Seen But Not Examined: The Supreme Court's New Limits on Psychological Evaluation of Children in Custody Disputes
*Wagitha Narayan
There is a paradox at the heart of every custody dispute: the more a court tries to understand a child, the more likely it is to hurt the child it is trying to protect. Each interview, counselling session, or psychological test, one professional stranger at a time, asks a child to relive the collapse of her family. Yet the other instinct is equally perilous: a court that does not pay attention could never uncover the manipulation, coaching, or abuse that lies behind a child's rehearsed answers. May courts refuse to examine children, or does restraint sometimes mask the very harm that the law is meant to avert?
In Sheetal Vasant Thakur v. Chirag Arora, decided on 11th June 2026, the Supreme Court squarely faced this dilemma and in what appears to be the first framework of its kind, laid down a detailed, twenty-point framework for when and how children caught in custody and access disputes can be subjected to psychological evaluation. The closest prior guidance, in Sakshi v. Union of India, (2004) 5 SCC 518, had addressed a child's trauma while testifying in criminal trials but had never before been extended to custody or access proceedings. The ruling is worthy of note not because it resolves the paradox, but because of how honestly it manages it. This piece argues that the framework is doctrinally sound but institutionally fragile: it assumes trauma-trained, neutral child psychologists are readily available to Family Courts, and that assumption does not yet hold true in India. Whether restraint operates as protection or as an excuse for inaction depends less on the judgment than on what the State does with it.
The Case and What the Court Held
The facts explain the Court's anxiety, and they illustrate the very danger the judgment guards against. In 2015, the parties married and immigrated to the United States, where a daughter was born; the marriage broke down in 2019, with the mother returning to India with the child amid allegations of domestic violence against the husband. She later alleged that the father had sexually abused the daughter when she was around two years old, leading to FIRs under the POCSO Act. When the father sought renewed access, the Family Court in Pune declined to order a fresh psychological assessment, noting that the child was already receiving psychological care. The Bombay High Court went further: it replaced the four-member panel, largely nominated by the father, with a new panel including specialists in “parental alienation syndrome” and “false memory creation.” A ten-year-old already in therapeutic care, and a suspected victim of the very parent contesting custody, was thus to be assessed by a panel effectively assembled by the accused.
The Supreme Court set aside those orders and, exercising its parens patriae role, issued twenty guidelines for all courts handling such requests, built on three ideas that each do distinct work. First, minimum intrusion: evaluation should be exceptional, conducted by a single, neutral, court-appointed expert rather than a panel, directly answering the panel-stacking this very case produced (para 97(iv), (vi)–(vii)). Second, sequencing: the parents' psychological state is assessed first, and the child's condition is introduced through her treating psychologist wherever possible; only if direct evaluation is recorded as actually necessary is the child herself evaluated (para 93, Directions (A)–(D); para 97(xix)). This is the framework's most significant move: it shifts the burden of proof from the child's demeanour to the adults' conduct, since dysfunction is more likely to surface in how parents behave than in what a frightened child says. Third, reasoned necessity: every order must state why evaluation is needed and why less intrusive alternatives will not do. As the Court itself put it in Sheetal Vasant Thakur, “the justice delivery system cannot treat the child as a mere evidentiary object subjected to repeated forensic or psychological scrutiny at the instance of contesting litigants.”
Notably, the Court did not invent new doctrine so much as apply old commitments consistently: the child-friendly architecture of the POCSO Act, whose Sections 24, 33(5), 36 and 39 respectively mandate minimum exposure, protection from repeated testimony, and insulation from the accused, was read into civil custody proceedings (paras 42–49, 97(xi)). The concern is old: in Thrity Hoshie Dolikuka, the Court refused to interview a child caught between warring parents as early as 1982, citing the strain of repeated interviews on her sensitive mind.
What Sheetal Vasant Thakur adds is procedural architecture. Gaurav Nagpal had already read a child's welfare broadly, in its full emotional and psychological sense; this judgment goes further by holding that the same broad conception of welfare must also govern how courts are permitted to investigate it: welfare is no longer only a substantive standard for the final decision, but also a procedural limit on the methods used to reach it.
Why The Court Reached for Restraint
The Court's caution responds to a genuine problem, one this case demonstrates: an expert report can function as litigation currency, capable of undermining the other parent's credibility or lending scientific weight to an otherwise unsubstantiated claim. This is exactly the dynamic visible in the father's proposed panel of alienation and false-memory specialists. The rise of parental alienation claims sharpens this danger. Parental alienation, understood as a factual pattern of one parent's conduct affecting a child's feelings toward the other, is real and has been recognised in Vivek Singh v. Romani Singh(para 18); but the label is also an expanding one, since a child's ordinary hesitation can be recast as coaching. The Court in Ramneesh Pal Singh (paras 20–23) had already cautioned that “parental alienation syndrome” is not itself a diagnosable condition, but a question of fact to be established through specific alienating behaviours rather than invoked as a label: a distinction between alienation as a factual pattern the Court must find and “syndrome” as an unproven clinical category. The present case shows how easily that warning is ignored in practice: the panel the father proposed was already stacked with alienation and false-memory experts, turning a welfare inquiry into a tool for discrediting the child's allegations. The Court's insight is that the very process of discovering welfare can destroy it. Minimum intrusion, on this reading, is not judicial squeamishness: it is an evidentiary discipline that removes the incentive to weaponise a child's mind.
The Uncomfortable Counter-Question
But restraint has a dark side. Some harms are visible only to a trained eye: sustained coaching, memory contamination, and abuse are precisely the injuries a busy Family Court is least equipped to detect, and a calm child in the courtroom can be the most dangerous kind of silence. A rigid rule against evaluation risks favouring the parent who is abusing or manipulating the child: the best evidence money cannot buy is the child saying nothing at all.
The judgment does not overlook this danger; its design responds to it. The Court rejected the mother's absolute proposition that expert assistance is barred wherever POCSO allegations are pending, holding this would unduly restrict courts in obtaining informed assistance. What the framework limits is not the courts' final power to investigate, but the sequence in which they may do so. The sequencing rule does not close off inquiry into the child's condition; it redirects it: prioritising the adults' behaviour and the child's treating psychologist as the first sources of information, and reserving direct examination for cases where the Family Court records that these less intrusive routes cannot adequately assist it. The court retains continuing jurisdiction, and the assessment can be revisited as the child grows. The framework is a filter, not a wall: it does not prevent courts from uncovering hidden harm, but it requires litigants to justify why examining the child directly is truly necessary.
Has the Balance Been Struck?
In this author's view, the balance struck is defensible, but its defensibility is doctrinal, not institutional. Minimum intrusion assumes a single competent, trauma-trained child psychologist is available whenever intrusion is genuinely required; sequencing assumes professionals exist who can meaningfully assess adult family dynamics before the child is ever approached. Most Indian Family Courts have access to neither: as recently as December 2023, the Delhi High Court had to direct its Registrar General to depute even one clinical child psychologist to each Family Court complex in the capital, precisely because none was available there as a matter of course, and family law practitioners have flagged a comparable, nationwide shortage of trauma-trained child psychologists attached to Family Courts. The judgment itself concedes this: it notes the absence of any Indian study on children's mental health in custody proceedings, relying on a solitary NIMHANS study from 2025, and accepts that judges are trained in law, not in reading human emotion. This is a genuine vulnerability: a framework built around expert availability, announced in a system that has not yet built that expertise, risks curdling into avoidance. A judge without access to a competent psychologist can simply record that evaluation is “not necessary” and move on, and the harms the Court sought to address go unaddressed, not because the law failed, but because the infrastructure never arrived. Capacity constraints can hide behind the language of principle.
Practical Recommendations
Two steps could close this gap, each tied to a specific failure identified above. First, the minimum-intrusion principle presupposes that a real, neutral, trauma-trained child psychologist is available for the court to appoint; this case shows what happens when that supply gap is treated as an afterthought, since the High Court's own effort to name a single expert foundered on the absence of one in Jalgaon and produced, as a workaround, the very panel this judgment set aside. State Judicial Academies and NALSA could jointly maintain a verified, periodically renewed district-level panel of trauma-trained child psychologists, so that “the expert appointed by the court” names someone who actually exists at that court, and not merely an aspiration. Second, the sequencing principle presupposes that a judge can meaningfully test whether a parent's or child's condition genuinely warrants direct evaluation, a competence the Court itself concedes judges are not trained in (para 97(xx)). Basic instruction in recognising developmental and trauma indicators, built into judicial training under the Family Courts Act, would let judges interrogate an expert's reasoning rather than simply defer to it, addressing the risk that a judge unable to test an opinion either rubber-stamps it or avoids evaluation altogether out of unfamiliarity. Neither recommendation requires legislative amendment: both are matters of administrative will and institutional capacity-building.
Conclusion
The immediate effects for Family Courts are real: evaluation requests must be considered, not automatically granted (para 97(ii)–(iii)); reasons must be documented (para 97(iii)); parents are assessed first (para 93, Direction (A); para 97(xix)); single neutral experts are preferred over panels (para 97(vi)–(vii)); and evaluative reports stay confidential and cannot determine criminal guilt (para 97(xii)–(xiv)). That will visibly alter the texture of custody practice.
But the deeper contribution returns to the tension with which this piece began. Half a century ago, the Supreme Court in Rosy Jacob held that “the children are not mere chattels; nor are they mere play-things for their parents”, a principle that fixed the purpose of custody law without addressing what enforcing that law does to the child. Sheetal Vasant Thakur completes that thought: a child is not an evidentiary object either, and a court can protect her by declining to examine her, provided the refusal is a considered act of protection, not an unexamined act of convenience. That proviso is this piece's central point. The Court has done its part by making restraint a reasoned, reviewable choice rather than a reflexive one; whether it remains principled now depends on whether the State builds the panel of experts this judgment assumes already exists. Until it does, the line between restraint and neglect will be drawn not in judgments, but in under-resourced Family Courtrooms.
Comments