1 1 An abbreviated version of this paper was read before the British Psycho‐Analytical Society on 5 November, 1958.

      2 2 Although in this paper I shall usually refer to mothers, it is to be understood that in every case I am concerned with the person who mothers the child and to whom he becomes attached, rather than to the natural mother.

      3 3 This account is adapted from those given in Robertson and Bowlby (1952) and Robertson (1953a).

      4 4 In the previous paper and the early drafts of this one the term ‘denial’ was used to denote the third phase. The problem of terminology is discussed more fully after the phase of detachment has been described.

      5 5 As explained in my previous paper (Bowlby 1958), ‘the terms primary and secondary refer to whether the response is built‐in and inherited or acquired through the process of learning.’

      6 6 See section ‘Ingredients of Separation Anxiety’ below.

      7 7 For this abridged version a critical examination of psycho‐analytical theories relating to separation anxiety has been omitted. It is being published as a separate paper: Bowlby, J. (1960). Separation anxiety: A critical review of the literature. Journal of Child Psychology and Psychiatry, 1 (4), 251–269.

      8 8 In an earlier paper (Bowlby 1957) I have used the term ‘instinctual response’ to refer both to the behaviour and to the hypothetical internal structure which, when activated, is presumed to lead to the behaviour. To avoid this confusion I am now using the term ‘instinctual response system’ for the hypothetical internal structure, and limiting the term ‘instinctual response’ to the active behavioural response, including both the motor behaviour pattern and its physiological and psychological concomitants.

      9 9 A similar view, though coupled with a materially different theory of instinct, was advanced by McDougall (1923).

      10 10 The possibility that a single emotion, fright, may accompany more than one instinctual response system suggests that Lorenz’s hypothesis that each response is accompanied by an emotional state peculiar to itself may need modification.

      11 11 See the discussion by Thorpe (1950, p. 390). Several workers (e.g. Hinde 1954a) have shown that, paradoxically, strangeness evokes both escape and curiosity, and that there is a complex balance between the two competing response systems.

      12 12 Such situations have been termed ‘consummatory situations’ by Bastock, Morris and Moynihan (1953) and by Hinde (1954b). In my view, however, partly because of the usefulness of the verb ‘to terminate’, a preferable term is ‘terminating situation’.

      13 13 ‘The tendency for instinctual responses to be directed towards a particular individual or group of individuals and not promiscuously towards many’ (Bowlby (1958, p. 370).

      14 14 The term ‘haven of safety’ has been introduced by Harlow and Zimmermann (1958). In describing their very interesting experiments with rhesus monkeys they write: ‘One function of the real mother, human or sub‐human, and presumably of a mother surrogate, is to provide a haven of safety for the infant in times of fear or danger.’ See also Harlow (1958).

      15 15 The secondary drive theory, which they invoke to account for the child’s tie to his mother and for separation anxiety, is not necessary to an interpretation of their data.

      16 16 In his experiments with goat kids Liddell (1950) has demonstrated the very different responses to fright‐evoking stimuli of identical twins according to whether or not they were with their mother: the twin with his mother roamed about naturally and seemed relaxed, whereas the one without his mother froze almost immobile in a corner of the room.

      17 17 Of Schaffer’s twenty‐five subjects, sixteen were aged over 28 weeks and nine 28 weeks and under. Of the two younger infants who deviated from the usual behaviour, one was already 28 weeks of age, and so on the margin of the older age group, and the other was thought to be missing his dummy.

      18 18 Anthony writes: ‘It