There are two senses of *contrastive* explanation that are confused with each other too often.

Consider we ask

“Why did my friend buy a house?”.

There are two types of contrast, one of which is at the conceptual core of explaining while the other is not, although the latter might be methodologically useful.

The first type of contrast introduces a counterfactual alternative (Y*) to the actual outcome (Y).

“Why did my friend by a house rather than not?”.

Explaining this contrast requires some actual X and counterfactual X* are described such that: had X* instead of X been the case, Y* instead of Y would have been the case.

“Because my friend had the money and saw the house as a good investment. Had she been poor, she would not have bought the house”.

This type of contrast is conceptually required by explanations.

The second type of contrast compares an actual outcome to another actual outcome, where the two outcomes differ.

“Why did my friend by a house and I did not?”

Here we introduce a difference between me and my friend that accounts for the different outcome.

“Because my friend has a real job (and thus has the money) while I am a philosopher (and thus poor)”.

This type of contrast is dependent on the first type. The answer above presupposes that

“Had my friend been a philosopher, she would not have bought a house”

is true. Otherwise, the difference in our jobs does not explain anything. To see this, consider some arbitrary difference:

“My friend rather than I bought a house because she has black shoes and I have red shoes”. This does not explain anything. Why? Because it is not true that

“Had my friend had red shoes, she would not have bought the house”.

The second type of contrast is, therefore, is not something we introduce to provide an explanation. In itself, it is not explanatory at all. Only if explicated with the structure of the first type of contrast, the second type of contrast is explanatory.

However, the second type of contrast is important in order to formulate explanatory hypotheses. To compare two situations with different outcomes is a way to isolate explanatory causes. However, to make the final judgement of explanatory relevance, we need the first type of contrast.

Bonus: There is also a third sense of contrast: It is sometimes said that we need to “contrast” different possible causes with each other to find the actual one. For example, my friend might have bought the house because she saw it as a good investment; because it was expected by her family; because she was forced to. Once we know that, actually, her family did not expect her to buy a house and that she was not forced, we know that she bought it because she saw it as a good investment.