Cognitive Therapy

The cognitive model is a model of information processing. People do not respond to situations; they respond to their construal of situations, and construal is shaped by structures laid down through experience. The theory is therefore about the architecture of meaning-making, and it is deliberately thinner than psychoanalytic personality theory - a feature its proponents defend and its critics attack.

Three levels of cognition

Beck grouped core beliefs into three broad categories - helplessness ("I am incompetent, trapped, out of control"), unlovability ("I am unwanted, defective, bound to be abandoned"), and worthlessness ("I am bad, immoral, toxic"). Most clinical presentations cluster around one, with the others recruited under stress.

Modes and the diathesis–stress logic

Schemas are latent structures, not permanent states. The diathesis-stress account holds that a schema laid down early lies dormant until matched by a congruent life event, at which point it activates and dominates processing - which is why an intelligent, functioning person can, within days of a specific loss, become someone who genuinely cannot see evidence of their own competence. Later Beckian theory organises schemas into modes: coordinated cognitive, affective, motivational and behavioural packages that switch on together.

Each disorder has a characteristic cognitive content - the cognitive content-specificity hypothesis, one of the model’s genuinely testable and largely supported predictions.

Cognitive distortions

The named distortions are teaching tools rather than natural kinds. Their purpose is to give the patient a vocabulary for catching processing errors in flight - the label is not the point, the noticing is.

All-or-nothing thinking

Categories with no middle. "If I’m not the best on the team I’m worthless."

Overgeneralisation

One instance becomes a rule. "I always ruin things."

Mental filter

Attending only to the negative detail while the rest of the picture is discarded.

Disqualifying the positive

Actively converting positive evidence into nothing. "They were only being kind."

Mind reading

Assuming knowledge of others’ thoughts without evidence.

Fortune telling

Predicting the negative outcome as if it were established fact.

Catastrophising

Going to the worst case and treating it as the likely case.

Emotional reasoning

Taking the feeling as evidence of the fact. "I feel like a fraud, so I must be one."

Should statements

Rigid demands on self or others, producing guilt or anger.

Labelling

Attaching a global identity to a specific behaviour. Not "I made an error" but "I am a failure."

Personalisation

Taking responsibility for events one did not control.

Magnification / minimisation

Inflating one’s errors and deflating one’s strengths - the binocular trick.

Beliefs change when they are treated as hypotheses rather than facts, and then tested with evidence the patient gathers and finds credible. Two routes, used together.

The behavioural route generally produces larger and more durable belief change than the verbal route, which is why "behavioural experiment" and not "thought record" is the workhorse of contemporary cognitive therapy. Reasoning one’s way to a new belief tends to produce the "I know it in my head but not in my gut" complaint; running an experiment that violates the prediction does not.

What actually mediates outcome

The dismantling literature is uncomfortable for a strict reading of the model. Jacobson’s 1996 component analysis found behavioural activation alone as effective as full CT for depression. Studies of mechanism have repeatedly failed to establish that cognitive change precedes and causes symptom change rather than accompanying it. A serious contemporary position holds that cognitive therapy works, that its behavioural components are doing a large share of the work, and that the cognitive account of *why* it works remains under-determined. Learning the model and holding this reservation simultaneously is what distinguishes a sophisticated practitioner from a technician.