Contribution
The first attempt to turn Danieli’s framework, Trauma and the Continuity of Self (TCMI), set out in her 1998 handbook, into a testable model. It puts in one multivariate model what earlier studies examined one at a time: the survivor’s age at the time, what they went through, where and how they resettled, and the family they built after. It asks how much of the child’s state runs through the parent’s way of living after the trauma, and how much runs past it.
Key insight
What passes to the child is the parent’s way of living after the trauma, not the trauma itself. In the framework’s words, “the parent’s fixity becomes the child’s biopsychosocial milieu” (p. 640). The data give that claim a shape. Internment matters to the child only through the parent’s victim style. The post-war family matters more than the war. And the one factor that reaches the child directly is not an injury but a broken link: the child’s sense that the murdered grandparents are not their own.
Assumptions
- TCMI. Identity spans many spheres: biological, intrapsychic, familial, communal, cultural, national. Trauma ruptures their flow through time, a state Danieli calls fixity. The conspiracy of silence, “societal indifference to, avoidance, and denial” of survivors’ experiences (p. 640), deepens it.
- Survival strategies become styles. In the extreme they “generalize to a way of life” (p. 640) and shape how survivors raise children.
- The child is a good enough witness. Parents’ styles, the child’s impacts and the family history are all reported by the child. The authors cite Yehuda et al. (2006) for children as reliable informants on parental PTSD.
- The child’s state is reparative. The outcome is named for “the core, perhaps unconscious, motivation of the second generation to undo and repair the past and heal their parents and themselves” (pp. 640–641).
- Missing data are missing at random. Part 3 completers and non-completers did not differ on any parent or child measure, which the authors take to justify mean substitution (p. 644).
Key results
- The sample (p. 642).
- The website had 7,222 visitors; 2,809 viewed the survey and 789 completed Part 1. The analysis kept 422 adult children with complete Parts 1 and 2 and some of Part 3. Grandchildren were excluded.
- 72% daughters, 78% North American and 22% Israeli, mean age 60 (SD = 8.2, range 38–84), 55% with graduate degrees.
- Anonymous, collected 16 June to 24 December 2012, in English and Hebrew, through survivors’, offspring and general Jewish organizations and word of mouth.
- The measures (pp. 642–643), from the three-part Danieli Inventory.
The style scales were derived by exploratory factor analysis of the
survey’s own Part 1 data.
- Victim style, 30 items (α = .92–.93): stuck in the loss and trauma rupture, overprotectiveness, emotional volatility.
- Numb style, 18 items (α = .89): emotional barrenness, conspiracy of silence in the family, intolerance of weakness.
- Fighter style, 12 items (α = .69–.70): valuing mastery and justice, valuing and maintaining Jewish (group) identity.
- Reparative adaptational impacts, 36 items (α = .91–.92), six subscales: insecurity about one’s competence, reparative protectiveness, need for power or control, obsession with the Holocaust, defensive psychosocial constriction, immature dependency. Linked to diagnoses of major depression, PTSD and generalized anxiety at ds = 0.48–0.89, in an earlier paper.
- Broken generational linkages: the mean of two items, “I rarely think of my parent’s murdered parents as my own grandparents” and remembering the family’s Holocaust experiences “only in bits and pieces” (p. 643).
- Descriptives (pp. 643–644). Impacts averaged 3.00 (SD .66) on a 1–5 scale, spanning nearly the whole range. Mothers’ victim, numb and fighter styles averaged 2.85, 2.79 and 3.58; fathers’ 2.68, 2.87 and 3.55. Of those with data, 84.7% had two survivor parents, and 47.4% of mothers and 45.5% of fathers had been interned without hiding or escape.
- Phase 1: what predicts the parents’ styles (pp. 644–646; Tables 3
and 4).
- Family history explained little: 6%, 4% and 5% of mothers’ victim, numb and fighter styles; 3%, 1% (ns) and 3% (ns) of fathers’. Internment without hiding or escape went with stronger victim style (mothers β = .22, fathers .28). Adult survivor mothers scored higher on all three styles than child survivors (victim β = .18).
- Family milieu explained more: 16%, 11% and 25% for mothers; 12%, 11% and 19% for fathers. Settling in Israel went with weaker victim style (β = −.22 for each parent) and much weaker fighter style (−.43, −.37). Broken generational linkages went with stronger victim (.17, .16) and numb style (.25, .25).
- Phase 1: what predicts the child (p. 646).
- Without the parents’ styles, Holocaust experiences (R²Δ = .026), setting and status (.009) and family configuration (.003) had no effect on the child. Social support (.054, p < .05) and continuity (.055, p < .001) did; broken linkages β = .22.
- On the six parent styles: mothers’ and fathers’ victim styles, β = .31 and .33; mothers’ numb style .15 (p < .05), fathers’ .04 (ns); fighter styles .10 and −.07 (ns). Daughters reported more impacts than sons. R² = 52%.
- Phase 2: the path model (pp. 646–647; Table 5). Numb and fighter
styles were dropped, as uneven or negligible.
- All-indirect model: χ²(22) = 31.30, p = .09, RMSEA = .032, CFI = .988, with one significant residual, between broken linkages and the child (z = 2.43). Freeing that path: χ²(21) = 25.01, p = .247, RMSEA = .021, CFI = .995. R² = 15% and 13% for mothers’ and fathers’ victim styles, 51% for the child.
- Victim style to child: mothers .38, fathers .34.
- Indirect effects of internment without hiding or escape: mothers’ .05, fathers’ .06. Mother being an adult survivor: .00.
- Total effects of milieu: Israel −.16, surviving maternal grandmother −.11, SES −.07, family size −.07, religious continuity −.07. Mother’s survivor-group affiliation −.02 (ns).
- Broken linkages: direct .09, indirect .12, total .21, the largest of any independent variable.
- The double dose (pp. 646–647). Having two survivor parents had no direct or indirect effect on the child. Each parent’s victim style had an independent, additive one. The authors read the double dose as psychosocial, not a matter of survivor status.
- The authors’ conclusion (p. 646). The study “demonstrated empirically” that survivors’ experiences during and after the Holocaust affect their children. Their limitations section grants that the effects are “necessary but not sufficient for establishing causation” (p. 648).
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Survivors’ wartime experiences, chiefly internment, relate to their children’s reparative impacts only through the parents’ victim style | weak to moderate: a well-fitting path model, but correlational, one informant, and indirect effects of .05 to .06 | Phase 2, Table 5 |
| C2 | The post-war family milieu weighs more on the parents’ styles and the child than wartime history does | moderate within this sample: variance explained 11–25% against 1–6% | Tables 3 and 4 |
| C3 | Broken generational linkages are the strongest risk to the child, acting directly as well as through the parents | weak to moderate: a two-item measure from the same informant, and the direct path added after the first model’s residual | Phase 2 |
| C4 | Settling in Israel protected survivors and so their children | weak: one dichotomy, confounded with everything else about the two countries; the reason given is drawn from other authors | Table 4, Table 5, Discussion |
| C5 | The double-dose effect is the sum of two parents’ victim styles, not survivor status | moderate within the sample | Phase 1, Phase 2 |
| C6 | Survivors’ experiences during and after the Holocaust affect their children | weak: no comparison group of non-survivor families, so the design cannot show that survivors’ children differ from others | Discussion |
Concepts
- TCMI (Trauma and the Continuity of Self: A Multidimensional, Multidisciplinary, Integrative Framework): Danieli’s 1998 account of identity as many spheres continuous through time, and of trauma as a rupture in that continuity.
- fixity: being “stuck” at the rupture; its severity depends on the trauma’s timing, duration, extent and meaning, and on later victimization.
- conspiracy of silence: survivors’ response to society’s indifference, avoidance and denial; here also a factor of numb style inside the family.
- posttrauma adaptational styles: victim, numb and fighter, survival strategies generalized to a way of life, as the child sees them.
- reparative adaptational impacts: the child’s state, read as a drive to repair the past and heal the parents and oneself.
- broken generational linkages: the child’s distance from the murdered generation, in identity and in memory.
- continuity: links that keep identity whole across the rupture, measured as broken linkages and shared religious practice across four generations.
Connections
- Fridman (NOTE-1291). The two designs answer different
questions.
- Fridman compares survivors’ daughters with others and finds no difference. Danieli et al. have no comparison group, so they say nothing about whether survivors’ children differ from others. They show which survivors’ children report more. The authors do not cite the meta-analyses Fridman restates (LIT-1718, LIT-1770); they cite Fridman et al. 2011 only on age cohort.
- Fridman’s sample is Israeli, and here Israel is protective (total −.16). This is an inference: the null may be partly an Israeli result.
- On silence they diverge. Fridman’s Proposition 8 holds that survivors setting the trauma aside may protect their children. Here numb style, which includes the conspiracy of silence in the family, had a modest effect through mothers only (β = .15) and none through fathers. The stronger channel was the parent stuck in the loss.
- A group’s story buffers, the third time. Costa et al.’s “grateful nation” (NOTE-1255), Fridman’s state-building, and here Solomon’s remark, quoted by the authors, that in Israel the Holocaust was the “legacy of all” (p. 647). This is the first time the record has an association for it, but it is a country dichotomy, not a measure of meaning.
- Volkan (NOTE-1250). His depositing of a trauma in children by parents is this paper’s channel too: children absorbed the Holocaust through “osmosis” (p. 648). Fighter style includes keeping up Jewish group identity, and it had no effect on the child.
- Eyerman (NOTE-1246). Cultural trauma passes by category. This sample is defined by descent and recruited through survivor organizations, so it cannot separate the category from the family.
- Hirst, Cyr and Merck (NOTE-1305). There, knowing a parent’s memory went with stronger identification. Here, half the broken-linkage measure is remembering the family’s history only in fragments, and that goes with more impacts.
- The germline (Horsthemke, NOTE-1265). The authors propose adding genotype and methylation data to the inventory (p. 648). That would join the routes in one sample, not separate them.
Bearing on the record
- The intergenerational trauma question (QUESTION-297). The
paper bears on three of its routes. What follows are inferences; the
authors do not frame their results as routes.
- Parenting. This is the clinical tradition’s own quantitative test of the route, and it places the route within survivor families. What reaches the child runs through how the parent lived after (victim style), not through what the parent went through. With Fridman, both can hold: no general harm to survivors’ children, and variation among them that tracks the parent’s way of living. But nothing here shows harm against a baseline, so it cannot answer Fridman’s null.
- Silence. The question calls this route “parenting and silence”. On this evidence silence is the weaker part. Two works now fail to make it the main channel: Fridman, who argues that setting the trauma aside may have protected the children, and this paper, where numb style, which includes silence, had a mother-only effect and did not enter the final model.
- Identity. The question asks whether an ancestor’s injury becomes part of the descendant’s self. “I rarely think of my parent’s murdered parents as my own grandparents” is the nearest the record has to a measure of whether the ancestors are part of the descendant’s self. It is a family self, not a group’s. Descendants who count them as their own reported fewer impacts, and the effect was partly direct. Meanwhile the outcome scale counts “obsession with the Holocaust” as harm. So the question’s phrase covers two things that point opposite ways here: the ancestor becoming part of the self, which went with less harm, and the injury becoming part of it, which is part of the harm. The data are correlational and the direction is open: a distressed child may also be one who has pulled away from the family past.
- The group. Fighter style’s group-identity component and mothers’ survivor-group affiliation had no effect on the child. Israel, where the trauma was the nation’s, protected. On this sample, a shared group story did not carry harm to the child, and may have buffered it.
- The owner’s thesis (CLAIM-491). Indirectly. The outcome is named for a child’s drive to “heal their parents and themselves”, and its subscales include reparative protectiveness. The clinicians score a child’s care for a parent as part of the child’s own condition. That fits a self extended to the family, Brewer’s relational self (NOTE-1290), but the paper does not separate the child’s caring from any extension of the self, and is not built to.
- The triviality objection (CLAIM-492). The grandparents item asks whether the dead are counted “as my own”, not how much one cares about them. It is a candidate criterion of the self’s extent backward in time. But it is averaged with a memory item and never set against a measure of caring, so it shows nothing yet.
- Batson’s objection (CLAIM-494). Not addressed.
- A dated note is proposed for QUESTION-297, not for the claim (section 4).
- No THEORY filed.
- No ML instruction.
Limitations
- No comparison group. The design cannot show that survivors’ children differ from anyone’s (p. 648).
- One informant for everything. The child rated the parents’ styles, their own impacts and the family history, retrospectively. A distressed child may see their parents as more victim-like, which the authors grant could inflate the links (p. 648).
- Convenience sample. Volunteers reached through survivor and Jewish organizations, highly educated, mostly daughters.
- Attrition and imputation. Of 789 who completed Part 1, 422 were analysed. Missing values were replaced by the mean; among them, the nature of the Holocaust experience was unknown for 32.0% of mothers and 34.8% of fathers, and displaced-person camp for 45.0% and 47.4% (Table 2).
- Measures built and tested on the same data. The style scales came from factor analysis of this survey’s Part 1. Fighter style’s α was .69–.70.
- Model building. Variables were kept or dropped by significance in Phase 1, and the direct path for broken linkages was added after the first model’s residual.
- The conclusion outruns the method. “Demonstrated empirically” (p. 646) sits beside the limitations’ “necessary but not sufficient for establishing causation” (p. 648).
- Not independent of the handbook. The first author edited the handbook whose framework is tested; the third reviewed the handbook for ISTSS.
- Typos.
- The in-text citation “Fridman, Fridman, Bakermans-Kranenburg, Sagi-Schwartz, & van IJzendoorn, 2011” (p. 641) doubles the first author; the reference list has one.
- Religious continuity is coded “0 = no continuity, 1 = some continuity, and 3 = complete continuity” (p. 643); 2 is probably meant.
- Cuthbert and Insel (2013) is given the DOI of Cuthbert (2014), 10.1002/wps.20087 (p. 649).
- “Mahway, NJ” for Mahwah (p. 650).
- Table 2 gives no n for fathers’ number of places before resettlement (p. 644).
Open questions
- Do the two broken-linkage items act alike? Counting the murdered as one’s own and remembering their history are different things, and only the first bears on the self.
- Does the linkage effect hold with a second informant, or over time? If linkage predicts later impacts, the self-inclusion reading gains; if impacts predict later withdrawal, it loses.
- Would non-survivor families of the same origin show the same pattern? Without them, the paper’s within-group result cannot be set against Fridman’s between-group null.