Let us disentangle identification as it occurs in the structure of a neurotic symptom from its rather complicated connections. Supposing that a little girl (and we will keep to her for the present) develops the same painful symptom as her mother--for instance, the same tormenting cough. Now this may come about in various ways. The identification may come from the Oedipus complex; in that case it signifies a hostile desire on the girl's part to take her mother's place, and the symptom expresses her object love towards her father, and brings about a realisation, under the influence of a sense of guilt, of her desire to take her mother's place: 'You wanted to be your mother, and now you _are_--anyhow as far as the pain goes'. This is the complete mechanism of the structure of a hysterical symptom. Or, on the other hand, the symptom may be the same as that of the person who is loved--(so, for instance, Dora in the 'Bruchstück einer Hysterieanalyse'[40] imitated her father's cough); in that case we can only describe the state of things by saying that _identification has appeared instead of object-choice, and that object-choice has regressed to identification_. We have heard that identification is the earliest and original form of emotional tie; it often happens that under the conditions in which symptoms are constructed, that is, where there is repression and where the mechanisms of the unconscious are dominant, object-choice is turned back into identification--the ego, that is, assumes the characteristics of the object. It is noticeable that in these identifications the ego sometimes copies the person who is not loved and sometimes the one who is loved. It must also strike us that in both cases the identification is a partial and extremely limited one and only borrows a single trait from the person who is its object.