Ambivalence: When We Love and Hate the Same Person

You put the phone down after a conversation with your mother, and something is boiling inside you that has nowhere to go. It lasts a few minutes. Then the second wave arrives, more insidious than the first: guilt. How can I feel this way about my own mother? I love her. And somewhere in there, quietly, a conclusion settles in that may stay for years: something is wrong with me.

That conclusion rests on an assumption our culture rarely questions: that real love is clean. That where there's a genuine bond, there's no room for anger, and if anger shows up, either the love was fake or we're defective. Psychoanalytic thinking has argued the opposite for over a century. The capacity to feel love and aggression towards the same person, without the relationship collapsing inside us, is not the problem. It's one of the most significant achievements of psychological life.

A term with a history: from Bleuler to Freud

In 1910, the Swiss psychiatrist Eugen Bleuler coined the word Ambivalenz to describe opposing feelings, thoughts or wishes coexisting towards the same object. Freud adopted it almost immediately and gave it a central place in his theory. He had, in fact, already described the phenomenon years earlier without having a name for it: in the case of Little Hans, the five-year-old with a phobia of horses, Freud saw a boy who loved his father and competed with him at the same time, and who couldn't tolerate that internal collision.

In Totem and Taboo (1913), ambivalence becomes almost anthropological: Freud places it at the heart of the oldest human institutions. And in Mourning and Melancholia (1917) it acquires perhaps its most moving dimension. Every loving bond carries opposing currents within it, and this becomes painfully obvious when we lose someone. Grief isn't only sorrow for the loss. It's also a reckoning with a relationship that, like every real relationship, was never purely luminous.

It's worth pausing on this. From the very beginning, the term did not describe pathology. It described the ordinary, unavoidable complexity of every human bond.

A clarification is needed here, because the word has drifted into everyday language and lost its precision. Ambivalence doesn't mean "mixed feelings" in the sense of confusion, and it isn't indecision. It means two full, simultaneous currents towards the same person: I genuinely love and I genuinely resent, not a lukewarm blend of a bit of each. The clinical difference matters. Someone who "doesn't know what they feel" is often defending against their own ambivalence, because fog is easier to bear than the clarity of two feelings that seem irreconcilable.

Klein: splitting and the two "mothers"

For an infant a few weeks old, the world doesn't yet contain unified people. It contains experiences. There's the experience of warmth, milk, arms that hold: a good world. There's also the experience of hunger that goes unanswered, of cold, of absence: a world that threatens.

Melanie Klein, working analytically with very young children, argued that the infant initially organises these experiences as though they came from two separate sources: a "good" mother who provides and a "bad" mother who frustrates. This division is called splitting, and in Klein's theory it forms the core of the paranoid-schizoid position, the earliest way the psyche organises experience. "Position" here doesn't mean a stage that passes and ends. It means a mode of functioning we can return to throughout life, especially under pressure.

Splitting isn't a design flaw. For the infant it's necessary: it protects the image of the good mother, on whom survival absolutely depends, by keeping the "bad" away from her. As explored in the article on the emotional bond between mother and infant, this early relationship is the workshop where the foundations of psychic life are built. The trouble starts when a mechanism that once protected us goes on governing our adult relationships: when a partner is either ideal or a disappointment, a friend either loyal or a traitor, with nothing in between.

The depressive position: psychoanalysis's most misleading name

Few psychoanalytic terms have been so poorly served by their own label. Klein's "depressive position" doesn't describe depression, and it isn't something to be cured. It describes an achievement: the moment a child begins to grasp that the mother who satisfies and the mother who frustrates are the same person.

That discovery comes at a cost, which is why Klein chose a weighty word. If the mother is one person, then the anger I felt towards the "bad" mother struck the very person I love. Here a wholly new feeling is born, unknown in the earlier phase: guilt over the damage I fear I've done to what I love. And with it comes something equally important, the wish to repair: to mend, to reconnect, to care.

There's a mourning hidden in this transition too. The child says goodbye to the ideal mother, the one who would be only good, always available, with no needs of her own. In her place it gains a whole, real mother: less perfect, infinitely more true. Psychological maturity, in Kleinian thinking, isn't freedom from "negative" feelings. It's the capacity to hold them alongside love, for the same person, without the bond shattering.

Reparation, in fact, isn't confined to childhood. It's probably the most everyday adult expression of the depressive position. Every honest apology after a row, every "I didn't mean it like that, can we talk again", every attempt to mend a relationship we've damaged, rests on the same psychological capacity: to bear having hurt someone I love, without collapsing into guilt and without denying it. Couples who last aren't the ones who don't fight. They're the ones who learnt to repair.

Winnicott: the hatred inside maternal love

In 1949, Donald Winnicott published a paper that still startles readers with its bluntness. In "Hate in the Countertransference" he wrote something few would say publicly: the mother hates her infant before the infant can hate her. He then listed, with almost provocative calm, eighteen reasons why. The baby is ruthlessly demanding and gives nothing back. It interrupts her life, her sleep, her body. It treats her as an unpaid servant. It hurts her, literally, while feeding.

Winnicott wasn't indicting mothers. He was freeing them. His argument is deeply subversive: a mother who can recognise her own hatred without acting on it, who can hold it inside while continuing to care, gives her child something precious. She demonstrates, long before any words, that love isn't destroyed by difficult feelings. He noted, with his characteristic humour, that folk wisdom always knew this: even nursery rhymes contain images of falling and disaster, giving the feeling a safe, sung outlet.

For any parent who has felt guilty for losing their temper with their child, for wanting for one moment to be anywhere else, that 1949 paper offers a different reading. The feeling doesn't make you a bad parent. Denying it, on the other hand, can become dangerous, because what isn't acknowledged can't be regulated.

How we learnt that anger cancels love

If ambivalence is so ordinary, why does it frighten us? The answer usually lies in personal history. No child is born believing that anger cancels love. It learns it.

It learns it when its anger meets withdrawal: a mother who "freezes" and pulls away for hours after every outburst. When it meets punishment aimed not at the behaviour but at the feeling itself: don't you dare speak to me like that. When it meets frustration wrapped in guilt: I sacrifice everything for you and this is how you treat me. In every version, the message reaching the child is identical: love is conditional, and your anger threatens it.

A child can't risk the bond it depends on. So it does the only thing available: it cuts itself in two. On one side, the acceptable self, the one who smiles, agrees, causes no trouble. On the other, the hidden self, where the anger, the envy, the disagreement all get stored, everything that proved dangerous to the relationship. The long-term cost is considerable: the adult who emerges struggles to assert, to say no, to tolerate conflict without panic. The article on boundaries and differentiation examines this difficulty closely, since being close without disappearing depends on precisely the capacity that was injured here.

There's a second route to the same conclusion, and on the surface it looks like the opposite. In some homes, anger wasn't punished. It ruled. Shouting that shook the walls, explosions that terrified, a parent whose rage seemed capable of destroying everything. A child growing up there doesn't learn that anger is forbidden. It learns that anger destroys. Its conclusion is quieter but just as binding: I will never be like that. In adult life, its own anger frightens it, because it carries an image of what anger can do when let loose. Two entirely different childhoods, one identical adult belief: this feeling has no place inside love.

The psychoanalytic concept describing the healthy outcome is object constancy: the internal certainty that the other person remains whole and remains ours even when they disappoint us, even when we're furious with them. Those who never built it experience every conflict as total loss. A disagreement with a partner isn't a disagreement, it's the end. Criticism from a friend isn't criticism, it's betrayal.

What happens in the brain

Neuroscience offers an interesting, if incomplete, biological parallel to this psychological work. Intense emotion such as anger primarily engages subcortical structures, the amygdala foremost among them: fast, automatic, binary responses built for survival rather than fine distinctions. Holding two contradictory representations at once ("I'm furious with him" and "I love him") appears to depend on more evolved networks, with significant involvement of the prefrontal cortex and anterior cingulate, regions implicated in emotion regulation and conflict processing.

Research on cognitive and emotional ambivalence shows that maintaining contradictory attitudes is mentally expensive: it demands more resources, more time, greater tolerance of discomfort. That fits the clinical observation that under acute stress or exhaustion, even psychologically mature people slide back into black-and-white thinking. In neurobiological terms, splitting is the cheapest form of processing. Integration is the most costly and the most fragile.

One caveat is owed here. Psychoanalytic concepts and neuroscientific findings don't map onto each other one to one. The "depressive position" isn't located in any particular circuit, and no fMRI can show reparation. What can be said with confidence is more modest but not trivial: the picture neuroscience sketches, of a brain that slowly and effortfully learns to integrate contradictory signals, runs parallel to the picture psychoanalysis sketched decades earlier using entirely different tools.

Where we meet it: how unprocessed ambivalence shows up

In the consulting room, difficulty with ambivalence rarely announces itself by name. It arrives in disguise, and the disguises are recognisable.

In romantic relationships it often takes the form of idealisation and devaluation cycles. The new partner is everything that was missing: unique, perfect, finally. Until the first serious frustration, inevitable in any real relationship, demolishes the whole picture. The same person becomes a stranger, a mistake, he was never who I thought he was. Otto Kernberg described this cycle as central to borderline personality organisation, but in milder versions it turns up almost everywhere. It frequently coexists with intense fear of abandonment, which, as examined in the article on emotional dependency in relationships, has its own roots in attachment history.

In grief, ambivalence becomes perhaps the most silent source of pain. Few things are harder to admit than anger towards someone who has died: anger that they left, that they neglected their health, that they left unfinished business behind, that they were also harsh and not only loved. Freud saw as early as 1917 that mourning becomes harder and deeper when the relationship with the lost person was strongly ambivalent. The stages of grief rarely unfold in a straight line, and unspoken anger is one of the commonest reasons the process gets stuck. Grief that makes room for anger is, paradoxically, grief that can be completed.

In parenting, unprocessed ambivalence becomes guilt that isn't spoken aloud to anyone: I got so angry with my child that I frightened myself. And towards ourselves, finally, it takes the form of an internal black-and-white trial: either I'm worthy or I'm worthless, depending on the day. The person who feels omnipotent on the morning of a success and utterly defeated on the evening of a mistake doesn't have two selves. They have one self that hasn't yet learnt to hold both what it achieved and what it failed. Self-esteem that survives over time isn't built on perfection. It's built on the same capacity for integration this whole piece describes, turned inward.

The therapeutic approach: a space with room for both

As a psychotherapist, I've learnt to pay attention to a particular moment in therapy: the moment the person opposite me becomes angry with me. It can arrive as a complaint about something I said, disappointment about something I failed to grasp, sarcasm, lateness, a thought about stopping altogether. However paradoxical it sounds, that moment is often among the most valuable of the entire journey.

There's another ambivalence that appears in nearly every therapy, before we even get to other people: ambivalence towards change itself. The same person who pays, attends, works and sincerely wants to change is also the person who forgets appointments, changes the subject as we approach something, and passionately defends the very patterns that hurt them. This isn't hypocrisy or lack of willpower. Old patterns, however costly, were once solutions, and part of us parts with them as reluctantly as we part with anything that kept us alive. Therapy that recognises this doesn't rush to defeat the resistance. It respects it as the half-truth it is.

Because it's there, live rather than theoretically, that the central question of ambivalence gets tested: can a relationship survive my anger? Winnicott, in one of his last major papers (1969), described exactly this: the object, the other, becomes real only by surviving our "destruction". When a therapist receives anger without collapsing, without retaliating and without withdrawing, the person experiences something they may never have experienced before: a relationship that doesn't need protecting from their true feelings.

Psychodynamic psychotherapy doesn't aim to "clean up" feelings, removing the anger and leaving only the love. It aims at a humbler and more realistic change: enlarging the internal space so both can fit. In practice, that means working with transference reactions, with the history that taught each person anger is dangerous, and with the gradual, often slow construction of the object constancy that never got built earlier. The relationship with the therapist becomes the workshop. The rest of life becomes the field of application.

In plainer words

Getting angry with people we love doesn't mean our love is fake or that we're defective. It means the relationship is real. Psychoanalysis calls this coexistence of opposing feelings towards the same person "ambivalence", and treats it not as a problem but as a sign of mature psychological functioning.

As children, we divided the world into good and bad because we couldn't yet bear complexity. Growing up asks us to discover that the person who delights us and the person who wounds us are often the same, and that we can hold them whole. Those who learnt as children that anger is punished by the withdrawal of love find this harder. Psychotherapy is one of the places where that capacity can be built again, inside a relationship that proves in practice that it holds.

A closing thought

Perhaps the most consoling conclusion here is this: the guilt we feel when we're angry with those we love testifies, in its own distorted way, to how precious the bond is. We don't get angry like that with people who don't matter to us. The work isn't to erase the anger, nor to feel guilty about the love that accompanies it. It's to teach them to live in the same house.

Anyone who recognises themselves in these lines, in the cycles of idealisation, in the silent guilt of grief, in the fear that one conflict will destroy everything, might hold onto this: the capacity to love whole people, with everything they give us and everything they withhold, isn't a gift granted to a lucky few but a capacity that gets built, at any age, inside relationships that hold. Of all the findings of psychoanalytic thought, that may be the most hopeful.

Theoretical sources

For anyone who would like to read further, the main sources this piece draws on, alphabetically:

Eugen Bleuler, who introduced the term "ambivalence" (Ambivalenz) in 1910, initially in the context of his work on the psychoses.

Sigmund Freud, Totem and Taboo (1913) and Mourning and Melancholia (1917), where ambivalence emerges as a fundamental dimension of every bond and of grief.

Otto Kernberg, for his account of splitting and the idealisation-devaluation cycle in borderline personality organisation.

Melanie Klein, "A Contribution to the Psychogenesis of Manic-Depressive States" (1935) and "Notes on Some Schizoid Mechanisms" (1946), the foundational papers on splitting, the paranoid-schizoid position and the depressive position.

Margaret Mahler, for the concept of object constancy within the separation-individuation process.

Donald Winnicott, "Hate in the Countertransference" (1949) and "The Use of an Object" (1969), on hatred within care and the survival of the object.

Contemporary neuroscientific research on emotion regulation and conflict processing, with emphasis on the role of the prefrontal cortex and anterior cingulate.