Ockham’s razor
from 1320 The phrase is seventeenth-century; the practice is older
Where two explanations account for the same evidence equally well, prefer the one that assumes less.
- Active
- from 1320
- Where
- Oxford
- Also
- novacula Occami
- Status
- In use
- Tradition
- Devices and constraints
What it is
The razor is not an argument and proves nothing. It is a rule for choosing between explanations that the evidence cannot separate: take the one that posits fewer entities, fewer kinds of thing, or fewer unsupported assumptions.
It is a tie-breaker. Where the evidence does separate two explanations, the razor has no work to do and no authority.
How it works
Establish that two or more explanations genuinely account for the same evidence. Count what each one has to assume that the other does not. Prefer the leaner — provisionally, and only until evidence arrives that tells them apart.
How it is applied
- Two explanations account for the same evidence
- One posits an entity or assumption the other does not
- Nothing in the evidence requires that addition
- Prefer the leaner explanation, provisionally
The device applied
Proposition The missing biscuits do not require a poltergeist.
- Two explanations account for the empty tin: someone in the house ate them, or a poltergeist took them.
- Both are consistent with everything observed — the tin is empty either way.
- The first assumes only people already known to be in the house.
- The second assumes an entire additional category of being, for which nothing else in the evidence calls.
Therefore Prefer the first explanation, while noting that this is a preference and not a proof — a poltergeist has not been ruled out, only set aside.
This shows a device rather than an inference. The steps do not derive a new truth; they apply a rule of selection to two candidates the evidence had already left standing.
What it is good for
Stopping a theory from growing a new entity every time it meets an awkward fact. It is at its best in the sciences, where it functions as a brake on unfalsifiable additions.
Where it fails
It is routinely used as though it settled matters it cannot settle. Where the evidence does distinguish two explanations, invoking the razor is a way of avoiding the evidence, and the simpler account is often simply the wrong one — the world is under no obligation to be economical.
It also has no agreed metric. Simpler in what respect: fewer entities, fewer kinds, shorter statement, fewer free parameters? Two people can apply the razor honestly to the same pair of theories and cut in opposite directions.
Common misconceptions
That Ockham said it. The formulations usually quoted are seventeenth-century and later; what he wrote was that a plurality should not be posited without necessity, which is a good deal more careful.
Also that it says the simplest explanation is *true*. It says only that it is the one to prefer while the evidence is silent.
Where it came from
Associated with William of Ockham, who used the principle constantly in his campaign against superfluous metaphysical entities, though the idea is present in Aristotle and in Ockham's medieval predecessors.
What became of it
Survived into modern science as a working norm and into statistics as a family of formal penalties for model complexity.
Lasting influence
Visible in every modern account of theory choice, and made mathematically explicit in information criteria and minimum description length.
Why it still matters
It is the most widely invoked principle in this section and the most widely misused.
Put another way
A tie-breaker in a race, not a way of running it.
Texts
William of Ockham, Summa Logicae
Not where the razor is stated as a slogan, but where it is used relentlessly against superfluous entities.
Where to start
Read what Ockham actually wrote before reading anything about the razor. The gap between the two is the whole lesson.
The surprising bit
Ockham used it to defend theology, not to attack it — his target was the swarm of intermediate entities in scholastic metaphysics, not God.
In their own words
Plurality is never to be posited without necessity
Who held it
Duns Scotus
Connections
- Influenced
- Abduction
- Parallel to
- Sufficient reason
The verdict
Indispensable when the evidence has genuinely run out, and an evasion whenever it has not.