Yes, it's an unfortunate clash of terminology. When class A extends class B, it inherits all of B's fields and methods as well as those of B's parent, and B's parent's parent, and so on. However, as far as A is concerned, all of those fields and methods were inherited as a single unit directly from B. In that regard, it's single inheritence.
Multiple inheritence specifically refers to one class having two direct parents as opposed to a single parent and any number of grand-parents. In Java it would look like this:In the beginning it can seem somewhat arbitrary given that a class can implement multiple interfaces:However, implementing an interface doesn't involve the complexity of inheriting instance fields and method bodies. Instead, the interface merely provides a set of method signatures for which the implementing class must provide bodies (unless it's declared abstract).
It's just a matter of definition of terms. The first situation presents all sorts of nasty issues -- for example, dealing with ambiguities if class C inherits a method named foo() from both A and from B. But in the second situation, there's no ambiguity at all: C will get B's version of foo().