Mad is probably the commonest term of abuse to describe loathed ex-partners, once their Badness and Sadness have been fully aired. While we tend to be motivated by malice at such times, this insult can be closer to home than you might think.
I know of eight studies that tested the mental health and personalities of couples before they married and then followed them up. In all of these, premarital 'neuroticism' in either partner (including mild or severe depression) and 'lack of impulse control' in the male partner predicted divorce compared with couples without these traits.
For example, one study examined 300 married couples in 1940 before they had married and followed them up in 1980. Divorce was significantly more likely in couples where one or both had had emotional problems before they married. Divorce was commoner in those of both sexes who, 40 years before, had been high in neuroticism and in men with a lack of impulse control back then.
The author's conclusion was: 'In marital relationships, neuroticism acts to bring about distress, and the other traits of the husband help to determine whether the distress is brought to a head (in divorce) or suffered passively (in a stable but unsatisfactory marriage).'
That the problems predate the couple meeting is suggested by a study which examined a large sample of 16-year-old girls, before they had even met their husbands. High neuroticism at that age predicted subsequent increased risk of divorce.
On top of this, emotionally rocky folk tend to shack up with each other. Dozens of studies of patterns of attachment show that the insecure attract the insecure and that the secure tend to pair up together. Hundreds more studies suggest attachment patterns are caused by early childcare.
Which may provide a different take on your relationship problems: the next time you are about to scream 'nutter' at your partner, pause to consider that if they are like that, the chances are, so are you.
My conclusion is that the 40 to 60 per cent of all people who have significant premarital emotional problems (mostly due to disturbing childhoods) are far more likely to end up divorced than was the case in 1950. The de-stigmatised, cheaper, simpler divorce process makes it far easier for disturbed people to project their problems into partners, convince themselves that they are with the wrong one and move on.
Sadly, that mostly does not work: divorcees who remarry are twice as likely to divorce as marital first timers. It's great that incompatible couples can more easily move on. But for a large slice of the population, there is a grave risk of blaming relationships for personal problems that would exist whoever you are with, needing individual treatment, not couple counselling.
Sorry folks, but it very probably would be no better with someone else. If you cannot see this in yourself, just think of the way your friends endlessly make the same mistake in their choice of mates.
Next week: peculiarly modern aspects of gender rancour.
· Oliver James is unable to enter into any personal correspondence.