Evaluating an Argument
Chapter Six
Syllabus topic 1.2, "How to identify, evaluate, interpret and construct argument."
Pages 29 to 32 of 334
In one line
Evaluating an argument means asking two separate questions: are the premises acceptable, and does the conclusion follow from them?
In the wording a student can write in an examination: an argument is evaluated by testing its premises for truth and its inference for validity or strength, the two enquiries being wholly independent, so that an argument may fail on either or on both.
Why the two questions must be kept apart
This is the point at which most students go wrong, and the error is worth naming before anything else. Presented with an argument they disagree with, they attack the conclusion. Presented with one they agree with, they accept it. Neither is evaluation.
The two questions are independent in a strict sense. Whether the premises are true is a question about the world, answered by evidence, by science or by the law of the land, and logic has nothing to contribute to it. Whether the conclusion follows is a question about the argument, answered by logic alone, and it can be answered without knowing whether any premise is true.
An argument can therefore fail in two entirely different ways, and the two failures call for entirely different responses. If the premises are false, produce evidence. If the conclusion does not follow, no amount of evidence will help.
Evaluating a deductive argument
A deductive argument claims that its conclusion follows necessarily. It is judged by two words.
Valid means the conclusion follows: if the premises were true, the conclusion could not be false. Validity is a property of the form and says nothing about truth.
Sound means valid and having all true premises. A sound argument's conclusion must be true, and that is the whole point of the word.
So the evaluation of a deductive argument has exactly two steps. Ask whether it is valid. If it is not, stop; the argument establishes nothing and the truth of the premises is irrelevant. If it is valid, ask whether the premises are true. If they are, the conclusion is established.
The four possible combinations are worth having in front of you.
| Premises | Valid? | Verdict | Conclusion |
|---|---|---|---|
| True | Yes | Sound | Must be true |
| False | Yes | Unsound | May be true or false |
| True | No | Unsound | May be true or false |
| False | No | Unsound | May be true or false |
Only the first line establishes anything. Notice what the other three share: the argument gives you no information about the conclusion at all. It does not show the conclusion is false. A bad argument for a true proposition is still a bad argument, and the proposition may be proved by some other argument entirely.
Evaluating an inductive argument
An inductive argument claims only that its conclusion follows with some probability. "Valid" and "invalid" do not apply to it, and using them is a marked error.
Evaluating an Argument
Strong means that if the premises were true, the conclusion would probably be true. Strength is a matter of degree: an inductive argument can be very strong, moderately strong or weak, where validity admits of no degrees at all.
Cogent is the inductive counterpart of sound: strong, with true premises, and with no relevant evidence left out. That last requirement has no deductive counterpart and it is where most inductive arguments fail.
An example makes the difference plain. "Ninety-eight of the hundred consignments delivered under this contract were defective, so the ninety-ninth was probably defective too" is strong. Add the information that the ninety-ninth came from a different factory, and the argument is no longer cogent, though nothing in it has become false. Suppressing relevant evidence destroys an inductive argument in a way it cannot destroy a deductive one.
Interpreting an argument: the principle of charity
Before an argument can be evaluated it has to be interpreted, because arguments in the world are compressed, ambiguous and enthymematic. The rule for doing this fairly has a name.
The principle of charity: interpret an argument in the strongest form its words will reasonably bear. Where a sentence has two readings, one of which makes the argument obviously silly and the other of which makes it worth taking seriously, take the second. Where a premise is missing, supply the weakest one that does the job.
Two reasons, one honest and one practical. The honest one is that refuting an argument nobody made proves nothing; that error has its own name, the straw man. The practical one is that in law the opponent's argument will be restated by the court in its strongest form, and a pleader who has only prepared to meet the weak version will be answering the wrong case.
The limit of charity is that it is not invention. If the argument as made contains no premise capable of supporting the conclusion, the answer is that the argument fails, not that a better one could be constructed.
A worked example
"The witness cannot be believed. He gave a different account to the police. Anyone who changes his story is lying."
Interpret first. The third sentence, read literally, is a sweeping claim that is plainly false: people change accounts through error, through fear and through the passage of time. Charity asks whether a weaker reading is available. It is: "a witness who materially contradicts his earlier statement is not reliable on that point." That is a defensible proposition and it is what the speaker probably meant.
Now test the form. On the charitable reading the argument is:
Evaluating an Argument
1. A witness who materially contradicts his earlier statement is unreliable on that point.
2. This witness materially contradicted his earlier statement to the police.
Therefore this witness is unreliable on that point.
That is valid. Notice what has happened to the conclusion, though: the original said the witness "cannot be believed" at all, and the charitable reconstruction supports only unreliability on the point of contradiction. The argument, at its strongest, proves less than it claimed.
Now test the premises. Premise 1 is a proposition about how evidence is weighed and it is broadly correct. Premise 2 is a question of fact, to be established by putting the earlier statement to the witness. If it is not established, the argument is unsound.
Verdict. Valid on the charitable reading, but for a narrower conclusion than the one asserted, and sound only if premise 2 is proved. That three-part answer is what evaluation looks like when it is done properly, and no part of it is an opinion about whether the witness was in fact truthful.
Distinctions that carry marks
| Deductive argument | Inductive argument | |
|---|---|---|
| Claim made | The conclusion follows necessarily | The conclusion follows probably |
| Good version called | Valid | Strong |
| Good and true version called | Sound | Cogent |
| Admits of degrees | No | Yes |
| Destroyed by omitted evidence | No | Yes |
| Truth | Validity | |
|---|---|---|
| Belongs to | Propositions | Arguments |
| Tested by | Evidence, observation, law | Logic alone |
| Depends on subject matter | Yes | No |
What this does not mean
A valid argument is not a true one. Validity is a relation between premises and conclusion. An argument from false premises can be perfectly valid, and its conclusion may be nonsense.
Showing that an argument is bad does not show its conclusion is false. It shows the conclusion is unproved by this argument. Treating a refuted argument as a disproved conclusion is itself a fallacy.
Inductive arguments are not defective deductive ones. They are a different kind of reasoning with their own standard. Calling an inductive argument invalid is like calling a chair badly written.
Quick revision
Two independent questions: are the premises true, and does the conclusion follow?
Deductive: valid or invalid; sound means valid plus true premises. Only a sound argument establishes its conclusion.
Inductive: strong or weak; cogent means strong plus true premises plus no relevant evidence omitted.
Four combinations: only true premises with a valid form gives a guaranteed conclusion. The other three tell you nothing.
Principle of charity: read the argument in the strongest form its words will bear. The opposite fault is the straw man.
A bad argument for a true conclusion is still a bad argument, and the conclusion may be established some other way.
Evaluating an Argument
Test yourself
1. What are the two questions asked in evaluating an argument, and why are they independent?
Whether the premises are true, and whether the conclusion follows from them. They are independent because the first is a question about the world, answered by evidence or by law, while the second is a question about the argument's form, answered by logic and answerable without knowing whether any premise is true. An argument can pass either test and fail the other.
2. Define valid and sound, and say why both words are needed.
An argument is valid when the conclusion cannot be false if the premises are true, which is a matter of form alone. It is sound when it is valid and all its premises are in fact true. Both words are needed because validity alone establishes nothing: an argument from false premises may be perfectly valid, and only soundness guarantees the conclusion.
3. Why may "valid" not be used of an inductive argument?
Because validity means that the conclusion cannot be false if the premises are true, and an inductive argument never claims that. It claims only that the conclusion is probable, so the proper words are strong and weak, which admit of degrees as validity does not. Calling an inductive argument invalid criticises it for failing to do something it never attempted.
4. What does cogency require that soundness does not?
That no relevant evidence has been left out. A deductive argument is unaffected by additional information: adding premises to a valid argument cannot make it invalid. An inductive argument can be destroyed by a single further fact, so cogency requires strength, true premises and completeness of the evidence relied on.
5. State the principle of charity and give the name of the fault it prevents.
Interpret an argument in the strongest form its words will reasonably bear, choosing the reading that makes it worth answering and supplying the weakest missing premise that makes it work. The fault it prevents is the straw man, which is refuting a weaker argument than the one actually made and then claiming to have answered the real one.
6. An argument is shown to be invalid. What follows about its conclusion?
Nothing at all. Invalidity shows that this argument does not establish the conclusion; it does not show that the conclusion is false. The proposition may be true and may be provable by some other argument. Treating the refutation of an argument as the disproof of its conclusion is a distinct error and is common in first-year answers.
The rest of this subject
These notes are cut from the University's printed syllabus. Open the syllabus itself, or the past papers, for the same subject.