The scientific exploration of prepsychotic detection and intervention in psychosis has just commenced. To identify developing psychosis at prodromal stages, it is important to learn how patients and families perceive initial prodromes naturalistically. We must understand better what we are going to detect, because the essential components of this phase, particularly the subjective experiences, remain unsettled. In a series of 19 first episode DSM-IV schizophrenia patients, we explore prodromal phenomena in depth and identify potential core dimensions. On the basis of our findings, we describe experiences and behaviors that appear to be essential components of initial prodromes. The subjects reported serious difficulties interpreting and talking about prodromal experiences at the time these occurred, causing delayed identification. We report detailed reasons for this, pointing out vulnerable aspects of at-risk assessments. From eight proposed groups of experiences, two are highlighted as tentative core dimensions: "disturbance of perception of self" and "extreme preoccupation by and withdrawal to overvalued ideas." Four potential dimensions of prodromal behavior are also identified: (1) quit school, university, or job, or major school truancy, (2) marked and lasting observable shift of interests, (3) marked and lasting social passivity, withdrawal, or isolation, and (4) marked and lasting change in global appearance or behavior. We argue that the findings, the phenomena, and their significance in prodromes are valid because they are logical and coherent in light of clinical experience as well as the empirical literature of a full century.