Grain of Salt

Valid vs sound: two different tests of an argument

Valid vs sound arguments

Struck diagram on assay stock: Valid vs sound arguments

A valid argument is one whose form guarantees that the conclusion is true if the premises are; a sound argument is a valid argument whose premises are also actually true. Validity is a test of the wiring. Soundness is validity plus a separate test of the facts. Every sound argument is valid, and most valid arguments are not sound, which is why calling an argument valid is a much smaller compliment than it sounds.

The two words are not interchangeable, and in ordinary speech "that's a valid point" means something closer to "that's a true point", which is the opposite of the technical sense. In logic and philosophy, validity says nothing at all about truth. It says only that the premises and the conclusion are connected in a way that transmits truth if there is any truth to transmit.

Valid vs sound argument: the four combinations

Valid and sound divide arguments into 4 cases, and seeing all 4 at once is the fastest way to fix the distinction.

FormPremisesNameCan the conclusion be false?
ValidAll trueSoundNo
ValidAt least one falseUnsoundYes
InvalidAll trueUnsoundYes
InvalidAt least one falseUnsoundYes

Only the first row protects you. The other 3 are all called unsound, which is why "unsound" is a weak criticism on its own: it tells the arguer that something failed without saying whether the problem was the logic or the facts. A useful objection always names which.

Validity is a property of form alone

Validity is a property of the shape of an argument, with the content stripped out. 2 examples make this concrete, and both are deliberately uncomfortable.

A valid argument with a false conclusion. Every bicycle in the club's rack has a red frame. The bicycle at the end of the rack belongs to the club. Therefore the bicycle at the end of the rack has a red frame. The reasoning is faultless. If the first premise is false, and club racks are rarely uniform, the conclusion is false too, and the argument is valid the whole time. Validity survived a false conclusion because validity was never about the conclusion.

This is the fact that makes deductive reasoning dangerous in the hands of someone who has checked only the form. A valid argument transmits whatever it is given, including error, at full confidence.

An invalid argument with a true conclusion. The car park fills up on match days. The car park is full today. Therefore today is a match day. The conclusion may well be true, and the argument still gives you no reason to believe it, because the premises leave room for a full car park on a day with no match at all. This is the form called affirming the consequent, and its persuasiveness comes entirely from the fact that its conclusion is often correct. A true conclusion is not evidence of good reasoning.

Soundness: where the argument meets the world

Soundness adds the requirement that every premise is true, and that is where the real work is. Checking validity is an armchair operation: you can do it with the subject matter blanked out, and philosophy departments do exactly that by replacing terms with letters. Checking premises is not an armchair operation. It requires going and finding out.

Take the bicycle argument again. Establishing validity took 10 seconds. Establishing the first premise means inspecting every bicycle in the rack, and the moment you do that you have left logic and entered evidence gathering, where the questions are how the observations were made, by whom, and whether the record you are relying on is the observation itself or somebody's summary of it. Primary vs secondary sources is the page on that distinction, and it is the reason soundness disputes last longer than validity disputes: validity has a decision procedure and premise truth does not.

There is a useful asymmetry here. To show an argument invalid you produce one imagined situation where the premises hold and the conclusion fails. To show it unsound you only need to falsify one premise, and you do not have to say which conclusion is right instead. Neither move obliges you to prove the opposite claim.

Is inductive reasoning valid?

No. Inductive reasoning is never valid, and that is not a criticism of it. Validity is the standard for deductive arguments only, and asking whether an inductive argument is valid is a category error, like asking how heavy an hour is. By construction an inductive conclusion says more than its premises, so the premises can all be true and the conclusion false without any defect in the reasoning.

Induction has its own pair of terms doing the same two jobs. Strong is the counterpart of valid: the premises make the conclusion probable. Cogent is the counterpart of sound: the argument is strong and its premises are true as well. Deductive vs inductive reasoning covers the full contrast, including why deduction's stronger guarantee buys nothing new about the world.

Is circular reasoning valid?

Yes, and this is the sharpest illustration of what validity does not give you. A circular argument assumes its conclusion among its premises, so the premises cannot possibly be true while the conclusion is false. That is the definition of validity, so circular arguments pass the validity test perfectly.

The timetable is accurate, because the timetable says so, and what the timetable says is accurate. Valid. Also worthless, because the only way to accept the premises is to have already accepted the conclusion, so the argument moves nobody who was not already persuaded. Philosophy calls this petitio principii, begging the question, and classifies it as a fallacy of dialectic rather than a fallacy of logic: the defect is that the argument does no persuasive work, not that its wiring is faulty.

The lesson generalizes past circularity. Validity is a floor, not a ceiling. An argument can clear it and still be uninformative, question begging, or built on premises nobody should grant. Logical reasoning sets out the further conditions, relevance, acceptability and sufficiency, that a valid argument still has to meet before it earns anything.

The two tests, in the order to run them

Run the form test first, because it is cheap and it is decisive:

Can you describe a situation, however unlikely, in which every premise is true and the conclusion is false?

If you can describe one, the argument is invalid and you are finished; you do not need to argue about the facts at all. If you cannot, the argument is valid, and only then is the second question worth asking:

Which premise would I have to check to find out whether this is actually true, and how would I check it?

Most disagreements that feel unresolvable are stuck because the two sides are running different tests on the same argument, one attacking the form while the other defends the facts. Naming which test is in play usually shortens the argument by half. Critical thinking exercises has items where the two verdicts come apart, with the reasoning shown for each.

Where to go next

Proudly powered by WordPress | Theme: Amber Blog by Crimson Themes.