Since the 1993 murder on Merseyside of the toddler James Bulger, the single most profound shift in our understanding of what creates disturbed children comes not from the criminal justice system, but from neuroscience. It is the graphic evidence for something many parents instinctively feel: that love is as vital to children as breathing. Brain scans conducted in the 1990s on children from Romanian orphanages deprived of almost all human interaction show a virtual black hole where the part of their brains dealing with managing emotions should be. The scans suggest the infant mind is not born but made, building itself like a muscle over the first two years of life as parental attention triggers physical responses in the brain.

Put simply, love builds a child's ability to relate to others. Love withheld creates a dangerous vacuum. A propensity to violence, or at least to uncontrollable emotions, may be hard-wired into abused children.

This argument, popularised in Sue Gerhardt's book Why Love Matters (2004), has advocates including the former Tory leader Iain Duncan Smith, the Labour MP Tom Watson and Rowan Williams, Archbishop of Canterbury. And I thought of it when the recall of Jon Venables to prison reawakened the mob. Why, with all this sophisticated understanding of what we used to call "evil", can we not see Venables and Thompson as both victims and perpetrators? Why are their parents not in the dock of public opinion instead?

Cycle of abuse

Yet, as time goes on, doubts creep in. There is unease among some academics about the "scientification" of parenting policy popularised by Gerhardt: where abuse is extreme, cause and effect may be clear, but there is less evidence that routine neglect alters brain chemistry.

Venables and Thompson certainly came from tough backgrounds. Thompson's mother was an alcoholic who frequently left her seven children to fend for themselves; Blake Morrison's book on the Bulger case detailed how they "bit, hammered, battered, tortured each other". Venables's mother suffered from severe depression and repeatedly hit him - he was afraid of her, ranging his toys up on his bed in a pathetic attempt at protection. Such accounts are depressingly common among child offenders. Yet was this parenting bad enough to explain the peculiar brutality of their crime?

Too many British children grow up in such homes, without emerging homicidal. This is more than an academic debate: the lesson from the Romanian orphans was that children adop­ted when over a year old had poor pros­ects of recovery. The damage was done. The causes of Venables's and Thompson's terrible crime may reveal much, both about their prospect of rehabilitation and wider debates about tackling dysfunctional families.

The recall of Venables means the Bulger tragedy now bookends the New Labour years - the alpha and omega of its criminal justice policy. In opposition, Tony Blair's interventions around the trial set the tone for his being tough on crime and tough on the causes of crime. Now his successor faces questions about whether the killers were adequately punished. However, the bigger question is about those root causes. Who is confident that a child born into a dysfunctional family today will be supported, as Venables and Thompson were apparently not?

The New Labour years are littered with well meaning and expensive solutions: Asbos, Sure Start, parenting orders, behavioural contracts, social-service reform, speeches about parental responsibility. Yet dysfunctional families appear stubbornly impervious. From Shannon Matthews to Victoria Climbié, from Baby P to poor starving Khyra Ishaq, the list of children betrayed by their parents echoes down the years. It is too hasty to judge measures in the short term which - like Sure Start - were always meant to take a generation.

But it is clear some huge questions were tackled too late, with too little rigour. The Climbié case was an enormous wake-up call to social services, yet the death of Baby P in 2007 indicated how pitifully little had changed.

Intensive care

Why, having identified the issues raised by the Bulger killing, was Blair (and later Gordon Brown) unable to resolve them? It is tempting to conclude, bleakly, that they are just too intractable. The Conservatives seem unable to offer any solutions either, beyond proposed welfare reforms not easily distinguished from the Labour version, and the puny weapon of tax perks for marriage. Those politicians advocating more sweeping changes are afforded respect, but no money: Duncan Smith's joint report with the Labour MP Graham Allen on early intervention was widely praised but deemed too expensive to implement.

So, where do we go from here? If infant development in the first two years of life really is critical, that raises new ethical questions about early intervention. How early should early be? There is evidence that a mother-to-be's heavy drinking or drug abuse affects the unborn baby's brain; in some US states, recklessly endangering a foetus is a crime. Here, we are instinctively nervous about anything that might drive mothers away from medical care.

Questions don't stop at birth. If the first two years are paramount, why aren't children taken into care on average until the age of six? Shouldn't social workers move faster, as the Barnardo's chief, Martin Narey, recently suggested?

These are precisely the sorts of unpopular and complex questions politicians are hopeless at answering. Better to ask ethicists, psychologists, lawyers, theologians and parents. These issues deserve deeper consideration than eve-of-election crackdowns can provide.

As for Venables and Thompson, for all that they have been through over the past 17 years, they remain essentially as they were at their trial: outcast, profane, unknowable.