# Inferential and Non-Inferential Knowledge

<!-- type: concept | created: 2026-11-19 | updated: 2026-11-19 -->

![Marble bust of Aristotle, Roman copy after Lysippos](/codex/assets/aristotle-bust-altemps-after-lysippos.jpg)

## Intro

You walk outside in the morning. The grass is soaked, the sidewalk is dark, and water is dripping off the roof. You did not see any rain fall, but you say, "It must have rained last night."

You just made an **inference**. You took things you could see and reasoned to something you could not see. Knowledge you get this way is called **inferential knowledge**: you reach it by reasoning from other things you already believe.

Now stub your toe on the doorframe. You know you are in pain. You did not reason your way there. You did not notice evidence and then conclude, "Given these facts, I am probably hurting." You just know it, directly. That is **non-inferential knowledge**: knowledge you have without working it out from anything else.

The two words are easy to mix up:

- An **inference** is the conclusion you draw. "It rained" is the inference.
- **Inferential** describes how you got there. Your belief that it rained is inferential because you reasoned to it.

Most of what you know is a mix of both. You see a tree (direct). You remember breakfast (direct). You figure out that the tree will lose its leaves in the fall (inferred). Reasoning has to start somewhere, and it starts from the things you know directly.

This matters a great deal for believing in God. Some arguments, like the [Kalam Cosmological Argument](/codex/kalam-cosmological-argument/), reason step by step toward God. That is the inferential route. Many Christian philosophers also argue that a person can know God directly, the way you know your own mind or the world in front of you, without first needing an argument. That is the non-inferential route. A Christian can use both.

**The 30-second version:**

> Inferential knowledge comes from reasoning: wet ground, so it probably rained. Non-inferential knowledge comes directly: you know you are in pain without reasoning to it. Every chain of reasoning has to start from something known directly, or it never starts at all. Arguments like the Kalam give an inferential route to God. Reformed epistemology says belief in God can also be known directly, the way you know the past through memory. Being inferred does not make a belief true. A conclusion is only as good as its starting points and the reasoning that links them.

## In full

A belief is **inferentially justified** when its justification depends on its being based on other beliefs, evidence or premises from which it is reasoned. A belief is **non-inferentially justified** (or *immediately* justified) when it has some positive epistemic standing that does not derive from its being inferred from other beliefs. The distinction concerns the *source* of justification, not the content of the belief or how confident the subject feels.

The noun and the adjective divide the work. An *inference* is either the act of drawing a conclusion or the conclusion drawn. *Inferential* qualifies a belief, a justification, or a piece of knowledge as owing its standing to such an act. "Inferential knowledge" therefore names knowledge whose warrant passes through reasoning.

The distinction is the hinge of the **regress problem**. If every justified belief were justified only by inference from another justified belief, justification would require an infinite chain, a circle, or an arbitrary stopping point, which is the dilemma formalized as [Agrippa's Trilemma](/codex/agrippas-trilemma/). [Foundationalism](/codex/foundationalism/) answers by holding that some beliefs are non-inferentially justified and that every inferentially justified belief traces back to them. [Coherentism](/codex/coherentism/) denies that any belief is basic and locates justification in the mutual support of a whole system. Infinitism accepts an endless chain of reasons.

Candidate classes of non-inferential belief include:

1. **Introspective beliefs** about one's own present mental states (pain, thought, mood).
2. **Perceptual beliefs** about the immediate environment ("there is a tree").
3. **Memory beliefs** about one's own past ("I had breakfast this morning").
4. **Self-evident truths** grasped on understanding them (2 + 2 = 4; nothing can be red all over and green all over at once).
5. **Testimonial beliefs**, on some accounts, though whether testimony is basic or inferred is itself disputed.

Classical foundationalism, associated with [Descartes](/codex/rene-descartes/) and Locke, restricted the foundations to what is self-evident, incorrigible, or evident to the senses. [Alvin Plantinga](/codex/alvin-plantinga/) argued that this criterion fails its own test, since the criterion is neither self-evident, nor incorrigible, nor evident to the senses, nor inferred from beliefs that are. He proposed a broader, modest foundationalism in which belief in God can be among the properly basic beliefs. That is the program of [Reformed Epistemology](/codex/reformed-epistemology/).

## Three kinds of inference

Inference comes in three main forms. Each links premises to a conclusion with a different kind of strength.

1. **[Deduction](/codex/deductive-reasoning/).** If the premises are true and the form is valid, the conclusion cannot be false.
   - The textbook example: *All men are mortal. Socrates is a man. Therefore Socrates is mortal.*
   - Strength is all or nothing. A valid deduction with true premises is sound, and a sound argument guarantees its conclusion.
2. **[Induction](/codex/inductive-reasoning/).** The premises make the conclusion probable, usually by generalizing from observed cases.
   - Example: every swan observed so far in this region is white, so the next one will probably be white.
   - Strength comes in degrees, and new evidence can overturn the conclusion. See the [Problem of Induction](/codex/problem-of-induction/).
3. **[Abduction](/codex/abductive-reasoning/)**, or inference to the best explanation. The conclusion is the hypothesis that best explains the evidence.
   - Example: the ground is wet, the roof is dripping, the car is beaded with water, so it probably rained last night.
   - Strength depends on how well the explanation fits compared with its rivals (a sprinkler, a burst pipe, a heavy dew).

The wet-ground case shows why the type matters. Written as a deduction ("If it rained, the ground is wet; the ground is wet; so it rained"), it commits [Affirming the Consequent](/codex/affirming-the-consequent/), because a sprinkler would also wet the ground. Written as an abduction, it is a reasonable inference that stays open to correction.

## Inferred does not mean true

Being inferential is a fact about how a belief was reached. It says nothing by itself about whether the belief is correct.

- **Garbage in, garbage out.** A valid inference from a false premise can deliver a false conclusion. *All birds can fly; penguins are birds; so penguins can fly* is valid and false.
- **Form matters.** True premises joined by an invalid form do not guarantee the conclusion. That is what a formal fallacy is.
- **Probable is not certain.** Inductive and abductive conclusions inherit only the support the evidence gives them.
- **Non-inferential beliefs can also be wrong.** Perception is fooled by illusions; memory misfires. Direct justification is usually understood as defeasible: it holds unless something overrides it. See [Epistemic Fallibilism](/codex/epistemic-fallibilism/).

So the useful question about any belief is twofold: where does its justification come from, and is that source in good working order?

## Apologetic use

The distinction gives a Christian two routes to rational belief in God and a way to keep each one honest.

### The inferential route: natural theology

The [Kalam Cosmological Argument](/codex/kalam-cosmological-argument/) is a clear inferential case.

1. Whatever begins to exist has a cause.
2. The universe began to exist.
3. Therefore the universe has a cause.

The conclusion is reached by reasoning, so it is only as strong as its premises. It also reaches only so far. "The universe has a cause" is not yet "the God of Abraham, Isaac and Jacob." [William Lane Craig](/codex/william-lane-craig/) adds a second stage of argument about what such a cause must be like (timeless, spaceless, immaterial, powerful, personal), and identifying that cause with the Christian God requires further evidence, chiefly the historical case for the resurrection. An apologist who knows the argument is inferential will present it as a link in a cumulative case and will not overclaim what the first syllogism delivers.

### The non-inferential route: Reformed epistemology

[Alvin Plantinga](/codex/alvin-plantinga/) argues that belief in God can be **properly basic**: rational to hold without being inferred from other beliefs, just as beliefs such as "I see a tree" or "I had breakfast this morning" are rational without argument. On his account, in *Warranted Christian Belief* (2000), the belief is produced by a cognitive faculty working properly in the environment it was designed for.

Plantinga draws on John Calvin's doctrine of the [*sensus divinitatis*](/codex/sensus-divinitatis/) (*Institutes of the Christian Religion* I.3), a natural sense of God that creation, conscience and the events of life set off in us. Paul's words fit the same picture:

> For the invisible things of him since the creation of the world are clearly seen, being perceived through the things that are made, even his everlasting power and divinity; that they may be without excuse: ([Romans 1:20](/codex/romans-1-20/), ASV)

The text speaks of God's power being *perceived* through creation. On the Reformed reading, the person who looks at the night sky and believes God made it is responding to an occasion that triggers belief, and need not be running a hidden argument. See [Romans 1.18-21](/codex/romans-1-18-21/) and [Innate Knowledge of God](/codex/innate-knowledge-of-god/).

Properly basic does not mean beyond challenge. A basic belief can face defeaters, such as the problem of evil, and the believer may need to answer them. Basicality settles where the belief's justification comes from, and leaves open whether objections to it succeed.

### Evidentialism

[Evidentialism](/codex/evidentialism/) holds that a belief is justified only to the extent that it fits the believer's evidence. In W. K. Clifford's strong version, it is "wrong always, everywhere, and for anyone, to believe anything upon insufficient evidence" (*The Ethics of Belief*, 1877). Read narrowly, as a demand that every belief be inferred from evidence, the principle meets the regress at once: the evidence itself must be known somehow, and it cannot all be known by further inference.

A Christian can grant much of what evidentialism wants. Evidence matters, and the codex is full of it. The point is that "evidence" includes experiences and direct awarenesses as well as arguments, and on that broader reading the theist who believes non-inferentially is not believing without evidence.

### Cross-examination

When a skeptic says, "Unless you can prove God exists, belief in God is irrational," these questions locate the real disagreement:

1. "Do you believe other people have minds? What argument did you infer that from?"
2. "Do you believe the world existed five minutes ago? Can you prove it without relying on memory?"
3. "Is the rule *only believe what you can prove* something you proved? From what premises?"
4. "If some beliefs are rational without argument, what makes belief in God the one that cannot be?"
5. "If I give you an argument, will you judge it by its premises and its form, or by whether you like the conclusion?"

The first three show that the skeptic already holds non-inferential beliefs and must. The fourth moves the burden to the real question: why belief in God should be excluded from the class of basic beliefs. The fifth sets up the inferential route on fair terms.

<!-- SCHOLARSHIP:START -->

## Scholarship

Published academic works this page draws on, each matched to a publisher, index or library catalog record.

- Alvin Plantinga, *Warranted Christian Belief* (Oxford University Press, 2000). [doi:10.1093/0195131932.001.0001](https://doi.org/10.1093/0195131932.001.0001)

<!-- SCHOLARSHIP:END -->

## See also

- [Epistemology](/codex/epistemology/), the parent field and its three master questions
- [Foundationalism](/codex/foundationalism/), the structure that makes non-inferential beliefs the base of knowledge
- [Coherentism](/codex/coherentism/), the main rival structure
- [Agrippa's Trilemma](/codex/agrippas-trilemma/), the regress the distinction is designed to stop
- [Reformed Epistemology](/codex/reformed-epistemology/), the case that belief in God can be properly basic
- [Sensus Divinitatis](/codex/sensus-divinitatis/), Calvin's natural sense of God
- [Innate Knowledge of God](/codex/innate-knowledge-of-god/), the rich hub on natural awareness of God
- [Evidentialism](/codex/evidentialism/), the view that justification is fit with evidence
- [Kalam Cosmological Argument](/codex/kalam-cosmological-argument/), a model inferential argument for God
- [Deductive Reasoning](/codex/deductive-reasoning/), [Inductive Reasoning](/codex/inductive-reasoning/), [Abductive Reasoning](/codex/abductive-reasoning/), the three forms of inference
- [Affirming the Consequent](/codex/affirming-the-consequent/), the fallacy that turns a good abduction into a bad deduction
- [Logic](/codex/logic/), the study of valid inference
- [Problem of Induction](/codex/problem-of-induction/), why inductive conclusions stay open
- [Epistemic Fallibilism](/codex/epistemic-fallibilism/), knowledge without infallible grounds
- [Subjective Experience vs Subjective Reality](/codex/subjective-experience-vs-subjective-reality/), what direct awareness of an experience does and does not establish

<!-- COMMON-QUESTIONS:START -->

<div data-pagefind-weight="5">

## Common questions this page answers

**Q: What does inferential mean?**

Inferential describes a belief or piece of knowledge you reach by reasoning from other beliefs, evidence or premises. If you see wet ground and conclude it rained, your belief that it rained is inferential. The conclusion you drew is the inference.

**Q: What is the difference between an inference and inferential knowledge?**

An inference is the conclusion drawn, or the act of drawing it. Inferential knowledge is knowledge that owes its justification to such reasoning. "It rained" is the inference; your knowing that it rained, because you reasoned from the wet ground, is inferential knowledge.

**Q: What is non-inferential knowledge?**

It is knowledge that is justified directly, without being reasoned out from other beliefs. Standard examples are awareness of your own pain, perceptual beliefs such as "there is a tree in front of me," memory beliefs such as "I had breakfast this morning," and self-evident truths such as 2 + 2 = 4.

**Q: Why does knowledge need non-inferential beliefs at all?**

Because inference needs starting points. If every belief were justified only by another belief, justification would run on forever, go in a circle, or stop arbitrarily. Foundationalists hold that the chain ends in beliefs justified directly, and that all inferred knowledge rests on them.

**Q: Does a valid inference guarantee a true conclusion?**

Only if the premises are true. A valid deductive form preserves truth, so true premises guarantee a true conclusion, but false premises can yield a false one. Inductive and abductive inferences make their conclusions probable to the degree the evidence supports them, and new evidence can overturn them.

**Q: What are the main types of inference?**

Deduction, where true premises in a valid form guarantee the conclusion; induction, where observed cases make a generalization probable; and abduction, or inference to the best explanation, where the conclusion is the hypothesis that best explains the evidence.

**Q: Is belief in God inferential or non-inferential?**

It can be either. Natural theology, such as the Kalam cosmological argument, reasons to God inferentially. Reformed epistemology, developed by Alvin Plantinga, argues that belief in God can also be properly basic, held rationally without inference, much as perceptual and memory beliefs are. Many Christians hold it in both ways at once.

**Q: Is believing in God without an argument irrational?**

Not by any standard the skeptic applies to his own beliefs. Everyone holds some beliefs without argument, including belief in other minds and in the reality of the past. Plantinga argues that belief in God can belong to that class when produced by properly functioning faculties, while still needing to answer any defeaters raised against it.

</div>

<!-- COMMON-QUESTIONS:END -->
