Methodic doubt
from 1641 The Meditations of 1641; the procedure is sketched four years earlier in the Discourse on Method
Set aside every belief that could conceivably be false, not because it is false but to find out what is left standing when you cannot.
- Active
- from 1641
- Where
- The Dutch Republic
- Also
- doute méthodique
- Status
- In use
- Tradition
- Classical procedures
What it is
Methodic doubt is demolition with a purpose. Rather than examining beliefs one at a time, which would take a lifetime, Descartes attacks them by source: if a whole class of beliefs rests on something that can deceive, the whole class goes at once. The senses have deceived him before, so everything the senses supply is set aside.
The doubt is provisional and deliberate. Nothing is being denied. Each belief is suspended for exactly as long as it takes to see whether anything survives the suspension, and whatever survives is the foundation the rest can be rebuilt on.
How it works
Work by source rather than by item. Ask of a belief not whether it is false but what it rests on, and whether that support could fail. If it could, set aside everything resting on it.
The grounds escalate. Ordinary sensory error disposes of distant perceptions; the dream argument disposes of the immediate ones; a deceiving demon disposes even of arithmetic. The escalation is a search for the point where doubt cannot go further, and that point is the result — what survives is the doubting itself.
The moves, in order
- Take a belief and identify what it rests on, not whether it seems true
- Ask whether that support could fail, however unlikely the failure
- If it could, set aside every belief resting on it, provisionally and completely
- Escalate the grounds of doubt until nothing further can be set aside
- Take what remains standing as the foundation, and rebuild only what it will carry
The procedure
Proposition The stocktake is out by eleven units and the records cannot all be right.
- The count on the floor is eleven fewer than the system says. Four people insist their own figures are correct, and each of them is reasoning from the others' figures being correct too.
- Apply the doubt by source rather than by item. Every number that was hand-entered from a paper docket rests on somebody's transcription; transcription can fail; so all of it goes, provisionally, including the numbers that look obviously right.
- Every number generated by a scan at the door rests on the scanner having fired once per unit. A scanner can double-read. So that goes too.
- What is left standing is the sealed-carton count, because a sealed carton was never opened, never transcribed and never scanned individually — its contents are fixed by the supplier's pack and confirmed by weight.
- Rebuilt from the sealed cartons outward, the discrepancy resolves to a single pallet scanned twice on a wet morning, and the eleven units were never missing.
Therefore The error was found by refusing to trust the figures that were almost certainly correct, because "almost certainly" was exactly the standard that had let it hide for a week.
This is methodic doubt rather than ordinary checking because nothing was examined for plausibility. The sealed-carton count was not chosen for looking right; it was what remained after everything with a conceivable failure mode had been set aside.
What it is good for
Finding foundations, and finding faults that survive inspection precisely because they look reasonable. It is at its best where a system of beliefs is mutually supporting and the support is circular — a case where checking each item against the others will never reveal anything.
Where it fails
The doubt is not real, and Peirce said so bluntly: a doubt you have talked yourself into is make-believe, and make-believe doubt cannot do the work of the genuine article, which arrives uninvited and stops you acting. On this objection the whole procedure is a performance whose conclusion was chosen in advance.
What survives is also less than advertised. Lichtenberg's remark is that the most Descartes is entitled to is that thinking is going on, not that an I is doing it — the self arrives in the conclusion because it was smuggled into the grammar of the premise.
And rebuilding is where it stalls. Descartes gets from the thinking self back to the external world only through an argument for a non-deceiving God, and almost nobody who accepts the demolition has accepted the reconstruction.
Common misconceptions
That Descartes was a sceptic. The doubt is a weapon aimed at scepticism: it is pushed further than any sceptic had pushed it in order to find the one thing that cannot be moved, and the entire second half of the Meditations is construction rather than demolition.
Where it came from
Descartes, in the Meditations on First Philosophy of 1641, though the strategy has older instances. Augustine had answered the Academic sceptics with a version of the same fixed point, and al-Ghazali, five centuries before Descartes, describes suspending every belief in a crisis of doubt that ended only when certainty was restored from outside argument entirely.
What became of it
Rebuilt rather than retired. Husserl took the demolition and dropped the theology, calling his own version a return to the Cartesian starting point; the analytic tradition kept the demand for foundations and spent a century arguing about whether anything can serve as one.
Lasting influence
It set the shape of modern epistemology, which for three hundred years took the central question to be what a solitary mind can know from the inside — a framing that is Descartes's doing rather than philosophy's natural starting point.
Why it still matters
It is the correct procedure whenever a set of claims support each other and you suspect the whole set. Auditors, air-crash investigators and debuggers all use it, usually without the name.
Put another way
Condemning every beam in a house to find the one the house is actually standing on.
Texts
René Descartes, Meditations on First Philosophy
The first two Meditations carry the whole procedure. The remaining four are the reconstruction, and they are the part almost nobody finds convincing.
Where to start
Read Meditation One. It is nine pages, it needs no prior philosophy, and the demolition is complete by the end of it.
The surprising bit
Descartes thought the doubt was dangerous enough to need a warning. He advises readers not to attempt it except once in a lifetime, with time set aside and no practical business pending.
In their own words
I think, therefore I am
Who held it
Meditations One and Two
Connections
- Relies on
- The thought experiment
- Set against
- Pyrrhonian suspension
- Influenced
- The epoche, Conceptual analysis
- Parallel to
- Pyrrhonian suspension
The verdict
The most productive act of destruction in philosophy, and it has never been clear that anything was rebuilt.