It takes her a long time to get it out ("Ah hem, the theoree by Anne Elk - that's not an elk but Anne Elk ..." etc), but eventually she states: "My theoree is that dinosaurs are small at one end, wider in the middle and smaller again at the other end, and that's my theoree, and it's mine."
I hope that the reader will not feel that I am suffering from Elkish obviousness or prolixity but I also wish to put forward a theoree. It explains why violence keeps increasing when property crime has been plummeting for five years.
Three days before the last election I had the joy of putting this theoree to an exhausted and dispirited Michael Howard on Newsnight. Probably because he had already had a terrible kicking during the campaign, he did not even try to interrupt me.
Jack Straw was also there, but if he had a different idea he was keeping mum. Perhaps wisely, now that it is his turn to explain these things, he leaves it to experts. It was the head of the Home Office Research Unit, a distinguished researcher called Chris Nuttal, who provided the gloss on the latest batch of statistics, released last Tuesday.
He announced that cases of violence against the person, which make up three-quarters of violent crime, had risen by 14,000 in 1997 from 1996. On the other hand, theft (which accounts for nearly half of all crime) had fallen by 9 per cent.
Nothing switches me off more than large numbers of numbers in a newspaper article, and, for most of you, this is probably truest of all in disputes about the supposed meaning of crime statistics. I can only beg forbearance in what follows.
The Home Office's simplest tactic has always been to emphasise percentages rather than raw figures. In percentage terms, the latest rise in violence does not seem all that spectacular - about average for the period since 1950 of steadily increasing crime. Last year's rise in violence against the person was "only" 5.8 per cent.
However, raw figures paint a different picture. In 1950 there were 6,000 crimes of violence against the person recorded, whereas last year there were 253,000. A 1 per cent increase in 1951 would have required 60 more crimes but in 1997 it required an extra 2,530 - 39 times more.
The key comparison is between the periods 1980-86 and 1987-1997. Each year between 1980 and 1986, there was an average of 4,000 more crimes of violence against the person. But for the 10 years between 1987 and 1997 the average annual increase was 13,000 - more than three times as much (for anyone wanting still greater detail, I can heartily recommend OW James's Juvenile Violence in a Winner-Loser Culture, written in 1995 and published by Free Association Books).
What, then, I hear you cry, is your bloody theoree? Ah hem. It has two prongs.
Violence is caused fundamentally, by being male (85 per cent), young (75 per cent under 29) and from a low-income family (about 40 per cent of boys from the poorest families have been violent by the age of 32). We know from studies of twins and adoptees that the difference between violent and non-violent poor young men is not their genes. Rather, it is patterns of childcare.
The violent are more likely to have had three particular childhood experiences: severe and frequent physical abuse, usually dressed up as legitimate punishment and always administered in erratic fashion (yesterday's punished behaviour is rewarded today); witnessing parental disharmony, often physical, usually dad hitting mum; and irritability, usually in both parents, often appearing as depression in the mum.
None of this is disputed by most criminologists, and neither is the fact that parents living with the stress of low incomes (defined by any of the accepted methods) are more likely than the well-off to care for their children in this violence-inducing fashion. Indeed, given the pressures of childcare without a washing machine, car, telephone and with constant money worries, the only surprise is that more low-income parents are not like this.
The crucial next step in my theoree is as follows: if you increase the proportion of boys being raised in low-income families, then a few years later there should be a rise in the proportion of men who suffered the kind of childcare that leads to violence ... and increased violence statistics.
Now consider the following: in 1979, about 20 per cent of boys were raised in low-income families. By 1981, the proportion was nearly 33 per cent, and it has stayed there ever since. Given the fundamental causes of violence (see above), is it any wonder that the violence statistics (for juveniles as well as adults) began an unprecedented surge a few years later, in 1987, lasting to the present day? Thatcher's children have come out to play.
The second prong to my theoree concerns the direct impact of Thatcher's policies on low-income males (rather than the indirect effect on their parents and, therefore, the kind of childhoods they had). The generalisation holds that the more unequal the nation, the more violent. Throughout the developed world, the greater the gap between rich and poor, the higher a country's homicide rate. Likewise, homicide rates are higher in countries that do not develop employment and training policies to keep unemployment low or that spend little on the health of the nation's work-force.
This even applies within developed nations. Across America levels of aid to families with children vary considerably from state to state. The higher the levels of benefit, the lower the level of violence in that state.
During the 1980s Britain became more unequal than at any time since the 1930s. Hence, prong number two is that increasing inequality has been like throwing paraffin on the embers of potential violence which prior childhood abuse creates.
Put bluntly, my theoree is that no amount of communitarian rhetoric about "morality" and "the need for poor people to take responsibility for their behaviour" will change the unambiguous findings of hundreds of scientific studies: violence is caused by the degree of the inequality that politicians deliberately decide is best for their society.
Such decisions are the main reason why grossly unequal America is five times more violent than relatively equal Europe.
Meanwhile, it was business as usual for Mr Nuttal at the Home Office last Tuesday, his line as unchallenged as ever, as if there had been no change of government. I saw no newspaper or broadcast reports suggesting that there were fundamental problems with his spin.
He came out with the same old superficial waffle, muddying the water just like he did when his boss was Michael Howard, evading the fundamental questions: how come violence has kept going up and up since 1987? Do the Home Office and its ministers have a clue or are they just pretending not to know?
During its first term, I do not expect New Labour to be much different from John Major in drag. I can even see the necessity for a spot of political cross-dressing. What alarms me is New Labour's preference for America over Europe as a role model.
We can but pray that, in its second term, New Labour rips off its Yankee couture and reverses the Americanisation that has been Thatcher's very violent legacy.
Oliver James's book 'Britain on the Couch - Why We're Unhappier Compared with 1950 Despite Being Richer' is published by Century at pounds 16.99.