Perhaps the most common kind of inductive 
argument is called an inference to the 
best explanation. 
And to understand this kind of argument 
we need to think way back to the first 
week when we discussed the difference 
between justification and explanation. 
If somebody uses an argument to justify a 
conclusion then they're trying to give a 
reason to believe that the conclusion is 
true. 
And maybe you didn't believe that 
conclusion before you heard the argument. 
But if somebody gives an argument. 
To explain, the conclusion. 
Then they're taking it for granted, that 
the thing they're explaining is true. 
An they're trying to, understand. 
Why it's true. 
An the argument is supposed to help us 
understand, why it's true. 
The weird thing about an inference to the 
best explanation is it really combines 
these two. 
It uses the argument to justify belief in 
the conclusion, that is to give you a 
reason to believe that it's true, and yet 
a premise of the argument is that, that 
conclusion explains a phenomenon that you 
took for granted. 
Inferences to the best explanation are 
extremely common. 
For example, almost every detective story 
uses an inference to the best 
explanation. 
Just imagine that there are three 
suspects. 
And you know that one of those three 
suspects must have done it. 
Because the door wasn't jimmied open. 
So, whoever got into the room must have 
had a key. 
And there are only three people who had 
keys. 
But one of'em couldn't have been the one 
who did it. 
Because she has really small feet. 
And the footprints outside the door were 
really large and so it couldn't have been 
her. 
And the second person couldn't have done 
it because they have an alibi on the 
other side of the town. 
So it must have been the third person who 
did it. 
You notice what you did is, you can't 
explain why the door wasn't jimmied by 
the person who got in. 
Because. 
there were no marks, and that means that 
the best explanation, that must be that 
the person that did it, had a key. 
And the best explanation why the 
footprints were so large, is that it was 
a large person who wore large shoes. 
And the best explanation of why the alibi 
would place the person on the other side 
of town was that they were being honest. 
Assuming it's not their brother or 
they're not the kind of person who would 
lie, or so on. 
So the best explanation of why the jewels 
are missing from the room, or why there's 
a dead body in the room, is that this 
third suspect is the one who did it. 
So all the, detective stories work like 
that. 
It's almost always an inference to the 
best explanation. 
The best explanation is, he's the one who 
did it. 
One implication of this, is that no 
matter what Sherlock Holmes says, the 
form of reasoning that he was using was 
inductive reasoning, not deduction, as he 
claimed. 
It's not only detective novels. 
It's also science abuses inference to the 
best explanation. 
Take for example the greatest murder 
mystery in history. 
What killed the dinosaurs? 
Well some people think that mammals 
eating their eggs, 
not a chance. 
That doesn't explain why they were also 
mass extensions in the oceans, where the 
mammals weren't eating their eggs. 
Well some people think that, the best 
explanation is that. 
There were big volcanic eruptions in 
Southern India. 
That seems pretty unlikely too, because 
the volcanoes that we know didn't put out 
enough material to kill things all the 
way around the world. 
So, they think of the best explanation of 
what killed the dinosaurs is that a giant 
meteor hit in the Yucatan Peninsula, in 
Mexico. 
And that produced lots of soot in the 
atmosphere that killed the dinosaurs. 
As well as many other species at the time 
but luckily, not mammals. 
And so, we now have a best explanation in 
science. 
And a lot a scientific theories work that 
way. 
As a matter of fact, the point of most 
scientific theories is to provide, the 
best explanation of the experimental 
data. 
that has been observed by scientists 
throughout the world. 
And of course, we also use inference of 
best explanation in our everyday lives, 
to try to figure out what's going on 
around us. 
Here's an example. 
Whoa, what was that. 
A drop of water just hit my head. 
There must be a leak in the roof. 
Now that is an inference to the best 
explanation. 
Starts with an observation, of the water 
hitting my head. 
Then I think about, why the water would 
be coming down. 
And the best explanation I can come up 
with, is that there's a leak in the roof. 
So I conclude that there is a leak in the 
roof. 
Notice that this argument uses an 
explanation but it runs in the opposite 
direction from an ordinary explanation. 
In an explanation, the premises explain 
the conclusion and the conclusion 
describes the phenomenon that was 
observed. 
When there is a leak in the roof, water 
drops through the roof. 
When water comes through the roof, it 
drops onto whatever is in its way. 
I am in its way. 
Therefore, water drops onto me. 
This explains that. 
In contrast, an inference to the best 
explanation works in the opposite 
direction. 
The conclusion is what does the 
explaining. 
The premises are what get explained, and 
the point is to justify belief in the 
conclusion, not to explain it. 
Water drops onto me. 
Leak in the roof might explain why water 
drops onto me? 
No better explanation is available. 
Therefore, there must be a leak in the 
roof. 
Notice that the conclusion of the left 
argument is a premise, of the right 
argument. 
And, the conclusion of the right argument 
is part of a premise, of the left 
argument. 
So, the form is different. 
The function is different as well. 
In the explanation on the left we start 
out already knowing that the conclusion 
is true. 
In the interest to the best explanation 
on the right, we argue that the 
conclusion is probably true because it is 
the best explanation of the observation 
in the first premise. 
That's why we call it an inference to the 
best explanation. 
Is this argument valid? 
Yes or no? 
Right, it's not valid. 
Because it's possible for the premises to 
be true and the conclusion false. 
Now does it's invalidity make the 
argument bad? 
Yes or no? 
Right. 
It does not make it bad, because an 
inductive argument can be good, even if 
it's invalid. 
The next question's a bit trickier. 
Remember that an argument is defeasible 
if further information can make the 
argument weak. 
So, is this argument defeasible? 
Yes or no? 
Right. 
It is defeasible because further 
information could make this argument 
weak. 
So what? 
Does its defeasibility mean that this 
argument was weak even before we got that 
additional information? 
Yes or no? 
No. 
Because inductive arguments can be 
strong, even though they are defeasible. 
So now we know that this argument is 
invalid, and is defeasible but it still 
might be strong and good.