Accentuate the Positive

Doctor, machine, state say: oh yes you can!

On page 19 of a computer software manual it states that the software “has been programmed to understand the content of phrases”. This is a big claim and a huge leap from early computer software. ELIZA, the first computer program to simulate human conversation had an incarnation called DOCTOR that mimicked the therapeutic model of person centred counselling. This was a reasonably simple interaction in which statements could be mirrored back through questions. It certainly did not understand the content of what people were telling it, but in its simple reformulation of content, it did create the possibility for reflection on the part of the “client”. Many people did in fact open their hearts to DOCTOR, telling it their life stories, perhaps for this reason. The software that I started this paragraph with also has a medical context although perhaps a less simulated one than DOCTOR. The “Logic Integrated Medical Assessment” (LiMA), introduced in 2005 is used by the multinational company Atos Origin to assist doctors in assessing benefit claimants’ capacity to work. But in what sense does it understand content? Or more importantly perhaps how has it been programmed to understand? Also, in being used by the state, how is it implicated in its biopolitical machinations?

In Postscript on the Societies of Control, Deleuze follows Foucault in describing mechanisms of power and argues that we are in the process of replacing the disciplinary societies analysed by Foucault with what he terms the societies of control. The corporation is a key component of this new social formation. Atos Origin is currently paid £100 million a year by the UK government to test people’s “fitness for work” and has an annual revenue of €8.6 billion. Among other contracts, Atos also have a role in military intelligence training for the MoD and is “the Worldwide Information Technology Partner for the Olympic Games”. State sanctioned medical assessments for work capability, military intelligence and the Olympics, are all sites for the discipline of bodies and minds. With the introduction of digital technology and artificial intelligence (AI) into these areas, they are perhaps “pivotal points” between the disciplinary and control societies as described by Deleuze and therefore sites in which overlaps and contradictions between these two types of social formation become apparent.

The sickness and disability benefit system is currently undergoing a massive overhaul, with 1.5 million people being reassessed for what is now called Employment and Support Allowance (ESA). The change from Incapacity Benefit to ESA in 2008 was heralded as placing a more positive emphasis on what people are capable of rather than what they are incapable of. This stated intention is present in the language used: from incapacity to employment and support. However, the shift occurred just after the recession alongside increasing rhetoric about getting the “welfare cheats” off benefits. It was also accompanied by a much harsher test, resulting in terminally ill people being found “capable of work” and thousands of appeals overturning the decisions being made.

The increasing mechanisation and automation of the medical assessment started in the 1990s with the introduction of tick boxes and a score system to the test. LiMA takes this automation one step further. During the assessment, the assessor, a doctor, sits behind a desk with a computer on it while the claimant sits on the other side of the desk. The doctor physically examines the claimant, makes observations and is prompted by LiMA to ask a series of seemingly banal questions about the claimant’s lifestyle, such as: do you answer the phone when it rings? do you read or watch television? LiMA then mechanically constructs phrases that resemble the answers the claimant gives to these questions. These phrases are assembled through the use of drop down menus, allowing the assessor to choose combinations of words and numbers. LiMA compiles the phrases into a report and uses them to produce a score [Brighton Benefits Campaign]. These are the “phrases”that LiMA has been programmed to understand the content of. The assemblages of words and numbers constructed by LiMA are interpreted by the software to derive implications from the phrases that have been put together. This process is more like a translationthan an understanding. While understanding does involve interpretation it is one that draws on a vast array of different kinds of knowledge. The word understand is directly connected to embodied terms such as to grasp (OED). It also signifies the perception of intended meanings of words, actions etc. and therefore is totally bound up with our social relations. LiMA cannot of course understand in this sense. The implications that LiMA is programmed with are connected to ideas about ability and capability. If the claimant tells the doctor that she can make herself snacks or use a microwave, evidence of physical and mental capacities are derived from this statement. These capacities are abstracted from their context and used as functions of the body at work. The fact that someone can walk 50 m, open a door, take something out of a fridge etc are all taken as evidence of being able to do a certain amount of “work”.

This follows a certain definition of work that can be seen in the context of scientific management and anthropometrics. These disciplines abstract the human body and view it mechanically, aligning it with the division of labour. In the 19th Century, the physiology of labour was born in military and penitentiary environments, the disciplinary environments of enclosure, where labour was more simplified and could be observed in an instrumentalised and objectified way [Doray: 1988: 73]. As a precursor of scientific management, Etienne-Jules Marey’s early analysis of movement through the photographic image of the body in motion helped prepare the way for calibration and standardisation. The technology of the machines of the time, of energy and combustion, were transposed onto the human body. Representations of Foucaultian biopower, they distribute in space, order in time. Marey photographed disciplined bodies: workers, soldiers, gymnasts. Now, think Atos Origin: work capability, military intelligence, Olympics.

The structure of the Work Capability Assessment and the interpretations programmed into LiMA equate human activity with the ability to work which in turn is equated with the ability to labour under the wage relation. Wage labour is not just about the ability to do something, it also involves stress: competition, deadlines, expectations of when and how something might be done, ever increasing time pressures and the extension of the working day. These all have effects on the mind and the body that are not easy to measure. Just because someone can walk 50 m doesn’t mean they can do it to order. This also begs the question of what kind of work claimants are being assessed for. It seems to be just a generalised category called “work”, with no accommodation as to what kind of work the claimant might otherwise be suited for or what is available. Massive changes in the labour market have taken place since Fordism with labour becoming more cognitive and also more precarious. More work takes place behind a computer desk. In manual labour, the body is instrumentalised, in cognitive labour it is the mind and imagination. “The content of labour becomes mental, while at the same time the limits of productive labour become uncertain. The notion of productivity itself becomes undefined.” [Beradi: 2009: 75] Labour itself has become more abstract, modulated. In The Soul at Work, Bifo suggests that as cognitive labour has grown so have mental health issues such as anxiety and depression. Increases in precarious working conditions, temporary labour and continuous self-management exacerbate the situation with individuals taking on more of the burden. For Deleuze, this is indicative of the Societies of Control, with undulatory man, having to continually undertake retraining. In the post Fordist world where work is linked to desire and self-realisation, Bifo argues that we are suffering from a mass depression. And while mental capacities can be measured such as understanding and awareness of the person’s context and situation, mental health issues are much harder to assess or understand. This is especially the case in the context of an assessment that looks at the body at work in such an instrumentalised way. In The Soul at Work, depression is described as the soul on strike [p12]. This brings up issues of refusal and resistance. In the short story The Apostate, Jack London charts a similar parallel in the industrial age. A young man brought up working in the mills suffers from neurasthenia so bad that he literally cannot get up, his body giving up on him. It is a study in fatigue that raises questions about conscious and unconscious responses to working conditions. Neurasthenia itself was not a straightforward condition, a 19th Century complaint of the nerves that perhaps bears a resemblance to Chronic Fatigue Syndrome, suggesting a more complicated relationship between body and mind. It has been compared to shellshock caused by the “wearing effect of modern technology”, a process of psychic self protection to survive modern industrial labour [Leese: 2002: 16].

The physical examination that takes place during the Work Capability Assessment, based on observation and manipulation, reinforces the mechanistic view of the body. The medical examiner moves the limbs to see how they move, looking at upper and lower limb dexterity, and the body’s mechanical ability to function. What this misses are conditions in which more complex relationships between the mind and the body are at play. In the LiMA handbook, it states that if the claimant is disabled by both physical and mental problems, the assessor will need to decide which one is the main problem, from a “functional perspective” and “curtail” the assessment of the “lesser problem” [p32]. This already separates the physical from the mental and ignores complicated interrelations between the two. As well as this separation, the assessment also does not take into account effects of chronic or variable pain on the body that may not be to do with problems with dexterity. Pain is subjective and very difficult to assess. It can be extremely debilitating, causing fatigue much sooner than if the body is not in pain. During the Work Capability Assessment, the claimant is not directly asked about pain, leaving any assessment of it to observation. If someone has been suffering from pain for a long time, there may not be any visual signs. There are assessment methods like the McGill-Melzack Pain Questionnaire [Melzack & Wall: 1982: 40) which could be used to give a better picture. This questionnaire asks patients to choose from a range of different kinds of descriptors including both sensory and affective adjectives. Terms like throbbing, pounding, shooting, cutting or burning; and tiring, sickening, nagging or vicious can be chosen to create a constellation of words that characterises their particular kind of pain. While it is impossible to exactly know what someone else’s pain feels like, these descriptions can give some kind of insight or understanding. Not only is pain not directly discussed during the Work Capability Assessment, but with the recent changes to the test, assessors have been asked not to take variable pain into account [CAB: 2010]. These omissions create more misunderstandings.

The results of the physical examination are fed into the final report along with any other observations the doctor might make. The intelligence distributed across the system, between the doctor and LiMA decides whether the claimant is capable of work or not. The relationship between the doctor and the software program could be seen as somewhat symbiotic. LiMA is a systematised formulation of the medically trained mind while at the same time the very structure of LiMA programmed with these particular interpretations, compels the assessor to think more like the program.

“Doctors pay more attention to the computer than the client; the system is inflexible and gives rise to inappropriate stock phrases in reports; options for investigation and findings are blocked off by the system inappropriately; doctors sign off reports without checking what they say, because the phrases have been generated by the system, not the doctor” [CAB 2009, p.4]

The autonomy of the doctor is undermined here by the rigid machinery of the software but also by that of the bureaucratic system. The relationship between the doctor and the software exacerbates a situation already existent within the technological system of the state. The phrases that LiMA constructs and “understands” always already exist before the claimant has spoken. The machine speaks for the claimant, translating and interpreting. The claimant is interpellated by the state through the machine, written by the machine.

And it makes mistakes. The results of these automated ‘cutups’ can be surreal. A recent report by the Citizens Advice Bureau revealed that 43% of assessments included serious levels of inaccuracy [CAB 2012]. Several Social Security judges when overturning cases ruled that reports contained “nonsensical statements”. And this is all the more horrifying precisely because it is used as evidence against the claimant by the state. When someone arrives at a Work Capability Assessment they are at the point at which they are unable to take part in wage labour because of medical issues and therefore not be productive to the state. This is where the state intervenes, the claimant’s medical issues, their personal and private life becomes the business of the state. Over the last three years 31 people have died while waiting for their appeals shortly after being found fit for work [Guardian 2012]. Rather than genuinely supporting people back into work it seems to be being used as a way of disapproving someone’s claim that they can’t work. In Kafka’s The Trial, the situation is at once banal and horrifying. The opacity of the crime, the arrest and the proceedings in general result in a frightening situation. It is one in which logic is obscured, there is no sense, no understanding. It is “the most fearsome of judicial forms” in which the law itself is in crisis, in transition. It is at a transitionary point between “the apparent acquittal of the disciplinary societies”, and “the limitless postponements of the societies of control” [Deleuze: 1998]. This creates a situation that ultimately results in cruelty. K’s guards know little about why he has been arrested, whether he is being charged or what will happen next. They claim to know only what they need to and no more. They are like cogs in a machine. Unthinking, they are perhaps an example of what Hannah Arendt calls “the strange interdependence of thoughtlessness and evil” [2003: 380] where the inhuman becomes inhumane.

Increasing complaints about the inhumanity of the Work Capability Assessment by the Citizens Advice Bureau, disability activists and others finally led to an independent review in late 2011. The Harrington review found the need to increase the transparency of the assessment and improve support and communications for people as they moved onto jobseekers allowance. It also called for checks to be introduced on the decisions to ensure fairness and consistency and for disability groups to be involved in providing guidanceforAtos healthcare professionals and Decision Makers. However “despite a successful trial in which Atos staff were deployed in benefits centres, which the government found was ‘an effective way of improving communications to discuss borderline cases’, Atos decided it would not be able to continue with the initiative because of its own ‘capacity pressures’”. [Social Welfare Advocacy: 2012]

The financial cost of providing more face-to-face communications was obviously too much for Atos. Automation in the name of streamlining, efficiency savings and speed often results in loss of personnel, in job losses. In this case providing more support for disabled people to go back into work themselves. Any change in the sickness and disability benefit system should be about providing real employment support rather than efficiency savings on the part of a corporation. This is indicative of the policy of small government, minimising the cost of benefits without taking responsibility for providing jobs. Employers in general are not very good at understanding or accommodating the needs of workers with health problems. They are often not keen to take on people who have a history of mental health issues for example and might need support and encouragement to do this. In the LiMA manual it is stated that assessors should emphasise what people can do, and while this should promote more understanding and support in getting people back to work, as we have seen, it is used as evidence against the claimant, to prove them wrong. Computers make mistakes precisely because they don’t understand. It is often difficult enough for people to feel that the medical profession understands their complaint or condition particularly if it is chronic or complicated. There also seems to be little understanding of how work is changing and the shift to more self managed working practices and the toll these can take. If work is becoming more modulated, the assessment still takes its core ideas from the disciplinary model. Complex health conditions are separated into mind or body “functional” problems by the assessment but this builds on problems already inherent in the medical model itself. These mismatches and misunderstandings create the possibility for harshness and cruelty. Rather than understanding the claimant, the assessment seems to work to obscure or mystify. Another aspect of understanding, is of course sympathy, or perhaps more importantly empathy. Without the ability to think and feel ourselves into the shoes of others, we are left only with categories instead of people, and ultimately inhumanity. For Deleuze, neither disciplinary or control regimes are harsher than the other although the modulations of control may be more complex. And with this in mind, I wonder how the need for real understanding and empathy could be harnessed.


Arendt, Hannah & Baehr, P.R. The Portable Hannah Arendt. Penguin Books, New York 2003

Beradi, Franco “Bifo”. The Soul at Work: from Alienation to Autonomy. Semiotext, Los Angeles 2009.

Deleuze, Gilles. Postscript on Societies of Control.

Doray, Bernard. From Taylorism to Fordism: A Rational Madness. Free Association Books, London 1988

Hales, Catherine. How We Became Post-Human: Virtual Bodies in Cybernetics, Literature and Informatics. University of Chicago press, Chicago 1994

Kafka, Franz. The Trial. Penguin books Ltd, Middlesex, UK 1977

Leese, Peter. Shell Shock: Traumatic Neurosis and the British Soldiers of World War 1. Palgrave MacMillan, New York & Hampshire, UK, 2002

London, Jack. The Apostate. First published in Woman’s Home Companion, 1906

Melzack, Ronald & Wall, Patrick. The Challenge of Pain. Penguin Books, London 1982

Gentleman, Amelia. Inaccuracies dog ‘fit to work’ test. The GuardianTuesday 10 January 2012.                               

Gentleman, Amelia. No Turning Back on Fit-to-work Test. The Guardian Tuesday 6 September 2011.


“Atos Computer Programme Manual (LIMA)”, n.d.

“Atos ‘Failed to Comply with Government Policy’ | Social Welfare Advocacy”, n.d.

“Decision-making and appeals in the benefit system”, September 2009.

“Right first time? An indicative study of the accuracy of the ESA work capability assessment reports”, Jan 2012

“Employment and Support Allowance: a New Harsher Test « Brighton Benefits Campaign”, n.d.

“Wca-review-2011.pdf”, n.d.

“Work Capability Assessment Independent Review – DWP”, n.d.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s