If the 21st century is teaching us anything, it is that violence increasingly rules the world. Physical violence means being demeaned, degraded and humiliated. It means terror, hurt and pain. For its proponents, on the other hand, violence may mean respect, power and authority.

But in a society like the UK, only a small proportion of people have any regular contact with violence. Perhaps that explains our misbegotten love affair with fantasised violence on the box, screen and videogame. However, the minority who are at particular risk, also tend to be those who are most vulnerable in terms of age, ethnicity, gender, sexuality, disability and poverty.

It is just these groups who are most likely to be users of social care services. There tend to be two ways of presenting the issue of violence in social care - according to who is seen as the victim. One scenario sees control as the solution; the other sees it as the problem. The first tends to go something like this:

Social workers operate at the heavy end of the market. They work with some very "vulnerable" and "disturbed" people. Social services "clients" are associated with "high rates of mental disorder". There are growing problems of antisocial, disaffected young people. The problem of violence is predictable. Social workers need protection. This requires controlled environments, adequate alarm systems, defensible space. To ensure safety may require armoured glass, metal grills, locked entrances and bolted down furniture. Training in physical avoidance techniques, in some cases, unarmed combat, may be helpful.

Then there is the social control model, which runs along the lines of:

Social care has moved a long way from the aspirations of the radical social work of the 1970s. Statutory social services have increasingly taken on the role, culture and appearance of the poor law and benefits office. Their soft cop role has become increasingly evident. Contact with clients has diminished. Interviews, tied to ever-narrowing eligibility criteria, have become increasingly mechanistic. In a broader context of defensive practice, organisational back-watching, "welfare to work" and the political prioritising of benefit fraud, distrust is in the air. "Clients" are likely to be increasingly desperate as they face gatekeepers, long waits, polite and not so polite refusals in dreary offices.

There's probably not an objective fact any of us could reasonably challenge in either of these conventional characterisations. But both rest on complex assumptions which encourage over-simplification. Neither takes us very far in understanding or addressing the problem of violence in social care. Violence is a two-way street anyway and we should never forget this. We may never know just how many service users experience explicit and more subtle forms of violence within social care, especially within its residential services. The focus must always be on violence as it affects both service workers and service users.

Violence is the offspring of alienation and inequality. It thrives where human beings don't. So long as social care operates in this territory - as it does - there will be the potential for violence. There is no need to pathologise service users or equate mental distress with automatic violence and threat, to take this properly into account. Here's where we can helpfully connect the micro-world of social care with the wider world it operates in.

We should also remind ourselves here, that service users and those who work with them are often devalued and treated as marginal. The two groups are natural allies.

Social care may not be able to isolate itself from the wider world of violence and has its own role in dealing with it, but it can do more to minimise violence in contacts between service users and workers. This will include a culture from the top based on real respect for clients and an organisational approach which rests on seeing each person as an individual. It also means encouraging the personal qualities in workers of warmth, empathy, openness, reliability and respect. Also in situations where people may be at their wits' end, where the only strategy some may know is violence, it means an even greater emphasis on developing interviewing and communication skills, to identify difficulties and help avoid them. Most of all, it must mean that workers are never left alone on their own to cope.

Peter Beresford is professor of social policy, Brunel University. He is active in the psychiatric system survivor movement.