Conclusions feel certain because we have confidence in them. But this confidence does not come from evidence or logic. Deductive and inductive reasoning are tools to help us solve everyday puzzles, such as a broken window.
Imagine walking into a room and finding a broken window. Phyllis and Fred were both in the area, but no one saw what happened. We know the window is broken, we know two people were nearby, and we want an answer.
At first, the problem may seem simple. We gather a few facts, connect them, and decide what must have happened. Yet the kind of conclusion we can honestly reach depends on how we use the information we have and how much we still do not know.
➡ For More See: How We Reach Conclusions: What Can Go Wrong.
Deductive and inductive reasoning methods
We can approach the broken window in two different ways.
1. Deductive reasoning begins with premises and asks what must follow if those premises are true.
2. Inductive reasoning starts with existing evidence. It then seeks the conclusion that best fits that evidence.
Both methods can help us reason well, but they do different jobs. The difference becomes easier to see when we stop talking about them as abstract ideas and use them to solve the same problem.
We already know that the window is broken and that Phyllis and Fred were nearby. What we do not yet know is who broke it, whether someone else could have been involved, or whether another cause is possible.
That missing information matters because deduction and induction handle uncertainty differently.
Using deductive reasoning to solve the broken window
Suppose we begin with this argument:
- Only Phyllis or Fred could have broken the window.
- Fred did not break the window.
- Therefore, Phyllis broke the window.
The argument seems clear. If only Phyllis or Fred could have broken the window and Fred did not do it, then Phyllis must have done it.
That is the strength of deductive reasoning. When the premises are true, and the reasoning is valid, the conclusion must follow from them. The answer is not simply likely. It is required by the information we started with.
The problem is that deduction does not prove its own premises.
We said that only Phyllis or Fred could have broken the window, but how do we know that? Perhaps those are simply the only two people we noticed. Someone else may have entered the room earlier. Something from outside may have struck the glass. The window may even have failed because it was already damaged.
The second premise also needs support. We said Fred did not break the window. Perhaps Fred told us he was somewhere else. His statement gives us information, but it does not automatically prove that the statement is true.
The deductive structure is still sound:
If only Phyllis or Fred could have broken the window, and if Fred did not do it, then Phyllis did. What has changed is our confidence in the premises.
That distinction is important. Deductive reasoning tells us what must follow from the premises. It does not tell us that we were correct to accept those premises in the first place.
If we find that no one else could have broken the window and prove Fred didn’t break it, then our conclusion gets much stronger. Under those conditions, deduction gives us exactly what it is supposed to give us: a conclusion that must follow.
Trouble begins when uncertain information is treated as certain. This happens before deductive reasoning even begins.
Using inductive reasoning to solve the broken window
Now suppose we approach the same broken window without beginning with the claim that either Phyllis or Fred must have done it.
Instead, we begin with what we actually know. The window is broken. Phyllis and Fred were both somewhere in the area. We do not know exactly when the glass broke, who was closest to it, or whether another cause is possible.
Inductive reasoning prompts us to collect more evidence. This way, we can better decide how much we should trust any explanation we encounter. We might ask:
- When was the window last known to be intact?
- Where were Phyllis and Fred during that period?
- Was the glass pushed inward or outward?
- Were there witnesses?
- Is there camera footage?
- Was anything found near the window?
- Could a ball, branch, or other object have caused the damage?
Each answer changes the picture. Suppose a witness saw Phyllis near the window shortly before the glass broke. That increases the possibility that she was involved, but being nearby does not prove she broke it.
Now suppose Fred says he was across town, but another witness remembers seeing him in the building. His explanation becomes less reliable, so the possibility that he was involved becomes stronger.
Then we notice that most of the broken glass is inside the room. That may suggest something struck the window from outside. Suddenly, the idea that either Phyllis or Fred broke it from inside becomes less convincing.
Inductive reasoning allows the conclusion to move as the evidence moves. We may begin by thinking Phyllis is the most likely cause. New evidence may shift our attention toward Fred. Another discovery may point away from both of them.
The conclusion is not being forced into certainty. Its strength changes with the information available.
That does not mean induction is guesswork. We are still using evidence and reasoning. The difference is that we are not claiming more than the evidence supports.
At some point we might reasonably say:
Phyllis probably broke the window.
That statement is different from:
Phyllis definitely broke the window.
The word probably does not make the first conclusion weak. It tells us that the evidence points toward Phyllis while some uncertainty remains. The conclusion matches what we actually know.
When past behavior enters the investigation
Now suppose someone tells us that Phyllis has broken windows before.
That sounds important, and it may be relevant. A history of similar behavior can sometimes affect how likely an explanation seems. But it still does not tell us what happened to this window.
Phyllis may have broken ten windows in the past and still be innocent this time. Fred may never have broken a window before and still be responsible for this one. The danger appears when history begins replacing present evidence.
If we already expect Phyllis to be responsible, we may give every fact that points toward her extra weight. We may also dismiss evidence that points somewhere else. We are no longer simply asking what happened to the window. We are beginning to reason from what we already believe about Phyllis. Her history can remain part of the investigation, but it cannot become the investigation.
We still have to ask where she was, what happened to the glass, what other evidence exists, and whether another explanation fits the facts better. Bias can distort that kind of judgment even when the reasoning method itself is sound.
➡ For More See: Reasoning Errors: How Bias and Fallacies Distort Belief.
Then another piece of evidence appears. A camera shows that Phyllis walked past the room several minutes before the window broke. Fred entered the area shortly afterward, but the camera does not show what happened next.
The case has changed again.
Phyllis’s past history matters less now. The current evidence does not show her by the window at the key time. Fred becomes more interesting, but the video still does not prove that he broke anything.
We can now think Fred is more likely involved. Or, we might still feel there’s not enough info to pick one option. Inductive reasoning leaves room for that answer. Sometimes the best conclusion is not Phyllis did it, or Fred did it.
Sometimes the most accurate conclusion is:
We do not know.
We may never know how the window was broken.
How certainty can fail
The broken window shows how easily certainty can appear before it has been earned.
Our deductive argument sounded certain:
- Only Phyllis or Fred could have broken the window.
- Fred did not break it.
- Therefore, Phyllis broke it.
If those premises were established, the certainty would be justified. But when we looked more closely, we discovered that neither premise was as secure as it first appeared.
The certainty came partly from the way we framed the problem.
We decided there were only two possible people before proving that there were only two possibilities. We accepted Fred’s innocence before establishing it. Once those ideas became premises, the conclusion looked unavoidable.
The reasoning did not create the mistake. The uncertain premises did.
Inductive reasoning looks at the broken window differently. It doesn’t force us to assume we know all the possibilities.
As evidence points toward Phyllis, our confidence in that explanation can increase. If later evidence points toward Fred, confidence can shift. If we discover that a baseball came through the window from outside, both earlier conclusions may disappear.
This is why probability can sometimes give us a more accurate answer than certainty.
Suppose all the evidence strongly points toward Phyllis, but we still cannot rule out another cause. Saying that Phyllis probably broke the window recognizes both parts of what we know. The evidence supports her involvement, but it does not prove it beyond every reasonable possibility.
Saying that she certainly broke the window would sound stronger, but it would not make the reasoning stronger.
It would only make our statement more definite than the evidence allows.
Certainty can fail when we turn an assumption into a fact. It can fail when we decide that an unknown possibility cannot exist simply because we have not discovered it. It can fail when strong evidence is seen as proof. It can also fail when confidence in an explanation is mixed up with the evidence for that explanation.
The broken window also shows why deduction and induction are not opponents.
Suppose the investigation continues and gives us much stronger information. Camera footage proves that only Phyllis and Fred entered the locked room during the period when the window broke. Another camera shows Fred remained in a different part of the room and could not have reached the window.
Now our premises have changed. We can reasonably establish:
- Only Phyllis or Fred could have broken the window.
- Fred could not have broken the window.
With those premises established, deductive reasoning can now take us the rest of the way:
Therefore, Phyllis broke the window.
Earlier, induction helped us investigate an uncertain situation. As more evidence appeared, it allowed us to adjust which explanations seemed most likely. Once the evidence established the facts for a deductive argument, deduction could reveal what must follow.
The two methods can therefore work together.
Induction helps us reason through evidence that is incomplete or still developing. Deduction helps us determine what must follow when the needed premises are established.
The important question is not which method is better.
The question is how much confidence the method and the evidence allow us to place in the conclusion.
That same problem appears far beyond broken windows. People form opinions on relationships, politics, religion, spirituality, health, history, and more. Some facts are clear, but many things are still uncertain.
A conclusion can sound final long before the evidence makes it final.
Learning to recognize that difference helps us avoid turning what is likely into what must be true.
➡ For More See: A Rational Framework for Investigating Religious Claims →
Conclusion
We began with a broken window, two people nearby, and a question that seemed simple: who broke it?
Deductive reasoning showed us that if only Phyllis or Fred could have done it and Fred did not, then Phyllis must be responsible. The logic was clear, but the certainty depended on whether those starting premises were actually true.
Inductive reasoning kept the window open as an investigation. We collected evidence and looked at other causes. We adjusted our confidence as new facts emerged. When the evidence was unclear, we accepted that the answer could be uncertain.
Neither method failed.
The problem appeared when we tried to make the conclusion stronger than the information supporting it.
Deduction tells us what must follow when its premises are established. Induction shows what the evidence strongly supports, even with some uncertainty.
Sometimes the broken window can be solved with certainty. Sometimes the evidence only tells us who probably broke it. Sometimes the most accurate answer is that we still do not know.
Good reasoning means being able to tell the difference.
References
- Deductive and Inductive Arguments. Internet Encyclopedia of Philosophy.
- Validity and Soundness. Internet Encyclopedia of Philosophy.
- Inductive Logic. Stanford Encyclopedia of Philosophy.
- Logic and Probability. Stanford Encyclopedia of Philosophy.
- Abduction. Stanford Encyclopedia of Philosophy.
- Bayesian Epistemology. Stanford Encyclopedia of Philosophy.
- The Oxford Handbook of Thinking and Reasoning. Oxford University Press.
- Thinking and Reasoning: A Very Short Introduction. Oxford University Press.
- Human Reasoning: The Psychology of Deduction. Psychology Press.
- Inference to the Best Explanation. Routledge.
- The Foundations of Scientific Inference. University of Pittsburgh Press.
- Scientific Method. Stanford Encyclopedia of Philosophy.
References
About this project
The Seeker Project 4 Spiritual Exploration (SP4SE) is a project of Americans In Alliance. The mission of SP4SE is spiritual education. The mission of Americans In Alliance is environmental conservancy. Click the links below to learn more and support our efforts.