If you'd like to participate in an interesting theological experiment let me point you to Leah's Christian versus Atheist Turing Test.To given some background, let me first explain a bit about the history of the Turing Test.
The Turing Test was the idea of Alan Turing. Turing was a famous cryptographer (he helped the British crack the German Enigma code), mathematician, and computational theorist. Many consider Turing to be the father of computer science and artificial intelligence.
Turing conceived of his test as a means to answer the question about when a computer program could be considered the functional equivalent of a human being. Turing's test goes like this. Imagine a human and a computer program hidden in two rooms, their identity unknown to a human judge. The judge, however, can communicate with both the program and human via a keyboard and screen. Basically, the judge can live chat with the program or human.
Given this setup the goal of the Turing Test is to have the judge interact with both the human and program at great length and in great depth. After the judge is satisfied she then guesses which room contains the program and which the human.
According to Turing, if the judge can't decide which room contains the human--that is, if the judge can't discriminate between the program and the human--then the program "passes" the test and, for all practical purposes, should be treated as a human being.
Now, you might not agree with the logic of the Turning Test, but Leah isn't applying the test to computer programs. She's applying it to atheists and Christians.
Leah's test was inspired by Bryan Caplan's notion of an Ideological Turing Test where conservatives and liberals are asked to simulate the positions of their rivals. Caplan's pushing back against Paul Krugman's contention that liberals can simulate conservative views but that conservatives can't return the favor. The issue here goes to one of empathy and perspective-taking. Who is better at seeing with the eyes of their ideological rivals?
That is an interesting empirical question. More often than not, in ideological debates we often see people react to "straw men" formulations of their rival's opinions. Is that just a way to score easy political points or does it reflect a failure of imagination?
Running with this idea, Leah has constructed an Christian versus Atheist Turning Test. The test has two parts with Christians simulating the answers of atheists and atheists simulating the answers of Christians. Here is the list of questions Leah created for the two tests.
Right now Leah is having a group of Christians simulate atheist responses to a series of questions. Mixed in are actual atheists' responses to the same questions. The empirical question is if a judge can tell the difference. That's where you can participate. Go here to read the various responses and then go here to cast your vote about which responses are from actual atheists and which are from Christians simulating atheism.
In the coming round the test changes with atheists trying to simulate the answers of Christians.
Should be a very interesting experiment.
This looks great. But I do have one question:
ReplyDeleteWhat if you're a Christian atheist?
like like like!
ReplyDeleteAlan Turing appears in an Experimental Theology post. My life is complete.
ReplyDelete<3 <3 Turing.
... I'm sorry for that double posting ... have NO CLUE how that happens ... I've tried several times to prevent it ...
ReplyDelete... I'M HALTING ...
Oh, good grief. I can't tell the difference between smug, one-note atheists and Christians who think that all atheists are smug and have only one note to play. (This is kind of embarrassing for everybody.)
ReplyDeleteHave I missed something? Is this not simply a measure of congruence - the ability to answer questions 'as if' you were another person without losing the quality of 'as if' - as Carl Rogers put it. This may tell us whether the 'average Christian' or the 'average atheist' is more empathic, but surely there will be much more variability within these groups than between them...?
ReplyDeleteIn Self-Organised Learning (SOL) practice, a standard measure of how well one person understands another on a certain issue is to take a series of elements (specific meaningful experiences e.g. a time when I felt the presence of God) and a series of related constructs (contrasting pairs of value descriptors e.g. holy vs. mundane) and to see how closely each person is able to predict on which pole of each construct each element will be placed - producing a grid of 1s and 0s. This strikes me as an ideal methodology for this kind of research (as well as a potential empirical measure of AI).
The other strength of SOL is in analysing differences. It seems to me that it will be in the discrepancies between held beliefs and construed beliefs that the value will lie.
Was this "wall of text" written by a computer? It seems too random (and too long) for a human response.
ReplyDelete"... this strikes me as an ideal methodology for this kind of research (as well as a potential empirical measure of AI) ..."
ReplyDeleteYes! I agree. (God - please let this go through without my former wall of text). The problem with this test on an AI level is that a real Turing test at that level – assuming the universe of atheist and Christian views is finite – we could do an arbitrarily long longitudinal Turing computation (but very, very looong) with enough tape in the Turing machine to identify all the Christian and atheist memes – but if we really did such a long Turing computation, then the whole tape would risk – risk – getting sucked into a black hole. This is not a Turing test!What the researcher did was swap out an arbitrarily long Turing computation for a flat and flatlander-like survey-feedback faux-Turing tape – with the survey-feedback type questions not extended arbitrarily – but (for AI purposes), even this flatlander Turing tape with flatlander answers risks moving away and away from the researcher (again for AI purposes) over loooong periods of time! – and real Turing answers to the questions of that test would risk – risk – moving beyond the cosmological (event-horizon?) horizon of the researcher (in a real Turing test done on flatland space)! All the data would be lost. For AI purposes. So the researcher used an arbitrary stop-function. For this test. It’s still a great test! It reveals mutual hypocrisy both ways I’d say. Reducing our Christian judgments (mea culpa - I include myself!) against atheists to biased sound-bytes! Clever. And well done. Cheers,Jim