Success at passing in the Imitation Game shouldn't be seen as confidence trickery, hoaxing, or any other kind of pulling the wool over the judge's eyes. Confidence tricks work by giving the judge -- or mark -- such a strong reason for wanting to believe the trickster that they subconsciously repair any mistakes in the performance. In such cases, the trickster can do a poor job and get away with it. In the Imitation Game, however, judges know from the outset that one of the players is pretending. They are alert to the slightest mistake and they ask difficult questions that can be answered only if the interactional expert has genuine expertise.
* * *
That the earliest imitation games compared men and women has greater significance than might be first apparent: Andrew Hodges' biography suggests that Turing had gender identity on his mind because he was gay. Back then homosexuality was a crime in Britain and, in 1954, at the age of 42, the brilliant Turing was hounded to suicide, eating a cyanide-laced apple.
In Turing's time, the lack of understanding of homosexuality among the wider population would have enabled few straights to pass as gay in imitation games. Nowadays, however, heterosexual people's increased knowledge and understanding of homosexual cultures would enable at least a few more of them to succeed. If only we could go back to the 1950s, multiple imitation games could be used to test this idea. We obviously can't, but something similar in fact is being tried.
Collins and his collaborators are developing a new and much more complex incarnation of the Imitation Game under sponsorship of the European Research Council. They run games on topics like pretending to be gay, pretending to be a Christian, pretending to be a member of an ethnic minority, and so forth, in different regions of Europe. The idea is to find out if the game can be used as a tool for measuring differences in the extent to which these groups are mutually understood in different societies.
At Cardiff University, students found it easier to pretend to be gay than Christian.
At Cardiff University, students found it easier to pretend to be gay than Christian. A measure, called the 'identification ratio', or 'IR', was developed to make numerical comparisons. The IR is right answers minus wrong answers divided by the total number of trials. For the blind pretending to be sighted the IR was 0.13; for the sighted pretending to be blind the IR was 0.86. For straight students pretending to be gay, the IR was 0.4; for secular students pretending to be active Christians, the IR was 0.7. This gives some indication of how secular a country Britain has become.
Though this result was striking, with the numbers of students involved, the difference between 0.4 and 0.7 isn't quite statistically significant. The research is now moving on from tests with quasi-controls where big differences could be expected, to cross-national comparisons of a single condition where differences are going to be smaller and far larger samples are required. Since playing large numbers of imitation games is time-consuming and difficult organizationally, the experiments have been taken apart and redesigned with the production line, ethos in mind. Initially, some games are played to produce sets of good judge-questions and good expert-answers. Then, the sets of questions are presented to much larger numbers of pretenders. Finally, the sets of questions and pretender-answers are recombined with original expert-answers, and sets of completed dialogues are sent to fairly large numbers of judges. This is a better way to test for differences between populations because the representative sample of pretenders is key. Initial results suggest that the technique may work, but it is in its early days.