Keynote Address – Dr Felicity Gerry KC prepared and developed with Phillipa Stafford
11th Global Summit on Feminist and LGBTQ+ Studies · Berlin · 24 September 2026
This keynote address is taken from a forthcoming book chapter on reflective practice for professionals when working with women in criminal justice. It explores my career with an emphasis on how trafficked women who commit offences are systematically misidentified as criminals rather than recognised as having a right not to be prosecuted or punished when subjected to exploitation, coercion, deception, and abuse of vulnerability such that their alleged crimes are caused by others.
Drawing on research findings from the Catch 22 project in Victoria, Australia the contention in this keynote address is that the erosion of women’s autonomy characterises trafficking and that this should be understood as negating core elements of criminal liability, rather than selective altruistic ‘protection’ or merely mitigating punishment. Research has already identified the ‘non-punishment principle’, as an emerging human right not to be prosecuted. Implementation depends on understanding how justice systems function poorly, I suggest, particularly for women, although much of what I say equally applicable to the LGBTQ+ community but focus of my research is on women.
I have spent over thirty years as a criminal barrister, sometimes representing women who are accused of crime or being alleged to be complicit, women who are, in my case examples, also victims of crime. Women trafficked, coerced, groomed, controlled. The overwhelming practice reality is that women do not fit the shape the law has built. And across jurisdictions there is the same silent architecture: a set of assumptions about what a woman is supposed to be, sexually and morally, and a legal and social system that punishes her the moment she departs from it. That is what I mean by gender constructs leading to sexual shame. Not shame as a private, individual feeling, though it is certainly that, for the women I represent, but shame as a structural tool. Shame that is built into charging decisions, into trial approaches and sentencing remarks, into the supposed defences that exist and the ones that, conveniently, do not.
Today I want to move from diagnosis to prescription. I will use the stories and the theory that have shaped my career to argue for three specific legal changes, changes that any jurisdiction could adopt, that would at least arguably begin to take shame out of the law’s architecture.
Removing the Binary
Criminal law likes binaries. Victim or offender. Innocent or guilty. Credible or not. But the women in criminal justice rarely fit either of these boxes neatly. They are what the literature has identified as ‘victim-perpetrators’, especially women who commit crimes, caused by others and thus because of their own exploitation. When a legal system forces a binary onto a woman who is both harmed and accused, it reaches for the oldest gendered shortcut available to it: Is she a good woman or a bad one? Is she chaste, passive and deserving of protection, or is she autonomous, complicit, and therefore fair game?
This is the ‘ideal victim’ problem, and it is a profoundly gendered one. The ideal victim is pure, passive, and, crucially, sexually blameless. The moment a woman’s story involves sex work, sexual exploitation, a relationship with the very man who trafficked her, or simply a level of resilience and assertiveness that does not read as fragile, she stops looking like a victim to systems and is framed as a participant. Prosecutors, judges and (in some jurisdictions) juries reach for stereotypes and gendered narratives to reframe her as the architect of her own harm.
I want to give you a few examples from my own practice, because I think the specifics matter alongside the theory which, in my PhD was that the dominant strategy is silence, so discovering shame in criminal justice is a difficult exercise.
Christine Keeler, and the sentence that outlives the century
I am currently part of the team which has filed an application for a posthumous pardon for Christine Keeler. In 1961, at nineteen years old, she had a sexual relationship with a government minister, John Profumo, who lied to Parliament about the affair. The evidence is emerging that she was just one young woman in a network. Keeler (CK) was the one who was called a “prostitute”, despite her protestations to the contrary. She was the one prosecuted and imprisoned, for allegedly lying about events surrounding an assault in which she was, in fact, the victim, and in which her attacker (LG) himself later admitted his guilt in open court. CK was prosecuted and imprisoned, even though her denial in court of the presence of two additional witnesses at the crime scene was immaterial (and thus not perjury), and indeed both witnesses ultimately did confirm that they had each seen the attack by LG on her take place. She pleaded to that offence (of allegedly lying) under extraordinary pressure, in the middle of a national political scandal not of her making and then spent the rest of her life trying to explain a miscarriage of justice that the politics, legal system and the public had already decided it apparently understood.
Sixty years on, we are still cleaning up after a construct: a young woman who had sex is a liar; a powerful man who lied to Parliament is knighted for services to charity. If we are serious about a national conversation on freedom from violence and abuse for women and girls, that injustice should be named for what it is; a woman labelled as sexually loose and dishonest, used to frame all women as sexually loose and dishonest. The shame did not belong to her. It was assigned to her, and it has stuck for more than sixty years.
The ‘madam’, and the defence that would not apply
I once defended a woman charged as a ‘madam’ in a sex trafficking network. She had herself likely been trafficked. Many are deceived by men who claim to be intimate partners. She moved up a chain of command until the evidence was she was the one telling other women what was expected of them. Two other women had been deceived into travelling to the UK, stripped of their papers, raped, and forced into sex work. My client did not give evidence. Her appeal, built on medical evidence of her own vulnerability and coercion, was dismissed. She was sentenced to eight years.
In England and Wales there is now a statutory defence for people compelled by trafficking to commit crime but conduct in a human trafficking network is expressly excluded for sexual exploitation offences. I have been sitting with that since the Modern Slavery Act 2015 UK came into force. The one category of trafficking-related offending that is built entirely around exploitation is the one category where the law refuses to ask whether the accused was herself a victim. That is not an accident of drafting. It is a gender construct written directly into a statute: where women in trafficking networks, especially sex trafficking are treated as too shameful, too compromising, for them to be ‘protected’ in the way we now examine drug offences or document fraud. The shame is not incidental to the law. In schedule 4 of the MSA which excludes that defence, the shame is the law.
This case is also, I think, the clearest illustration of why the existing defence in the UK is built on the wrong model. Section 45 of the Modern Slavery Act asks whether a person was compelled to commit the offence, and whether that compulsion is attributable to the trafficking, and whether a reasonable person in the same situation with the same characteristics would have had no realistic alternative but to act as she did. The compulsion model borrows its shape from duress: a single moment of coerced choice. It does not describe who caused the crime. Grooming, debt bondage, the removal of documents, threats, loss of identity or when ‘choosing’ to supervise other trafficked women is the only route that felt survivable are not single compelled acts but a causal process that hollows out autonomy. A compulsion model cannot see that process. A causation model can and it is being used elsewhere.
The woman alleged to be complicit in a stabbing when she froze
Across every case, as my PhD emphasis on game theory mathematically found, the dominant strategy of exploitation is silence, through fear of traffickers, and silence through fear of the very authorities who hold the power to ‘protect’. Language barriers compound it. But so does shame. Women who have been exploited do not appear as accused persons in a police station and calmly narrate what happened to them in the linear, consistent, emotionally regulated way that our courts expect of a credible witness. And when they cannot do that, when the account comes out in fragments, out of order, incomplete, the system does not ask what trauma happened to her. It asks what is wrong with her account as an accused person. This is not exclusive to trafficking cases
I recently defended a young woman charged with murder and manslaughter after her co-defendant fatally stabbed a man in the course of a dispute over money he was alleged to owe them from a car rental investment scheme. The prosecution said because she did not react, she was was party to a plan, and it pointed to her presence, her proximity, and her lack of reaction before and panic after as apparent proof that she knew what was coming and had signed up to it. What the unchallenged expert evidence actually showed was a woman who had suffered violence and abuse for over a decade, and whose passivity and “freezing” were not signs of planning but the well-documented response of a traumatised person to a rapidly unfolding, terrifying event. Hindsight and the retrospective reading of footage distort reality; the law demands discipline, not inference dressed up as proof. The jury acquitted her of both murder and manslaughter, but the structure to shame was there. A woman’s stillness in the face of violence is not evidence of her autonomy. Read as evidence of complicity, freezing becomes just one more assumption that a “real” bystander to violence would resist, flee, or intervene, and that a woman who instead goes silent and still must therefore be complicit in it.
This is where gender constructs do their quiet, devastating work. A woman who freezes is read as complicit. A woman who stays is read as consenting. A woman who does not explain her role in a crime immediately is read as unreliable. Every one of those readings is a story about how an ‘ideal’ victim is supposed to behave, and every one of those stories has shame sitting underneath what she is not able to say out loud.
The suicidal mother
In Victoria, Australia, a woman I represented, suicidal, escaping a violent partner, briefly and momentarily unwell with her own child in a moment of crisis in which the child was, medical evidence showed, entirely unharmed, was ultimately given what our law calls a ‘non-conviction’ outcome. Section 8 of the Sentencing Act there allows a court, having found the charge proved and to sentence but to decline to record a conviction, so the offending is acknowledged and the sentence is real, but the woman is not permanently branded a convicted criminal for conduct that was itself the product of her exploitation and crisis.
Vulnerability, autonomy, and the trap in between
Martha Fineman’s vulnerability theory argues that vulnerability is not a special condition of a few unfortunate people. She says it is a universal and constant feature of being human, produced and worsened by social structures. I find that framework useful, but I want to push it further in this address. The structures that produce shame for women are not neutral. They are built on a very old, very gendered demand that women be simultaneously vulnerable enough to deserve ‘protection’ and autonomous enough to bear full responsibility for anything that goes wrong, and never both at once.
A woman cannot be both a survivor of exploitation and a person who exercised any agency within it, in the eyes of a system that still seeks perfect victims. So, when a trafficked woman finds ways to survive, takes on a role in a criminal network, manages other women, stays with an abuser to protect a child, has sex with someone more powerful than her, that form of survival is misread as autonomy, and that misread autonomy becomes the peg on which shame, and often a prison sentence, hangs.
This is precisely why a compulsion model of defence fails women, and why the law needs to extend ‘protection’ not only the person at the bottom of a network but all those alleged to be complicit. A ‘madam’, a lookout, a driver, a person who collects money, an alleged accessory of any kind, may be exercising exactly the misread survival-autonomy I have just described. If the defence only ever reaches the most visibly powerless person in the crime, it simply relocates the ideal-victim test one rung up the ladder, and shame does the rest.
Why this belongs in a room about gender and sexuality studies
As a lawyer and perhaps an accidental theorist I have reached the view that criminal courtrooms are one of the purest laboratories we have for watching gender constructs convert into shame in real time. Nowhere else is the ideal-victim myth tested so directly, when a woman’s liberty is at stake. When intersectionality is added; race, class, immigration status, sexuality, disability, the shaming intensifies rather than dilutes. Women who are not white, women who are queer, women who are poor, are given even less benefit of the doubt when their stories do not fit a passive, chaste template. I am noisy – so I think the same is true of a noisy woman lawyer. Gender constructs are not abstract. They show up as a schedule excluding a defence. They show up as an absent of defences altogether. They show up as a sixty-year-old newspaper label that a woman’s family is still trying to have overturned after her death.
Three legal changes
I do not think the answer is simply more sympathy, because that frames women as weak. I think the answer is structural: An ethics of care built into legal frameworks, not bolted on afterwards. Restorative justice approaches that ask not, ‘which crime’ but ‘who was harmed and how’, and that make room for a woman’s account to be heard on its own terms, rather than measured against a template of the ideal victim. Where voluntariness and automatism laws recognise trauma and dissociation as complete defences. Where non-punishment principles include human trafficking networks. Where a fragmented, delayed, inconsistent account by an accused woman can be accepted as the fingerprint of trauma and shame.
From that structural view, and from the case studies I have just given you, I want to leave you with three concrete legal changes. None of them is radical. Each already has a foothold, however partial, in an existing jurisdiction somewhere. What is missing is the will to put them together.
- Defences for trafficked persons who commit crime should be built on a causation model, not a compulsion model. As the ‘madam’ case shows, the existing English and Welsh defence under section 45 of the Modern Slavery Act 2015 asks whether the accused was compelled, in the manner of duress, at the moment of the offence. That question is too narrow and too blunt for what trafficking actually does to a person. A causation model asks a different, prior question: did the trafficking cause or negate the elements of criminal liability, in particular the voluntariness of the act and the fault element, rather than merely excuse an otherwise complete offence. This is not a semantic distinction. A compulsion model waits for harm. A causation model treats sustained exploitation as capable of negating choice and identifying others as the culpable cause. Voluntariness and automatism doctrines already do this work in other contexts; there is no principled reason they cannot do it here.
- That causation-based defence should extend to those within human trafficking networks and to all alleged accessories, for any offence. The current UK statutory defence expressly excludes conduct within a human trafficking network for sexual exploitation offences under schedule 4. That exclusion written directly into legislation is the one category of offending built entirely around exploitation is the one category the law refuses to examine. A woman who is trafficked into supervising other trafficked women, who drives, who collects money, who is charged as a principal or (in any crime) as an accessory, has not thereby stepped outside the class of person the defence exists to ‘protect’. Ultimately, modern Slavery defences should be complete defences to any crime but especially in human trafficking networks.
- Third, jurisdictions should legislate a non-conviction outcome of the kind found in section 8 of Victoria’s Sentencing Act 1991. This allows a court to describe what had happened honestly, without either denying its seriousness or permanently branding the accused mother with it. It is a small legislative mechanism, but my example is what it looks like when a system chooses description over shame. Other jurisdictions without an equivalent non-conviction power should legislate one, and it should be drafted, or its use expressly encouraged by sentencing guidance, so that it reaches trafficked and exploited women whose alleged wrongdoing is tangled up with their own victimisation, including where that tangle is human trafficking or as an accessory.
Together, these three changes move the law from asking ‘was she compelled’ to asking ‘what did this do to her’; from protecting from abuse of vulnerability in a trafficking networks to protecting everyone caught in its causal chain; and from a system that must choose between silence and a permanent criminal record to one that can describe what happened or how she came to be assisting in some way and still let an accused woman move on. That is what actual protection looks like so that criminal justice is not a further exercise of survival.
Conclusion
Gender constructs tell us who is allowed to be hurt. They tell us which women get to be victims, and which are labelled as something else; manager, liar, slut, accomplice. And because sex remains the site where female virtue and female danger are still measured against each other, shame is very often where those constructs do their sharpest work.
A causation‑based “voluntariness” model as a complete defence for trafficked persons who commit crime, extended to everyone within a trafficking network and to all alleged accessories, and paired with a legislated non-conviction outcome of the kind Victoria already has in section 8 of its Sentencing Act 1991, alongside safe pathways for women to disclose offending or complicity without incurring criminal liability, should be a short step forward. By focusing on women’s lived experiences, we can frame the criminal law’s approach from individual blame to structural accountability and calls for survivor‑informed reform. Game theory allowed me to suggest that justice is the equilibrium between the extremes of commission of crime and not prosecuting that person for that crime. The essential repositioning is that women can move from objects of punishment to subjects ‘protected’ by law that functions to recognise women’s autonomy and remove shame.
Phillipa Stafford is an early career lawyer, artist and researcher. She was the recipient of the Victorian Supreme Court Award and the Deakin Honours Prize for her honours thesis on gender in international criminal law. Phillipa works in criminal law in lutruwita/Tasmania.
