Philosophy & Ethics · Reasoning & arguments
What makes an argument good?
A good argument needs a conclusion that follows from its premises, and premises that are true. Learn to check each separately, and to catch an argument that sounds right and proves nothing.

What you will learn
- Identify premises and a conclusion in a short argument
- Tell giving a reason apart from repeating a claim
- Reconstruct an argument by supplying the unstated premise it depends on
- Apply the definition of deductive validity
- Tell a valid argument apart from a sound one
- Use validity in reverse to show that at least one premise must be false
Lesson 1 of 4 · 6 min
Find the missing sentence in any argument
"She trained for months, so she'll win." "The shop lights are off, so nobody's inside." You hear arguments like these all day, and each sounds complete. Each also leaves a sentence out, and that sentence is often the one worth arguing about.
To find it, you need the parts of an argument first. Take one with nothing missing. "This room is available, because every room on this floor is available, and this room is on this floor."
The claim being supported is the conclusion. Here it's that this room is available. The statements offered in support are the premises. In philosophy, an argument is just that pairing of a claim with its reasons. No raised voices are needed.
Write the argument out one line at a time. Premise one says every room on this floor is available. Premise two says this room is on this floor. The conclusion says this room is available.
Notice that the conclusion came first in the original sentence. Position doesn't mark it. Its job does. It's the line the others are there to support, and words like "because" and "so" tell you which is which.
Laid out this way, you can examine the support instead of being carried along by a fluent sentence.
The word "because" can also promise a reason and deliver nothing. "The room is available because it's free to use" says the same thing twice, if "available" and "free to use" mean the same here. The claim has been restated, not supported.
A real reason is a separate fact that makes the conclusion more believable. "The room is available because today's booking sheet is empty" gives you one. You could check the sheet without already knowing the answer.
Everyday arguments also leave premises unsaid. "Members get in free, and Jo is a member, so Jo gets in free" quietly assumes that Jo's membership hasn't expired and that the offer applies tonight.
Saying those assumptions out loud shows where a disagreement lies. One person may doubt that Jo is still a member. Another may accept that and doubt the offer covers tonight's show. Those are different objections, and each needs a different answer.
Try to recallA friend says, "The proposal is fair because it's just." What would you ask to find out whether that gives a reason or only restates the claim?
Ask what makes it fair. You're looking for a feature of the proposal, such as how it splits the costs, that you could check without already agreeing it's fair. Here "just" is only another word for "fair."
You can dig out an unstated premise by routine instead of by hunch. Write down the premises that were actually said. Write the conclusion under them. Then ask what else would have to be true for those premises to reach that conclusion, and write it as a full sentence.
The writing matters. An assumption can pass unnoticed while it's silent and look far too strong once it has a line to itself.
Try it on "She trained for months, so she'll win." The premise is about training and the conclusion is about winning, so the missing line has to join the two. It would say that anyone who trains for months wins.
Written out, that's plainly false. Every runner in the race may have trained for months, and only one can win. Swap in something believable, such as "training for months makes winning more likely," and the argument now supports a weaker conclusion. She's more likely to win.
That swap is the fair way to do it. Supply the weakest premise that still gets the argument where it's going, not the silliest one you can think of. You're trying to find out whether the argument works, and beating a version nobody meant tells you nothing.
If the speaker would accept that premise, you've made their case clearer. If even the weakest version is hard to believe, you've found the real disagreement.
So when an argument sounds right, slow it down. Find the conclusion, list the premises that were said, and write out the missing sentence in full. Often that sentence is where the disagreement was hiding all along.
Lessons in this topic
Find the missing sentence in any argument
An argument is a claim plus the reasons given for it. Learn to pull them apart, spot a reason that only repeats the claim, and write out the premise nobody said aloud.
A valid argument can be false from start to finish
Validity is about whether a conclusion follows. Truth is about whether the premises are right. See why logicians keep the two apart, and how a false conclusion can expose a false premise.
A true conclusion can come from bad reasoning
An argument can have true premises and a true conclusion and still prove nothing. Learn to expose the gap by building a counterexample, step by step.
Why logicians swap words for letters
Replace an argument's content words with letters and you see its logical form, the thing that makes it valid or invalid. Grammar doesn't always show that form honestly.
Three quick stories from this topic
A lame duck isn't a duck
Logic lets you conclude a ferocious tiger is a tiger, but not that a lame duck is a duck
"Tony is a ferocious tiger" and "Clinton is a lame duck" have the same grammar. Only the first means a tiger that is ferocious. Because grammar hides differences like that, logicians write arguments in symbolic languages where it can't.
Why the inference fails
From "Tony is a ferocious tiger" you may validly infer that Tony is a tiger, because the sentence has the form "x is an A that is F." The lame-duck sentence doesn't have that form, so the parallel inference to "Clinton is a duck" isn't valid, however natural it looks.
Why logicians build artificial languages
Because grammatical form can diverge from logical form, contemporary logicians use symbolic languages in which the two coincide. The study of which deductive forms are valid is called formal or symbolic logic.
True lines, broken argument
Every statement in an argument can be true while the argument itself is still invalid
"All popes reside at the Vatican. John Paul II resides at the Vatican. Therefore, John Paul II is a pope." All three lines were true when that was written. But nothing in the premises says only popes live there, so they don't guarantee the conclusion.
The same pattern proves the Earth is a basketball
All basketballs are round. The Earth is round. So the Earth is a basketball. That argument has the same form as the pope argument, and the form is invalid whatever words fill it.
A bad argument doesn't make its conclusion false
An invalid argument can still end on something true. Finding the flaw shows only that this argument hasn't established its conclusion.
Time-travelling toasters
An argument concluding that all toasters are time-travel devices can be perfectly valid
Its premises are that all toasters are made of gold and that everything made of gold is a time-travel device. Both are false. Validity only asks whether the conclusion would have to be true if the premises were, and here it would.
What validity checks
An argument is valid when it's impossible for its premises to be true and its conclusion false. The premises don't have to be true for that to hold.
Sound asks for more
A sound argument is valid and has true premises, so its conclusion is true. The toaster argument is valid but not sound.
Sources
- Validity and Soundness, Internet Encyclopedia of Philosophy
- lame duck, Cambridge Dictionary, Cambridge University Press




