False Allegations of Incest
Let me approach this discussion with the basic premise that all mental health professionals have the child’s best interests as their primary consideration. All mental health professionals wish to protect children from sexual abuse. No mental health professional seeks to expose a child to sexual abuse victimization.
If there are differing viewpoints on how to accomplish this goal, they are good-faith expressions of “equally valid poles of a dialectic” (Linehan, 1993), and the goal should be to move toward synthesis of differing perspectives rather than adopting adversarial positions.
Authentic sexual abuse occurs far too frequently. We must act to protect children from sexual abuse victimization.
And… on occasion,
A narcissistic/(borderline) parent will induce or elicit false allegations from the child of sexual abuse victimization by the other parent as a means to exploit these child allegations to achieve power over the situation and the other parent. When this occurs, not only does the child lose a loving and affectionally bonded relationship with a normal-range targeted parent, the child also loses the potential protective influence that the normal-range psychological organization of the targeted-rejected parent can have in lessening the distorting pathogenic influence of the narcissistic/(borderline) parent on the child’s development.
Furthermore, for a narcissistic/(borderline) parent to induce or elicit from a child false allegations of sexual abuse against the other parent represents extremely distorted parenting that rises to the level of severe psychological abuse of the child. Failure to respond to this type of psychological child abuse when it is present is to abandon the child to the severely distorting effects of the psychological child abuse of the narcissistic/(borderline) parent that will have a long-term destructive impact on the child’s psychological development, likely influencing future generations of the family as well through the transmission of the effects of the child abuse to the next generation through the future pathogenic parenting of the current child with his or her own children.
Our goal should be to protect the child from ALL forms of child abuse, particularly the severe forms being considered in this discussion. It is NEVER acceptable to abandon a child to any form of child abuse.
Assertion 1: Child Sexual Abuse Allegations are Not Developmentally Normal
Under no circumstances does a child ever spontaneously develop a false belief that a parent sexually abused the child. It doesn’t happen. Ever.
Children may develop on their own symptoms of hyperactivity or inattention, or anger control problems, or cognitive delays, or social problems, or phobic anxieties. All of these types of symptoms can sometimes arise endogenously to the child’s own developmental course.
Children NEVER spontaneously develop, on their own, a false belief that a parent sexually abused the child.
When a child asserts that a parent sexually abused the child, there are only three differential diagnostic possibilities:
1) Authentic child abuse incest by the parent
2) Extremely distorted pathogenic parenting by a narcissistic/borderline parent in which this parent induces or elicits the child’s false belief of sexual abuse,
A) In order to exploit the child’s symptoms to achieve power over the situation or other parent, or
B) As a result of semi-psychotic decompensation of the narcissistic/borderline parent into a delusional belief in the (false) threat represented by the other parent, with the parental delusional belief then being transferred to the child through severely distorted parenting practices of the narcissistic/borderline parent.
“When particular schemas are hypervalent, the threshold for activation of the constituent schemas is low: they are readily triggered by a remote or trivial stimulus. They are also “prepotent”; that is, they readily supersede more appropriate schemas or configurations in processing information. (Beck et al., 2004, p. 28)
“The conceptualization of the core pathology of BPD as stemming from a highly frightened, abused child who is left alone in a malevolent world, longing for safety and help but distrustful because of fear of further abuse and abandonment, is highly related to the model developed by Young… that some pathological states of patients with BPD are a sort of regression into intense emotional states experienced as a child. Young conceptualized such states as schema modes…” (Beck et al., 2004, p. 199)
3) The child, typically an adolescent, lacks a moral conscience, typically as a result of extraordinarily poor parenting, and is consciously, intentionally, and independently using a false allegation of sexual abuse against a parent as an intentional manipulation to achieve a desired goal or outcome.
This is not a spontaneous development of a false belief, this is simply a conscious lie perpetrated by the adolescent for manipulative gain.
Exclusionary caveat: The adolescent’s actions are not the product of influence from an allied and supposedly “favored” parent to allow the parent (or parent-child alliance) to then exploit the child’s allegations for manipulative gain (i.e., Causal Origin 2).
In my clinical practice I have known adolescents who admitted to me that they colluded with peers on how to create marks on their body to substantiate false allegations they made about physical abuse from a parent. In one of these cases the motivation of the adolescent was to obtain a custody change from the current parent to the desired parent. However, it wasn’t so much that the desired parent was in a cross-generational coalition with the child as much as the desired parent was exceedingly lax and permissive as a parent, which was a parenting style preferred by this adolescent over the current parent’s more structured expectations. Other adolescents in my clinical practice have threatened their parents with filing false child abuse allegations against the parent if the parent did not capitulate to the child’s demands in some area. These families had a history of highly dysfunctional relationships.
In all three causal origins for sexual abuse allegations made by a child against a parent, there is extremely bad parenting somewhere within the family. Sexual abuse allegations made by a child against a parent are ALWAYS evidence of extremely bad parenting occurring somewhere within the family. The only question is where.
Personal Estimates of Prevalence
Note: There is no existing reliable data to support these estimates. These are personal estimates based solely on my clinical judgment.
From the domain consisting of all sexual abuse allegations made by a child of incest by a parent, my estimates of the prevalence for the three different origins for the child’s sexual abuse allegations against a parent are:
1. Authentic child abuse
Estimate of the likely prevalence of this causal origin for the child’s allegations: I would estimate that authentic sexual abuse of the child is the causal origin of the child’s allegations in approximately 95% of the cases of children’s alleged sexual abuse by a parent.
2. Extremely distorted pathogenic parenting by a narcissistic/borderline parent that induces or elicits from the child a false allegation of sexual abuse against the other parent
Estimate of the likely prevalence of this causal origin for the child’s allegations: I would estimate that pathogenic parenting by a narcissistic/borderline parent that induces or elicits from the child a false allegation of sexual abuse against the other parent is the causal origin of the child’s allegations in approximately 1% – 2% of the cases of children’s alleged sexual abuse by a parent.
3. An intentional adolescent lie as a manipulative or retaliatory action against the parent
Estimate of the likely prevalence of this causal origin for the child’s allegations: I would estimate that an independent intentional lie created by an adolescent as a manipulative or retaliatory action against the parent is the causal origin of the child’s allegations in approximately 3% to 5% of the cases of children’s alleged sexual abuse by a parent
In any individual case, the estimated population prevalence is not a relevant consideration.
For example, in the 1% or 2% of the cases in which the child’s allegation of sexual abuse incest against a parent is induced and elicited by the pathogenic parenting practices of a narcissistic/borderline parent, in those 1% – 2% of cases the likelihood that the child’s symptoms were induced or elicited by the pathogenic parenting practices of a narcissistic/borderline parent is 100% – in those cases.
Is the specific case under consideration one of these rare cases? Perhaps. And if this specific individual case is one of those rare cases, then the probability that the child’s allegations are the induced or elicited product of pathogenic parenting practices by a narcissistic/borderline parent is 100% (i.e., if this specific case is one of those 1% – 2% of cases).
So population prevalence is not a consideration in evaluating any specific case. Each case is individual.
This is a foundational construct in psychological testing. Just because the population prevalence of any specific issue, such as learning disabilities, ADHD, or mental retardation, is an infrequent occurrence in the general population doesn’t mean that this specific child doesn’t have the issue in question.
General population prevalence is not relevant to the assessment of any individual case. The assessment of each specific case is individualized to the specific data of that individual case.
There are technical considerations in establishing the criteria by which we make decisions. In some cases the data allows us to make decisions based on 100% certainty, but this is extremely rare in decisions about psychological issues. When 100% certainty is not available, there are two possible types of errors we can make in our decisions based on the data
When a child makes sexual abuse allegations against a parent, there are two possible hypotheses that can be supported or disconfirmed by the data,
“Null Hypothesis” – no sexual abuse occurred
“Clinical Hypothesis” – sexual abuse did occur
Type I Error: A Type I error is when we erroneously accept the Clinical Hypothesis that sexual abuse occurred when, in truth, there was NO sexual abuse of the child. This is called a “false positive” decisional error, when we erroneously say something took place when it actually didn’t occur.
Type II Error: A Type II error is when we erroneously accept the Null Hypothesis that NO sexual abuse occurred, when in actuality the child was sexually abused by the parent. This is called a “false negative” decisional error, when we say nothing took place when it actually did occur.
These two types of decisional errors are interrelated. The lower the probability of making one type of decisional error, the higher the probability of making the other type of decisional error.
In establishing the criteria for making a decision from data that does not allow 100% certainty, the question is which type of error is worse in the context of the issue under consideration. So if we are making a decision about whether a child has been sexually abused by a parent, is it worse to,
A) Erroneously conclude that the child was sexually abused when in actuality no sexual abuse took place (a Type 1 Error).
The consequences of this decisional error would be that we would needlessly and erroneously terminate the child’s relationship with a normal-range and affectionally available parent, and we would abandon the child to the custody of a narcissistic/borderline parent who is engaging in extremely distorted parenting practices that will severely distort the child’s emotional and psychological development (Causal Origin 2), or
We will terminate the child’s relationship with a parent in collusion with the child’s manipulative intent of attaining a desired goal or outcome (Causal Origin 3)
B) Erroneously conclude that the child was NOT sexually abused when in actuality the child was sexually abused by the parent (a Type II Error)
The consequence of this decisional error would be that we DO NOT terminate the child’s relationship with a sexually abusive incestuous parent, thereby abandoning the child to continued sexual abuse victimization.
One of the particularly devastating consequences of this decisional error with regard to child sexual abuse allegations against a parent is that we communicate to the child that we do not believe the child’s report is accurate when the child is, in actuality, telling us the truth. This is psychologically devastating to the sexually abused child. For the child to overcome the personal shame and family secrecy surrounding sexual abuse victimization and to come forward in disclosing the abuse, and then not to be believed and to be abandoned to their continued victimization, is psychologically devastating to the child, compounding the severity of the trauma for the child.
The decision as to which is worse, the impact of Type I errors or the impact of Type II errors, with regard to any particular issue is a value judgment regarding the comparative impacts. Based on this value judgment we then set decisional criteria that minimize either the likelihood of making a Type I decisional error (thereby increasing the likelihood of making a Type II decisional error), or we establish decisional criteria that minimize the likelihood of making a Type II error (thereby increasing the likelihood of making a Type I decisional error), or we seek decisional criteria that provide some sort of balance between the risks of making Type I and Type II errors in our decision making about the data.
But the criteria we establish for making our decisions is based on value judgements regarding the relative dangers of making a Type I decisional error (called a “false positive”) compared to the dangers associated with making a Type II decisional error (called a “false negative”).
Prosecution and Child Protection
The legal system in the United States has traditionally made a value judgment to minimize the potential for Type I errors (“false positive” decisions of convicting an innocent person). The decisional criteria in the legal system would prefer to set free numerous criminals (i.e., Type II errors of “false negatives”; saying the person is innocent when the person is actually guilty) in order to avoid (to the extent possible) convicting an innocent person (i.e., to minimize the risk of making a Type I error of a “false positive”).
This approach to establishing decisional criteria is based on our value judgments concerning our desired system of justice.
The impact of this approach within mental health and social service systems dealing with child allegations of parental sexual abuse, however, is troubling to many. The value judgment of the legal system to minimize Type I errors (“false positives” of saying there was child abuse when in actuality there was no child abuse) will correspondingly increase the probability of making Type II decisional errors (“false negatives” of saying there was no child sexual abuse when in truth there actually was child sexual abuse), meaning that many instances of actual child sexual abuse by a parent will not result in a proper child protection response as a result of our Type II decisional error.
Advocates for child protection rightly become extremely concerned about the rate and frequency of Type II decisional errors that leave children unprotected, even when, and especially when, the child discloses the sexual abuse. Parental sexual abuse of the child is surrounded by tremendous personal shame for the child, and is masked beneath the cover of family secrecy. For the child to break free from this family secrecy and personal shame associated with parental incest is a significant achievement for the child. But then for the child not to be believed and to be cast back into the abusive relationship with an incestuous parent compounds trauma upon trauma.
The psychological injury caused to the child by a Type II decisional error of saying there was no sexual abuse of the child when in truth there was, and when the child actually takes the momentous step of disclosing the sexual abuse to others, is extremely severe and devastating to the child. That child advocates within the mental health and family service fields are highly concerned about the frequency of Type II decisional errors is understandable and justified.
There are a certain percentage of cases of false child allegations of parental sexual abuse (i.e., Causal Origins 2 and 3).
Within the legal system, the wrongful conviction of a parent as a pedophile (i.e., a Type I decisional error of a “false positive” attribution), so that this parent is wrongly identified with the lifelong stigma as being a “sex offender” can have an extremely devastating impact on this parent. Given the heavy consequences for an innocent person of a wrongful conviction as being an incestous sexual pedophile, the legal system is justifiably reluctant to make Type I errors, and would prefer to allow some guilty people to go free rather than convict an innocent person. So the decisional criteria within the legal system are adjusted toward limiting the probability of making Type I decisional errors (i.e., a “false positive” attribution that convicts an innocent person).
Most people support this general legal philosophy of limiting to the extent feasible the probability of convicting innocent people for crimes they did not commit (i.e., of making Type I decisional errors of saying something happened when in truth it didn’t), and we are willing to adjust to the negative consequences associated with sometimes allowing the guilty to go free in order to minimize to the extent feasible the risk of convicting an innocent person.
Despite our best efforts, we nevertheless sometimes convict innocent people (i.e., make a Type I decisional error), and this is extremely distressing for both the legal system and the general public, and yet it is also unavoidable. The only way we can ensure with 100% certainty that we will NOT make a Type I decisional error of convicting an innocent person is to set our decisional criteria so far in favor of not making a Type I error that we make far too many Type II errors of allowing all, or nearly all, criminals go free rather than run even the barest of infinitesimal risks of possibly convicting an innocent person.
In our decisions we must balance the risks of making a Type I error (a “false positive”) against the risks of making a Type II error (a “false negative”). As we lessen the risk of one type of decisional error we increase the risk of making the other type of decisional error. That’s just the way it works.
While in general, we can appreciate the reluctance of the legal system to convict an innocent person, so that we accept and tolerate the release of burglars and thieves, and even murderers, when it comes to the sexual abuse of children we become disturbed by the possibility of exposing children to continued sexual abuse victimization in order to limit the risk of making a Type I decisional error of a “false positive” attribution that the parent is an incestuous pedophile when in truth no sexual abuse of the child occurred, especially when the likely population prevalence of “false positives” is so incredibly low (for example, my estimate of between 4% to 7% of all sexual abuse allegations made by children against a parent; Causal Origins 2 and 3).
While we are willing to tolerate the consequences of a “false negative” decision that results in the release of a thief or even a murderer, we are much more reluctant to tolerate the consequences of a “false negative” decision that results in the release of a parental pedophile, in which we will be re-exposing the child to continued sexual abuse predation by the parent.
Given the very low estimated population prevalence of “false positives” (i.e., parents who are accused of sexual child abuse when NO abuse occurred), if we wanted to eliminate the risks of making a Type II error (a “false negative” of saying there was no sexual abuse of the child when in fact the child had been sexually abused by the parent) because of the devastating effect of a Type II decisional error on the emotional and psychological well-being of the child, we could simply accept all child allegations of sexual abuse as valid. This would mean that we would likely make between 4% and 7% Type I errors of making a “false positive” attribution (i.e., saying there was sexual abuse of the child by the parent when, in truth, the parent did not actually sexually abuse the child). In accepting the consequences of making an estimated 5% to 10% (rounding off) wrongful convictions of innocent parents as being incestuous sexual predators when they are not, we can ensure that we do not re-expose ANY child to authentic sexual abuse victimization.
This would represent a value decision regarding the comparative negative consequences associated with making a Type I decisional error as compared to making a Type II decisional error. The decisional criteria we decide on regarding the acceptable probability of making Type I versus Type II decisional errors is based on our value judgments regarding the comparative damage of making each type of decisional error. The more we reduce the chances of making one type of decisional error, the more we increase the chances of making the other type of decisional error.
Since the estimated population prevalence of “”false positives” (i.e., false child allegations of parental sexual abuse) is so small relative to all child allegations of parental sexual abuse, it is conceivable that a reasonable discussion would consider the relative benefit of accepting all child allegations of parental sexual abuse as valid in order to eliminate the possibility of making Type II decisional errors of “false negative” decisions that continue a child’s exposure to an incestuous pedophiliac parent even after the child has disclosed the predatory sexual abuse victimization.
However, if this decision is made we should also recognize that we are accepting that there will be a certain, not insubstantial number of “false positive” Type I decisional errors in which we wrongly say that an actually innocent parent sexually abused his or her own child. Recognizing that we will be making such errors in a relatively substantial number of cases (my personal estimates are between 5% to 10% of all cases of child sexual abuse allegations), we would likely want to limit the damage to these wrongly convicted parents.
One approach to limiting the damage to innocent parents who are wrongly accused of sexual abuse that they did not actually commit is to separate the legal from the social service responses, so that the social service response to child allegations of parental sexual abuse would be to accept ALL child allegations of parental sexual abuse as valid, recognizing that approximately 5% to 10% of these allegations are not true, in order to eliminate the damage to children re-exposed to authentic sexual abuse incest because of an erroneous Type II “false negative” decision and response from us. Whereas the legal response of convicting the parent as a “sex offender” would be disconnected from the social service response, so that the legal criteria for conviction would maintain higher standards against making a Type I “false positive” decisional error of convicting an innocent person.
This approach would result in ALL child accusations of parental sexual abuse being accepted as valid by social service agencies, so that ALL child accusations of parental sexual abuse result in termination of parental contact with the child, but only a limited number of these cases would actually result in a legal conviction of the parent as a “sexual offender.”
While separating our social service response to child allegations of parental sexual abuse (limiting our risk of making Type II decisional errors) from our legal response to child allegations of parental sexual abuse (limiting our risk of making Type I decisional errors) would provide for interesting and lively discussion, such an approach to separating decisional criteria would be unlikely to withstand legal challenge in the Courts. But I believe this possibility at least merits considered discussion as we strive for synthesis of equally valid poles of the dialectic.
Low Base-Rate Phenomenon
When a condition is rare in the general population this can substantially affect the rate of “false positive” decisions (Type I errors) we make from our assessments. For example,
Say we are assessing for TRAIT X in the population.
The population prevalence of TRAIT X is 5% (i.e., a “low base-rate”) and our instrument for identifying TRAIT X is 95% accurate. No approach to identifying a psychological trait in a population will be 100% accurate and in most cases a sensitivity of 80% is generally considered excellent. But for the purposes of our example, let’s say we have an amazingly good assessment instrument that can accurately identify 95% of the cases of TRAIT X.
So out of 1000 people assessed for TRAIT X, the prevalence of TRAIT X will be 50 people (i.e., 5%). Our 95% accurate instrument will then correctly identify roughly 48 people with TRAIT X, and will miss only 2 people who actually have TRAIT X but who we said didn’t have the trait (i.e., “false negatives,” a Type II error). Correctly identifying 48 of the 50 people with TRAIT X (i.e., “true positives”) while only missing 2 people with TRAIT X (i.e., “false negatives”) is pretty good.
Great. So far, so good.
However, there are 950 people in our population of 1000 without TRAIT X, and our instrument has a 5% error rate, so our 95% accurate instrument will identify 5% of the 950 people without TRAIT X as incorrectly having the trait (i.e. “false positives”), so that our instrument will incorrectly say that 48 people have TRAIT X when they don’t (5% of 950).
This means that out of 1000 people, our 95% accurate instrument will correctly identify 48 people as having the trait (“true positives”) and will miss only 2 who have the trait (“false negatives”), and will also incorrectly identify 48 people as having the trait when they don’t (“false positives”). So essentially, our identification of TRAIT X is only 50% accurate, we correctly identify as many people as having the trait (“true positives”) as we incorrectly identify people as having the trait when in actuality they don’t (“false positives”).
This is the result of what’s called “the low base-rate phenomenon” – no matter how accurate our assessment instrument (even an insanely accurate 95%) we will nevertheless produce and inordinately high number of “false positives” because of the low-base rate of the characteristic in the population.
Out of all the cases of child allegations of parental sexual abuse, the base rate of false child allegations is going to be low (my estimate is around 1% to 2%, others may assert other prevalence estimates).
If our value system says that we should avoid “false positive” identifications (i.e., avoid Type I decisional errors), then any approach that identifies as many “false positives” as “true positives” is going to present a problem. If we are trying to identify cases of “false child allegations of sexual abuse” and out of 1000 cases of alleged sexual abuse we identify 50 cases of false allegations, but in order to do this we also incorrectly identify 50 children who were ACTUALLY sexually abused by the parent as NOT being sexually abused by the parent, that’s a lot of kids who were authentically abused who we are not protecting simply because of the low base-rate phenomenon. We protect as many children from child abuse as we expose to child abuse.
AND… if our assessment method is less that 95% accurate, then the number of decisional errors increases substantially. So in actual practice, in order to identify the 50 cases of false allegations of sexual abuse we may wind up erroneously identifying 200 or 300 children who were actually sexual abused as NOT being sexually abused by the parent. This may mean that for every one parent-and-child we protect from the pathology of a false allegation of sexual abuse, we expose as many as four or five other, authentically abused children to continued sexual abuse victimization.
At what point does the damage caused to the 200 or 300 children who were authentically abused outweigh the damage to the 50 children who were not abused and whose false allegations were the product of severely distorted parenting by a narcissistic/borderline parent?
And yet do we simply abandon these “false allegation” children to the severe psychological abuse inflicted on them by the psychopathology of the narcissistic/borderline parent in order for us to reduce the harm to other children who were authentically sexually abused? Is it moral to knowingly sacrifice one to save many?
Or do we still try to identify these “false allegation” children anyway, even though we know that this will lead to more Type II errors in identifying children who were authentically abused, so that we will be abandoning some authentically abused children to the psychopathology of their pedophile parent?
It would be wonderful if we could achieve 100% accuracy in our diagnosis of both “true positives” (children we identify as being sexually abused who were actually sexually abused by the parent) and “true negatives” (children who make false allegations of sexual abuse and who we identify as false allegations), but be we can’t. There will always be some error in our decisional criteria. Do we wish to limit Type I errors and increase Type II errors, or do we wish to limit Type II errors and increase Type I errors. This is a value judgment based on the relative damage we judge to result from each type of decisional error.
Another consideration in this discussion is that the legal system itself will bias decisional criteria to limit Type I errors (i.e., to limit the probability of incorrectly identifying children as being sexually abused by a parent when there was no sexual abuse), so that the rates of not identifying authentic sexual abuse will already be inordinately high. Since this is already the case, should we then try to identify actual cases of false allegations of child sexual abuse in order to rescue and protect as many of these children as we can, since trying to identify actual cases of false allegations of child abuse is not likely to affect rates of identifying authentic child sexual abuse cases since the Court is already not identifying these cases anyway by its decisional criteria to limit Type I errors.
Conclusion to Part I
The issues are complex. My assumption is that ALL mental health professionals want what is best for children, and that ALL mental health professionals want to protect all children from all forms of child abuse.
Differing views on how to achieve this represent “equally valid poles of a dialectic” (Linehan, 1993). We should work for synthesis of these poles in which we achieve consensus on a reasonable approach to very difficult and challenging issues, rather than engaging in continual unproductive adversarial conflict in which each side falsely “demonizes” the other as being callously unconcerned about protecting children from child abuse.
We all want to protect children from child abuse. The issues are difficult and challenging.
Moving forward, I hope to engage in productive dialogue regarding these complex and challenging issues in which reasonable people can disagree about approaches while still maintaining a fundamental agreement about the underlying desire to protect ALL children from ALL forms of child abuse.
In a future blog post (Part II) I will discuss my personal views regarding potential resolutions of the complex and difficult issues surrounding child allegations of parental sexual abuse. In this next blog post on the subject of false child allegations of parental sexual abuse, I will offer my thoughts on whether there are specific distinctive identifying features in the 1% to 2% of cases reflecting false child allegations of sexual abuse as a result of the pathogenic influence on the child of a cross-generational coalition with a narcissistic/(borderline) parent (Causal Origin 2). If there are distinctive and reliable identifying features for false child allegations of parental sexual abuse originating from the pathogenic influence on the child of a narcissistic/(borderline) parent, then we may be able to protect some of these children without inordinately affecting our ability to ALSO protect children whose allegations of parental sexual abuse are true.
I recognize ahead of time that no matter what position I take in this second blog post there will be those who disagree, and that the views of these people are reasonable and well founded. Down the road, they may convince me of the greater correctness of their views, or I may convince them of my views. At this point I believe the primary issue is engaging in reasoned and productive dialogue that recognizes that both poles in the dialectic represent reasonable and understandable positions that merit careful consideration and dialogue.
Craig Childress, Psy.D.
Clinical Psychologist, PSY 18857
Linehan, M. M. (1993). Cognitive-behavioral treatment of borderline personality disorder. New York, NY: Guilford
“Staff splitting,” as mentioned earlier, is a much-discussed phenomenon in which professionals treating borderline patients begin arguing and fighting about a patient, the treatment plan, or the behavior of the other professionals with the patient… arguments among staff members and differences in points of view, traditionally associated with staff splitting, are seen as failures in synthesis and interpersonal process among the staff rather than as a patient’s problem… Therapist disagreements over a patient are treated as potentially equally valid poles of a dialectic. Thus, the starting point for dialogue is the recognition that a polarity has arisen, together with an implicit (if not explicit) assumption that resolution will require working toward synthesis.” (p. 432)