There have been many attempts to define Radical Behaviorism. Most attempts are in terms of inclusion and exclusion, i.e. what radical behaviorists talk about, and what they do not talk about. More often definitions focus on solely on exclusion, providing a negative definition in which behaviorists are defined based on what they don’t do, rather than on what they do. However, this minimizes the profoundness of the approach. A simple, positive definition is: Radical behaviorists claim that all questions about psychology are questions about behavior. One is tempted to say something like “all interesting questions about psychology”, but that is unnecessary, as the converse of the above claim is also made: All questions that are not about behavior are not about psychology. These claims are historic and inclusive; the behaviorist is not trying to redefine psychology, rather to point out what psychology has always been. Thus, behaviorists are not, as is commonly believed, trying to deny the existence of phenomenon typically handled by psychology. Quite to the contrary, they are trying to argue the traditional questions can only be answered through careful observation and analysis of behavior.
For example: You might ask: “At the dinner table, my child is really good at telling me what happened in school each day. How does he remember what happened earlier in the day?”
The behaviorist asserts that your question can be answered completely, and without remainder, by explaining how your child’s behavior becomes a function of things that happened in the past. In particular, it sounds like you are interested in how his verbal behavior comes to be a function of what happened earlier. That is, to see how a child develops the ability to remember, we must (at a minimum) start with a child whose behavior is not a function of the past events, and observe them until their behavior is a function of the past events. If we determine what changed, we will know how he “remembers”.
You might protest: “But that doesn’t really answer the question. A child could correctly report what happened earlier without remembering! They could be responding to cues you gave them, making lucky guesses, or remembering the homework they just finished rather than the class lesson from hours ago.”
True enough. You have correctly identified one way in which the behaviorist’s version of the question is difficult to answer, but the term “function of” does not merely mean that the events correspond in any particular case. It is not enough to know that the verbal behavior matches a particular past event, the verbal behavior must be a function of those past events. To know that the child remembers the past event, we must demonstrate not merely that the two events correspond, but that the latter event changes as a function of the specific earlier events in question, rather than as a function of alternative past or present factors. 
It is likely that determining whether or not a given pattern of verbal answers is a function of earlier events will require us to observe many situations that have not yet been observed, and perhaps even situations that would not occur naturally, i.e. we may need to experiment, to manipulate variables. Yes, the empirical burden is difficult to meet; however, the fact remains that once I answer the question of how behavior becomes a function of past events, I will have answered the question of how children come to remember things.”
 This is just as it is not enough for physicists to show that a single object happened to fall towards the center of the earth, they demonstrated that objects move as a function of the center of mass of other objects – move the center of mass, and the behavior of nearby objects changes accordingly.