CHAPTER 9

A Priority and Inference

What Is This Thing Called Knowledge?
by Duncan Pritchard

Topics Covered in This Chapter

• A Priori and Empirical Knowledge
• The Interdependence of A Priori and Empirical Knowledge
• Introspective Knowledge
• Deduction
• Induction
• Abduction (Inference to the Best Explanation)
1. A Priori and Empirical Knowledge
What Does This Mean?
In philosophy, one of the most important distinctions is between A PRIORI knowledge and
EMPIRICAL (also called A POSTERIORI) knowledge. The key question is: Did you need to go
out and investigate the world to learn this, or could you figure it out just by thinking?

A Priori Knowledge Empirical Knowledge

Gained WITHOUT investigating the world Gained BY investigating the world

You can know it from your armchair (just by You must get up and look, ask, observe
thinking)

Example: "All bachelors are unmarried men" Example: "The Tropic of Cancer is in the
northern hemisphere"

Example: 2 + 2 = 4 Example: Water boils at 100°C at sea level

Also called: armchair knowledge Also called: a posteriori knowledge

Easy Way to Remember: The Armchair Test

Armchair Test: Can you know this just by sitting in your armchair and thinking, WITHOUT
going out to check anything? If YES = A Priori. If NO = Empirical.

How We Get A Priori Knowledge

A priori knowledge comes from:
• Reflecting on the meaning of words — e.g. "bachelor" means "unmarried man", so you
already know all bachelors are unmarried without checking.
• Logical reasoning — e.g. if A implies B, and A is true, then B must be true.
• Mathematical thinking — e.g. you do not need to count actual objects to know 2 + 2 = 4.

Important Point: One Way Does Not Imply the Other

A priori propositions CAN also be known empirically (e.g. you could look up in a dictionary that
'bachelor = unmarried man'). But empirical propositions CANNOT always be known a priori. You
cannot figure out from your armchair which hemisphere the Tropic of Cancer is in — you must
look it up.
Exam Tip: The distinction applies to JUSTIFICATION too, not just knowledge. A belief is a
priori justified if the justification required no worldly investigation. A belief is empirically
justified if the investigation of the world was needed to justify it.
2. The Interdependence of A Priori and Empirical Knowledge
The Big Idea
In real life, most of our reasoning uses BOTH a priori and empirical knowledge at the same
time. They are not completely separate — they work together.

The Detective Example (Professor Plum)

Imagine a detective investigating a murder. A witness tells him Professor Plum was in the pantry
at the time of the murder. The detective then concludes Plum was NOT in the hallway. Here is
the full argument:

Step Statement Type of

Knowledge

Premise 1 Professor Plum was in the pantry at the time of the EMPIRICAL
murder (told by a witness) (heard from a
witness)

Premise 2 A person cannot be in two places at once, so if Plum A PRIORI (just

was in the pantry, he was NOT in the hallway from thinking)

Conclusio Professor Plum was NOT in the hallway — so he is EMPIRICAL

nC innocent! (because Premise
1 was empirical)

Key Insight: Even though the conclusion is EMPIRICAL knowledge, the reasoning used an
A PRIORI step (Premise 2). This shows that empirical and a priori knowledge are deeply
INTERDEPENDENT — they support each other.
3. Introspective Knowledge
What Is Introspection?
Introspection means "looking inward" — examining your own thoughts, feelings, and mental
states, rather than looking outward at the world.

Two Examples from the Book

Example 1: Should I get into this relationship?
If you are unsure about a relationship, you might ask yourself: "Do I enjoy their company? Does
being around them make me happy or anxious?" You are not investigating the outside world —
you are investigating your own inner feelings. This is introspection, and it is a priori because it
requires no worldly investigation.

Example 2: Do I smell gas?

Your partner asks if you smell a gas leak. You "look inward" at your own sensory experiences to
check. This is also introspection — but notice something important here:

A Priori Component Empirical Component

You examine your experience WITHOUT Your ORIGINAL experience (being in the
going to get more information from the room, your sense of smell) came from the
outside world real world

The act of self-examination itself is non- The content of what you are examining is
empirical empirical

Bottom Line: Not ALL introspective knowledge is a priori. When you introspect an
experience that originally came from the world (like smelling gas), the result can be empirical
knowledge. But the process of introspection itself is always a non-empirical, inward-looking
activity.
4. Deduction
What Is Deduction?
A deductive argument is one where, if the premises are TRUE, then the conclusion MUST be
true. The premises GUARANTEE the conclusion. There is no way the conclusion can be false if
all premises are true.

Key Term — VALIDITY: An argument is VALID when it is impossible for the premises to all
be true and the conclusion to be false. In other words: true premises FORCE a true
conclusion.

Key Term — SOUNDNESS: An argument is SOUND when it is (1) VALID and (2) all
premises are actually TRUE. A sound argument gives you a very good reason to accept the
conclusion.

Professor Plum Example Again

Premise 1: Plum was in the pantry. Premise 2: A person cannot be in two places at once, so
Plum was not in the hallway. Conclusion: Plum was not in the hallway.
This argument is VALID because if both premises are true, the conclusion must be true. It is
also SOUND because (we assumed) both premises are true.

Valid but NOT Sound

You can have a VALID argument that is NOT sound — if one of the premises is false. For
example: Premise 1: Plum was in the pantry (FALSE — he wasn't). Premise 2: A person cannot
be in two places at once. Conclusion: Plum was not in the hallway. This is still VALID (the logic
works), but it is not SOUND because Premise 1 is false.

Valid Argument Sound Argument

True premises FORCE a true conclusion Valid PLUS all premises are actually true

Can have false premises and still be valid Cannot have false premises

Good LOGICAL structure Good logical structure AND true premises

Example: 'All cats are fish. Whiskers is a cat. Example: 'All humans are mortal. Socrates is
Therefore Whiskers is a fish.' (Valid but not human. Therefore Socrates is mortal.' (Valid
sound — premise 1 is false) AND sound)
Why Deduction Is Useful
• Deduction EXTENDS your knowledge — if you know the premises, you can infer new
conclusions.
• Even if you only have JUSTIFIED BELIEF (not certain knowledge) in the premises, a
valid argument extends that justified belief to the conclusion.
• Deductive arguments always preserve truth — if premises are true, conclusion must be
true.
5. Induction
What Is Induction?
An inductive argument moves from specific observations to a general conclusion. Unlike
deduction, the premises do NOT guarantee the conclusion — but they make it LIKELY or
PROBABLE. This means inductive arguments are non-deductive (also called "ampliative" —
they go beyond what the premises strictly say).

The Emu Example

Premise 1: Every observed emu has been flightless.

Therefore:
Conclusion C: All emus are flightless.

This is NOT valid — it is possible (even if unlikely) that an unobserved emu somewhere is NOT
flightless. But it is still a perfectly GOOD argument as long as the premise is made more
precise:

Premise 1*: LOTS of emus have been observed over MANY years in a WIDE RANGE of
environments, and they have ALWAYS been flightless.
Therefore:
Conclusion C: All emus are flightless. (Very likely to be true!)

What Makes an Inductive Argument GOOD?

• The sample must be LARGE — a couple of emus is not enough.
• The sample must be REPRESENTATIVE — observing emus only at one specific lake
does not count.
• Varied conditions — emus observed in many different environments, over many years.

Induction vs Deduction: A Quick Comparison

Feature Deduction Induction

Does conclusion follow YES — must be true if NO — could still be false

necessarily? premises true even if premises true
Can premises be true and NO (if valid) YES (possible but unlikely if
conclusion false? good argument)

Does it preserve truth? YES always Not always — fallible

Extends justified belief? YES always YES, if argument is good

Example All humans are mortal. Lots of observed emus are

Socrates is human. flightless. Therefore all emus
Therefore Socrates is are flightless.
mortal.

Important Limitation: Induction is FALLIBLE. You might know the premise (many emus
observed, all flightless) and yet the conclusion be false (because there is one flightless emu
somewhere you have not observed). So induction can extend justified belief but does not
always extend knowledge.
6. Abduction (Inference to the Best Explanation)
What Is Abduction?
Abduction is a type of non-deductive reasoning where you observe a phenomenon and infer the
BEST EXPLANATION for it. It is often called "Inference to the Best Explanation" (IBE).
Unlike induction (which goes from MANY observations to a general rule), abduction often goes
from a SINGLE observation to the best explanation of that observation.

The Curtain Example

Premise 1: There are feet visible under the curtain in the hall.
Therefore:
Conclusion C: There is someone hiding behind the curtain.

This reasoning is perfectly sensible! But notice: the premise does NOT entail the conclusion
(maybe the feet are dismembered, or it is a joke prop). We make this inference because
"someone hiding behind the curtain" is the BEST and most likely explanation of what we see.

Sherlock Holmes and Abduction

Famous Example — Sherlock Holmes: In A Scandal in Bohemia, Holmes looks at Watson

and concludes he has been out in wet weather recently and that his maid is careless. How?
Holmes notices six parallel cuts on the inside of Watson's left shoe — as if a maid carelessly
scraped mud off the sole. Holmes infers this is the BEST explanation of what he observes,
even from a single observation. He is not basing this on hundreds of similar shoes — he is
choosing the explanation that fits best.

When Abduction Looks Like Induction

Sometimes abductive inferences can be understood as "shorthand" inductive inferences. Why
do we infer someone is behind the curtain when we see feet? Because from PAST
EXPERIENCE, we know there is a regular connection between "feet under curtain" and "person
behind curtain." When you unpack it, the abductive reasoning rests on an inductive regularity.

When Abduction Goes Beyond Induction: The Corn Circle Example

Imagine you see a CORN CIRCLE for the very first time. You have no prior experience with
corn circles. Possible explanations include: (1) freak atmospheric conditions, (2) local
pranksters, (3) Martians. Since you have NO observed regularity to draw on, you cannot rely on
normal inductive reasoning. Instead, you choose the SIMPLEST and most CONSERVATIVE
explanation — one that requires the least adjustment to your existing beliefs. Most people would
prefer atmospheric conditions over Martians.

The Regulative Principle of Abduction: We tend to prefer explanations that are: (1)
SIMPLE — fewer new assumptions, (2) CONSERVATIVE — do not force us to abandon
many existing beliefs. We choose the explanation that best fits with what we already know.

The Problem with Abduction

There is a significant challenge: Why should we think that the simplest or most conservative
explanation is MORE LIKELY TO BE TRUE? There is no logical guarantee of this. The only
answer is: because in the past, simple and conservative explanations have tended to be correct
— but this is itself an INDUCTIVE argument. So:

Key Conclusion: Abductive inference ultimately depends on induction. If induction is

legitimate, then abduction is legitimate too. But as we will see in Chapter 10, there is a
serious problem with justifying induction — which means abduction faces the same problem.

Comparing All Three Types of Inference

Feature Deduction Induction Abduction

What it does Guarantees Generalises from Infers the best

conclusion from many observations explanation of one
premises observation

Is it valid? YES — truth- NO — ampliative NO — ampliative

preserving

Key requirement True premises Large, Best explanation

representative available
sample

Famous example Professor Plum Emu example Sherlock Holmes /

(logic) feet under curtain

Risk If premises false, Conclusion might be Chosen explanation

conclusion fails false might not be true

Depends on what? Logic alone Observed Background

regularities knowledge +
induction
Chapter Summary: Everything You Need to Know
Topic Key Points to Remember

A Priori Knowledge Gained without worldly investigation; armchair knowledge; e.g.

logic, maths, meaning of words.

Empirical Knowledge Gained by investigating the world; e.g. geography, science,

testimony.

Interdependence Most reasoning uses BOTH types; the detective example shows
an empirical conclusion reached partly through a priori
reasoning.

Introspection "Looking inward" at your own mental states. Can yield a priori
OR empirical knowledge depending on the situation.

Deduction Truth-preserving inference; valid = conclusion MUST be true if

premises true; sound = valid + premises actually true.

Induction Non-deductive; generalises from a large, representative sample;

fallible but extends justified belief.

Abduction Inference to the best explanation; often from a single

observation; ultimately depends on induction; prefers simpler,
conservative explanations.
Study Questions with Model Answers
Q1: Explain the distinction between a priori and empirical knowledge.

Model Answer: A priori knowledge is knowledge gained independently of any investigation

of the world — you can know it just by thinking or reflecting, as if from an armchair. For
example, knowing that all bachelors are unmarried men requires only reflecting on word
meanings. Empirical (a posteriori) knowledge, by contrast, requires an investigation of the
world. For example, knowing which hemisphere the Tropic of Cancer is in requires looking it
up. The same distinction applies to justification: a priori justification needs no worldly
investigation; empirical justification does.

Q2: What is introspection? Is it always a priori?

Model Answer: Introspection is the process of examining your own psychological states —
your thoughts, feelings, and experiences — rather than investigating the external world. It is
often a priori because it requires no worldly investigation. However, not all introspective
knowledge is a priori. When you introspect an experience that itself came from the world
(e.g. examining whether you can smell gas in a room), the resulting knowledge can be
empirical, because the original experience was gained through interaction with the world.

Q3: What is a deductive argument? What is validity? What is soundness?

Model Answer: A deductive argument is one where the premises entail the conclusion — if
the premises are true, the conclusion MUST be true. Validity means it is impossible for all
premises to be true and the conclusion false. Soundness means the argument is valid AND
all premises are actually true. Example: (1) All humans are mortal. (2) Socrates is a human.
Therefore: Socrates is mortal. This is valid (logic holds) and sound (both premises are true).

Q4: What is an inductive argument? How does it differ from deduction?

Model Answer: An inductive argument moves from observations of a large and

representative sample to a general conclusion. Unlike deduction, the premises do NOT entail
the conclusion — it remains possible (though unlikely, given a good sample) that the
conclusion is false even when premises are true. For example: many emus observed across
many environments are all flightless, so probably all emus are flightless. This is a good
inductive argument. It differs from deduction because the conclusion goes beyond what the
premises strictly guarantee. However, good inductive arguments still extend justified belief.
Q5: What is abduction? How does it relate to induction?

Model Answer: Abduction (inference to the best explanation) is a form of non-deductive

reasoning where we infer the best explanation for an observed phenomenon, often from a
single observation. Example: seeing feet under a curtain, we conclude someone is hiding
there — because that is the best explanation. Abduction differs from normal induction in that
it does not always rely on a large observed sample. However, abduction ultimately
DEPENDS on induction: we prefer simple and conservative explanations because,
inductively, such explanations have worked better in the past. So if induction is unjustified,
abduction faces the same problem.
Key Terms Glossary
Term Definition

A priori knowledge Knowledge gained independently of worldly investigation

(armchair knowledge)

A posteriori / Empirical Knowledge gained through worldly investigation

knowledge

Introspection Examining one's own psychological/mental states

Introspective knowledge Knowledge gained by introspection

Deduction Reasoning where true premises guarantee a true conclusion

Valid argument An argument where true premises necessarily produce a true

conclusion

Sound argument A valid argument whose premises are all actually true

Induction Reasoning from a large, representative sample to a general

conclusion

Abduction Inference to the best explanation; reasoning from an

observation to its most likely explanation

Inference to the best Another name for abduction; choosing the hypothesis that best
explanation explains observed facts

Ampliative reasoning Reasoning that goes beyond what is strictly entailed by the
premises (induction and abduction are ampliative)

Good luck in your exam! You've got this.