The author agrees with Dorpat's view, but prefers the term functional introject because specific functions of the analyst, rather than his person, are at issue. For introjection to occur, a boundary between self and object must be established in the patient's mind. Prior to that, preintrojective use of the analyst as a transitional object may be necessary. A case example is given in which countertransference reactions and resulting misinterpretations are seen as having usefully facilitated such a process.