The stages of grief, with the evidence put back in.
The five stages were never meant to describe the bereaved — and 50 years of research has replaced them with something more honest, and considerably more humane.
The short version
The five stages of grief — denial, anger, bargaining, depression, acceptance — were proposed in 1969 by Elisabeth Kübler-Ross to describe how terminally ill patients face their own deaths, not how survivors grieve a loss. The application to the bereaved came later, mostly through cultural transmission, and the empirical record has not been kind to the strict version of the model. What has held up: grief oscillates, most people are more resilient than we give them credit for, ongoing internal relationships with the dead are healthy rather than pathological, and a small but real share of bereaved adults — about one in ten — experience a longer, more identity-disrupting grief that benefits from professional support. Held lightly, the stage vocabulary still names emotions that show up. Held tightly, it tells most people they are doing it wrong.
What Kübler-Ross actually wrote
Most of what is repeated about Elisabeth Kübler-Ross's On Death and Dying is not in the book. The 1969 monograph drew on her observations as a Swiss-American psychiatrist working with terminally ill people at the University of Chicago hospitals, and it described five emotional responses she saw in the dying, not in their families[1]. The five — denial, anger, bargaining, depression, acceptance — were sketches, not a sequence; she explicitly cautioned against reading them as a fixed staircase, and the original text is full of patients who skip stages, return to earlier ones, or settle into one and never leave.
The migration of the model from dying to grieving happened almost entirely outside the academic literature. Hospice training, popular journalism, talk-show segments, eventually grief support groups — the framework had a clean five-bullet shape that was easy to remember and easy to repeat, and it filled a real cultural gap. Mid-twentieth-century Western life had little public vocabulary for loss. The five stages gave bereaved people something to point at.
Kübler-Ross herself spent the last decades of her career trying to undo the rigid, stepwise reading. In her final book, co-written with David Kessler and published the year after her death, she wrote that the stages were “never meant to help tuck messy emotions into neat packages” — and that they were “responses to loss that many people have, but there is not a typical response to loss as there is no typical loss.” The phrasing rarely makes it into the citations.
What the empirical record actually shows
The largest direct empirical test of stage theory is the Yale Bereavement Study, published in JAMA in 2007 by Paul Maciejewski and colleagues[2]. The researchers followed 233 bereaved adults from one to twenty-four months after a loss and tracked which of the five emotional states dominated at each interval.
Three findings, taken seriously, change the picture.
First, acceptance was the most-endorsed response from the very first measurement. Not at the end, where the stage model puts it. From the start. Disbelief, yearning, anger, and depression all peaked at expected points and declined — but acceptance was higher than any of them at every point in the study, including the first month after the loss. The implication is uncomfortable for the popular model: most bereaved adults arrive at some form of acceptance early, even when they are also feeling everything else.
Second, yearning — not depression — was the most pronounced negative emotion. The popular version of the model has “depression” as the heaviest stage. The data put yearning — the longing for the person who is gone — at the top, peaking around four months. Depression peaked later and at lower intensity. This matters clinically: yearning and depression respond to different kinds of support, and conflating them tends to push grieving adults toward antidepressant prescribing when what they actually need is space for the longing.
Third, the order varied. Many participants did pass through something resembling the five emotional states — but rarely in the predicted sequence, and rarely without skipping or revisiting. By the strict reading of stage theory, almost no one was “doing it right.” By any reasonable reading of human experience, almost everyone was.
The five stages survived their own evidence. They are now better understood as a vocabulary than as a map.
What does fit the data: the dual process model
The framework that has largely replaced strict stage theory in the academic literature is the dual process model of coping with bereavement, proposed by the Dutch psychologists Margaret Stroebe and Henk Schut in 1999[3]. Its central observation is straightforward: grieving is not a single emotional task but two, and the bereaved person oscillates between them.
Loss-orientation is the work of confronting the loss directly — weeping, looking at photographs, replaying the last conversation, missing the person, being with the absence. Restoration-orientation is the work of reorganizing a life around the absence — doing the laundry, returning to work, learning new roles the deceased used to fill, taking on new identities. The dual process model insists that healthy grieving requires both, and that the bereaved person moves between them, often within a single hour.
Two practical consequences fall out of this. The first is that an ordinary morning two weeks after a death is not denial; it is restoration-orientation, and it is one of the modes the model predicts. The second is that grief that gets stuck almost always gets stuck in one orientation, not in a stage. People who can only feel the loss never restore; people who only restore never let the loss in. The clinical signal is which side a person can't visit, not which stage they are on.
Subsequent reviews have generally supported the dual process model's descriptive accuracy, particularly its prediction that flexible oscillation correlates with better long-term outcomes than either pure rumination or pure avoidance.
Bonanno's resilience finding
The most uncomfortable thing the modern bereavement literature has to say is also the most well-supported. Most bereaved adults are resilient. Not stoic, not detached, not in denial — resilient, in the technical sense of maintaining stable functioning through and after the loss while still feeling the loss.
The psychologist George Bonanno spent two decades documenting this in longitudinal studies of widows, widowers, parents, and people exposed to mass-trauma events. His 2009 book The Other Side of Sadness is the accessible synthesis[4]; the underlying papers are in Psychological Science, Journal of Personality and Social Psychology, and American Psychologist across the early 2000s. Across studies, he and his collaborators found four characteristic trajectories after loss: resilience, the most common, with stable low-distress functioning; recovery, an initial period of distress that resolves over one to two years; chronic grief, a smaller group with lasting high distress; and delayed grief, a contested pattern where distress emerges later. Across most studies of common bereavement, roughly half of adults fall into the resilient trajectory.
This is not a finding that gets repeated much in popular grief writing. It is uncomfortable in two directions. To bereaved people who are not feeling devastated, it suggests that nothing is wrong — which can read as permission they didn't feel they had asked for. To bereaved people who are devastated, it can sound like a verdict. Both readings are mistakes. The data are descriptive, not prescriptive: they describe a population, not what any individual ought to feel. They simply mean that resilience after loss is not pathological numbness, and that capacity for ordinary functioning alongside grief is the rule rather than the exception.
For the late-night and early-morning hours of grief — the time before the world is awake and after the people are gone — Marcus is the Mindflex companion built for the existential weight of an ordinary loss. Not a substitute for grief work. A place to think out loud while the rest of the house is sleeping.
Continuing bonds: the finding that quietly changed clinical practice
Through most of the twentieth century, the dominant Western model of grief work assumed that healthy grieving required severance — a gradual emotional decathexis from the lost relationship, in Freud's original framing, after which the survivor's libido was free to attach elsewhere. The hospice and grief-counseling field absorbed this: ongoing inner conversations with the deceased were treated as evidence of incomplete grief work.
The 1996 anthology Continuing Bonds: New Understandings of Grief, edited by Dennis Klass, Phyllis Silverman, and Steven Nickman, replaced that picture with the opposite finding[5]. Across cross-cultural studies of bereaved children, widows, and parents, the editors documented that bereaved people routinely — and often for the rest of their lives — maintained ongoing internal relationships with the person who died. They talked to them. They asked what they would say. They consulted them on hard decisions. They kept their habits. None of this correlated with poor outcomes; in many populations, the opposite.
The reframing has held up. Subsequent research has refined the picture — some forms of continuing bonds (clinging, intrusive memory) correlate with worse outcomes; most forms (consultative, narrative, identity-integrative) correlate with better — but the broad finding stands. The bereaved person who still talks to the person who is gone is not stuck. They are doing what bereaved people, in most of human cultures and across most of recorded history, have always done.
When grief becomes “stuck”
About one in ten bereaved adults will experience a grief that is fundamentally different from the resilient and recovery trajectories — one that does not move on its own, that remains identity-disrupting, that interferes with most areas of life for many months. Beginning in 2022, the formal nomenclature for this in both the American and World Health Organization frameworks is Prolonged Grief Disorder[6].
The criteria, in plain English: persistent intense yearning or preoccupation with the deceased; identity disruption (“part of me died with them”); marked avoidance of reminders, or its opposite, complete inability to think of anything else; emotional numbness; difficulty re-engaging with life; the conviction that life is meaningless; intense loneliness specifically tied to the loss. The duration thresholds are at least 12 months in the DSM-5-TR and at least 6 months in the ICD-11. Clinical estimates put prevalence at around 9–10 percent of bereaved adults, with higher rates after sudden, violent, or stigmatized losses.
Prolonged grief is not a moral failure to grieve correctly. It is the empirical finding that some bereavements, for some people, in some circumstances, exceed the capacity of unaided grieving — and that targeted, evidence-supported approaches (in particular, prolonged-grief-focused cognitive-behavioral protocols developed by Katherine Shear and colleagues) help most of the people who try them. If grief still owns most of someone's life six to twelve months in, that is information, not failure.
What actually helps, by the literature
None of the following is a stage. They are practices that recur in the bereavement literature and in clinical experience as helpful for most people, most of the time.
1. Allow oscillation. If you are crying in the morning and laughing by lunch, you are not betraying anyone. The dual process model would call that exactly what is supposed to be happening.
2. Let the bond continue if it wants to. Talking to the person who is gone — in your head, in a notebook, at the grave, in the kitchen — is not a sign that you are stuck. The continuing-bonds research is unusually clear on this point.
3. Resist the urge to grade the grief. “I should be over this” and “I shouldn't be enjoying this lunch” are the same mistake. Both apply a stage-theory ruler to a process that does not have stages.
4. Watch for the markers, not the calendar. The question is not how many months. It is whether the grief is moving, whether you can still inhabit the parts of your life that matter to you, and whether, six months in, life feels at all possible. If the answer to those is no, that is the moment to talk to a grief counselor — someone trained specifically in bereavement, ideally familiar with prolonged-grief-focused approaches.
5. Take the body seriously. Sleep, food, water, and movement are not solutions to grief, but they are the difference between grief and grief-plus-physiological-collapse. The research on bereavement is clear that the second is much harder to come back from than the first.
6. Be skeptical of the timeline. Most cultural scripts for grief vastly underestimate how long meaningful loss takes. A year is a beginning, not an end. Two years is normal. Five years for a major loss is not unusual. None of this means anything is wrong.
Where Mindflex fits
A grief counselor is the person who walks the bereavement specifically with you. A friend or family member is the person who knew the one you lost. A grief group is the room of people who recognize the shape of what you are carrying. A reflection companion is something else entirely — the private space at 3 a.m., when the world is asleep and the mind is replaying, and the question is what you are trying to think and whether you can make any sense of it tonight.
Mindflex is a reflection companion. An AI developed by clinical psychologists in Berlin, for the internal conversation that comes before the external ones. For grief: the hours that are too quiet for a phone call, the months when the visible mourning has stopped and the inside hasn't, the small daily reckonings that don't fit anywhere else. Not a grief counselor. Not a medical device. Not a crisis service. Not the 2009 Mattel Mindflex brain-controlled levitation toy.
No account to start. iOS (Android coming).
Questions people actually ask
Why do I not feel anything?
Numbness is one of the most common early responses to a major loss, and it is not a sign that you didn't love the person. The nervous system tends to throttle emotional bandwidth in the first days and weeks of a major loss; the absence of feeling is often the body protecting the system from being overwhelmed. The numbness usually starts to lift on its own as the immediate shock recedes. If it persists past a few months and is accompanied by inability to engage with anything in your life, that is a marker worth taking to a clinician.
Is laughing soon after a death disrespectful?
No. The dual process model would say it is restoration-orientation doing exactly what it is supposed to do — giving the system a break from loss-orientation so that loss-orientation can return when the body has the bandwidth. Many cultures explicitly build laughter into funerals for this reason. The instinct that laughter betrays the loss is mostly a cultural script, not a psychological reality.
What if I didn't have a good relationship with the person who died?
Complicated relationships often produce complicated grief, and the literature treats this seriously. Loss of a difficult parent, an estranged sibling, a partner with whom things ended badly — these can feel as heavy as “clean” losses, and sometimes heavier, because grief now includes the loss of any chance for repair. The practical version is the same: name the actual feeling (which may be relief, anger, regret, or some impossible mix) without grading it. Ambivalent grief is grief.
How long is grief supposed to last?
Far longer than most cultural scripts suggest. The acute phase — in which grief is the dominant feature of daily life — commonly lasts six months to two years for a major loss. The integrated phase — in which the loss is folded into a continuing life — lasts the rest of the life. That is not pathology. The bereaved person who still misses someone twenty years on is not stuck. They loved someone, and the love did not stop because the person did.
Should I avoid reminders or face them directly?
Both, in oscillation, is the answer the research consistently gives. People who can only avoid reminders tend to carry the grief longer. People who can only face them tend to deplete themselves. The clinical aim is flexibility: the capacity to look at the photograph today and put it away tomorrow, both intentionally. If you find yourself pinned to one mode, that is the signal to vary it deliberately.
What if I am grieving someone who is still alive?
Anticipatory grief (for a dying loved one), ambiguous loss (estrangement, dementia, deportation), and loss of relationship without death are all recognized in the bereavement literature, and most of the same patterns — oscillation, continuing bonds, the slow integration — apply. The work of the psychologist Pauline Boss on ambiguous loss is the standard reference here. Grief without a clear endpoint is harder in some specific ways — the absence of ritual, the social uncertainty about whether you are entitled to grieve — and worth treating with the same care.
How do I know if my grief needs professional help?
The clearest markers are these: the grief is not moving at all six to twelve months in; you cannot inhabit the parts of your life that mattered before; you have persistent thoughts of joining the deceased; you are using substances to manage the grief; or at any point, you are having active thoughts of self-harm. Any one of these is reason to talk to a grief counselor or a primary-care clinician familiar with bereavement, not a sign of weakness. The 988 line in the U.S. and the Telefonseelsorge in Germany handle this exact conversation thousands of times a day.
References
- Kübler-Ross, E. (1969). On Death and Dying. New York: Macmillan. (50th-anniversary edition: Scribner / Routledge.) The original five-stage proposal — described as responses observed in terminally ill patients, not in the bereaved.
- Maciejewski, P. K., Zhang, B., Block, S. D., & Prigerson, H. G. (2007). An empirical examination of the stage theory of grief. JAMA, 297(7), 716–723. DOI — the Yale Bereavement Study.
- Stroebe, M., & Schut, H. (1999). The dual process model of coping with bereavement: rationale and description. Death Studies, 23(3), 197–224. DOI
- Bonanno, G. A. (2009). The Other Side of Sadness: What the New Science of Bereavement Tells Us About Life After Loss. New York: Basic Books. Synthesis of the resilience-trajectory research; the underlying papers are scattered across Psychological Science, Journal of Personality and Social Psychology, and American Psychologist (2001–2008).
- Klass, D., Silverman, P. R., & Nickman, S. L. (Eds.). (1996). Continuing Bonds: New Understandings of Grief. Washington, DC: Taylor & Francis. The volume that displaced the “detachment” model of grief work.
- American Psychiatric Association (2022). DSM-5-TR: Prolonged Grief Disorder; and World Health Organization (2022). ICD-11: Prolonged Grief Disorder (6B42). Both nosologies added the formal category in 2022. ICD-11 entry.