As judgement, and thus, purpose of executing our actions are supplied by our VP (which use our personal B-ROMs), when we want to understand other people's actions, we mimic their actions mentally using the first stage of action, as if we are doing the same action (reverse process of executing the action) to judge what they are. In other words, instead of making the judgement and then going to the first stage (in order to execute the action), we, based on our past experiences, execute the first stage by mentally mimicking the action, and then judging its purpose, e.g. "What would I be doing, based on the situation the other person is in, when I would be doing such an action".
This process is popularly known as simulating an action, but I will call it "Dry Running the Action", which is another name for the same, as I find it more suitable.
Having the same body structure as other people, dry running actions (as explained above) helps a person's VP decode purposes of their actions, which it does using past data stored in his personal B-ROM (or using summarized data from LB in case of repetitive actions), using contextual cues of respective situations. In short, by recalling what he would have done while performing action the other person is doing in a particular situation, a person is able to judge his purpose of executing it.
Repetitively judging interactions this way not only makes the process of decoding actions automatic
(more under title Repetitive Interactions and Condition Based Repetitive Interactions), but enables a person to extrapolate reasoning others use for different interactions (which plays a major role when growing children learn from others), using which, a person starts building CB-ROMs for them.
Such extrapolation of reasoning further helps a person decode mental states of others, like desires, beliefs, opinions, hopes, wishes, etc., while lack of same can lead to mental disorders
(more under title DF Analysis and Autism).