The activation of a frame is equivalent to recognizing familiar situations or things. But how do frames become activated? "We'll simply assume that every frame is activated by some set of recognizers. We can regard a recognizer as a type of agent that, in a sense, is the opposite of a K-line -- since instead of arousing a certain state of mind, it has to recognize when a certain state of mind occurs."
NB: As Minsky notes later in this section, the big issue here is fleshing out the mechanisms for frame recognition. Does New Connectionism provide any insights into how this problem might be solved?