or indirectly. An argument may be internally consistent, mathematically elegant, and logically valid while beginning from false premises or producing predictions that fail. Whenever rational ideas acquire empirical implications, those implications remain answerable to reality. Rational ideas require an empirical application.
TST is not pragmatism alone either. An idea may be useful, comforting, socially stabilizing, or effective without being true. Pragmatic success can contribute to the evaluation of an idea, especially by revealing how well a model performs, but usefulness does not determine truth. What works and what is true may overlap, but they are not identical.
TST is not institutional worship. Science, law, and journalism are Truth Hammers because they provide structured processes for testing claims beyond private judgment. Their conclusions do not become true because an institution declares them. Their epistemic strength in rational credences comes from justification. Institutions are valuable precisely because they can be corrected when reality pushes back.
And TST is not skeptical paralysis. Doubt is a tool, not a permanent residence. Radical doubt can reveal weakness in a belief, but doubt alone cannot rebuild knowledge. Justification grows through structured contact with reality: observation, measurement, reasoning, testing, correction, and continued revision. The inability to obtain absolute certainty does not prevent one from holding strongly justified beliefs.
TST therefore occupies a disciplined realist position. It affirms a mind-independent world while accepting that human descriptions of that world remain mediated, partial, and revisable. Reality stands; ideas attempt to align with it. Justification measures the quality of that attempt without pretending the attempt has become reality itself.
TST is disciplined realism.
Reality first.
Reason in service to it.
Justification provisional, not absolute.
Belief warranted, not merely declared.
The Math of Belief
Truth and belief operate differently. In a true-false system, a proposition is either true or false, but a mind’s degree of belief in that proposition is quantitative — you can measure it. Formal epistemology commonly represents credence on a scale from 0 to 1, where 0 represents complete confidence in falsehood and 1 complete confidence in truth, with any fraction in between representing a degree of belief. The same value can be expressed as a percentage from 0% to 100%. A credence of 0.8 means 80% confidence that the proposition is true; it does not mean the proposition is 80% true.
This gives us a simple way to distinguish truth, justification, and belief.
Truth remains binary relative to reality. Confidence remains graded relative to evidence.
For empirical and rational ideas, TST adopts classical bivalence: true or false. Truth within a given scope is either true or false. This can be represented like this:
[Truth] ∈ {0,1}
A proposition either corresponds to reality within its scope or it does not.
Justification, however, is graduated:
0% <= [Grades of Justification] < 100%
A belief may have no justification, weak justification, moderate justification, or extremely strong justification. But because justification belongs to a fallible mind working from incomplete information, it is not exhaustive. Our minds are on one side of the Split; reality, the standard for truth, is on the other.
Credence, a mind’s actual belief, can range across the full scale:
0% <= [Actual Belief] <= 100%
A mind can completely reject an idea or believe it with absolute certainty. People do both. But this describes what minds can do, not necessarily what a critical thinker should do.
For a fallible critical thinker, 100% certainty does not exist.
0% < [Critical Thinking Belief] < 100%
Even when confidence is extremely high, a critical thinker leaves room for correction. This applies even to our strongest rational beliefs. It is hard to imagine something like a logical contradiction being evaluated incorrectly, but the rule protects revisability because the evaluating mind remains fallible.
Let’s clarify a bit. We are talking about standard Bayesian conditionalization. The idea that assigning a proposition a probability of 0 effectively prevents ordinary evidence from resurrecting it, while assigning 1 similarly locks in certainty. In that sense, absolute credence mathematically closes the door to revision.
Reality may be absolute. The mind’s confidence is not.
Mathematics does not determine truth, nor does it automatically tell us which evidence deserves trust. But it can discipline the relationships among our degrees of belief and show how those beliefs should change as new evidence arrives.
Conclusion: The Belief Principle
The position defended here is that justification is provisional and graduated. A belief is justified to the degree that the evidence and reasoning supporting it are appropriate to the kind of idea being evaluated. Empirical claims must remain accountable to the material world. Rational claims must maintain logical coherence and valid empirical application. All ideas must remain open to disciplined testing, criticism, and revision.
Given this:
- Credence is how strongly I currently believe an idea.
- Credence basis is the grounds upon which my present credence actually rests.
- Justification is the epistemic support available after appropriate evaluation.
- Rational Credence is the degree of belief that the justification warrants.
- Calibration is the process of analyzing and validating justification.
The process is this.
TST Recalibration is a practice within Idea Evaluation of the TST Framework. It’s the deliberate identifying of the degree of alignment of a credence to a rational credence. You evaluate the available justification, identifying the rational credence that justification warrants, and then consciously deciding whether to revise one’s actual credence. A mind’s choice, a choice that does not require you to adopt rational credences.
The value is in the process is deliberately deciding who you want to become. This is akin to other methods like the Socratic method. Whether you use this process or another, the goal is the same: to align belief with reality.
This produces a reality-first structure of justification. Reality provides the ground. Reason organizes and extends what we learn. Disciplined testing exposes error. Fallibilism keeps belief open to correction.
The principle of justified belief is therefore simple:
belief should remain proportionate to the quality of its justification and answerable to reality wherever empirical implications exist.
To make the practical use of this theory concrete, here is how I apply it in everyday thinking. In chapter 18 of 30 Philosophers, I wrote:
A statement or belief stands as an idea and is either empirically true, rationally true, or irrationally false in a logical setting. And here’s the kicker, discerning its category is straightforward and uncomplicated.
In the book, I was essentially saying that classifying what is true and false is pretty easy. And I meant it. In practice, this feels easy to me. When I encounter a new idea, I first ask what kind of idea it is. Is it already empirical, rational, or irrational? If so, I’m done unless their is significant new evidence. If it is disproven, the classification is easy: disproven. If it has not yet earned empirical or rational status, it remains speculative. If it is potentially true, I ask whether it describes reality directly or indirectly. For indirect ideas, I turn to the applied direct versions. I then tend to lean on science because it’s pretty good at direct descriptions. Where science doesn’t say, I think scientifically when deciding my degree of belief. I compare that warrant to my actual credence, and I recalibrate when needed. In chapter 22 of 30 Philosophers, this is the empirical pragmatist worldview lens.
Minds do not possess absolute truth. They pursue disciplined alignment with it. We seek empirical grounding where available, rational justification where necessary, and revision whenever reality pushes back. That pursuit is what it means to believe well.