Skip main navigation

Testing for subjectivity

How do we know whether something else has a mind? Watch as Mark Solms describes the Turing Test and some of the problems associated with this approach
So we need an objective criterion for determining whether something else other than ourselves has a mind or not. There was a famous mathematician, Alan Turing, during the middle of last century who came up with an objective criterion. It’s called the Turing test. In the Turing test, you’re not allowed to see the being that you’re interacting with. You’re only allowed to communicate via a computer terminal. So you type your questions to the entity that you’re interrogating to determine whether or not it has a mind, and it responds to you on your computer screen. Let me give you an example of how the Turing test works.
And remember, what you’re trying to do is to decide whether the entity that you’re communicating with has a mind or not. The example I’m going to give you is of somebody called ELIZA, who’s a psychotherapist. But because it’s the Turing test, the only way that you can communicate with ELIZA is via a computer screen. And I’m going to read to you what my friend Oliver Turnbull– what happened in his interactions with ELIZA. He logs on, and ELIZA says to him, “How do you do? Please tell me your problem.” Oliver says, “Why can’t I be happy?” ELIZA says, “Do you think you should be able to be happy?”
Oliver replies, “I think I deserve more happiness than I have at the moment.” “Do you really think so?” “Yes. Life doesn’t need to be so cruel.” “You seem quite positive about that.” “I suppose I am, at least some of the time.” “Is that why you came to see me?” “No, I came because I need help.” “Is that the real reason?” “Actually, I came because I heard you were a computer and not a real therapist.” “Hmm. Do computers worry you?” “Hmm. Can I ask a question?” “Whether or not you can ask a question depends more on you than on me.” “The question is, how do mind and brain relate?” “I’m not sure I understand you fully.” “It’s the hard problem.
How does mental life arise from physiological events?” “Hmm. Please go on,” says ELIZA. Now, I don’t know what you make of that interaction. But a lot of people think that ELIZA has a mind whereas, in fact, ELIZA is an extremely simple computer programme. There’s some basic rules that the programme follows which gives these kind of stock psychotherapy phrases like “Please go on” or “Does that worry you?” and turns your questions into sort of reflective statements. It’s the school of Carl Rogers, that type of psychotherapy.
I, when told that people believe that ELIZA has a mind and then I’m told that, in fact, she’s a simple computer programme, I’m not convinced that the Turing test has worked. Or, to put it differently, I’m not convinced that ELIZA has a mind. Why is that? Well, I think this is stuff that we’re going to have to come back to. For one thing, I think the test is not really fair. It’s not really the right way of going about doing it, because you have to communicate verbally. You can’t, for example, do the Turing test on a dog. You can’t do it on any other creature other than human beings who talk.
Does that mean because they can’t answer questions they don’t have a mind? I’m not sure that’s what Turing intended. But also, why not see her? What’s wrong with seeing her? Why do you have to exclude information in order to determine whether or not somebody has a mind? And closely related to that, does it really not matter whether you have a body or not? Because that’s really what you’re trying to exclude. That’s what Turing was trying to do. He didn’t want the person doing the Turing test to be able to see whether they were interacting with a computer or not.
But I think seeing whether you’re interacting with a computer or not might be important in terms of determining whether or not the thing has a mind. Most important of all though is that the question I think the Turing test frames is the wrong sort of question. It starts in the wrong sort of place. It’s not about behaviour. And with the Turing test, which was developed during the height of behaviourism, that’s what you’re restricted to. It’s that same old thing. If you’re going to do this objectively, then you must only be able to measure the behaviours, to judge the behaviours. You can’t get to the internal states. You can’t get to the subjectivity. Does ELIZA have subjectivity? That’s really the question.
I think that’s the starting point. It has to be the starting point for any sort of objective criterion for determining whether something else has a mind. The behaviourist approach to psychology excluded the psyche because it was subjective. It wanted to focus on objective things, and the Turing test was based in this approach. Now, there was a great philosopher, Rene Descartes, in the 18th century who famously said, “I think, therefore I am.” What that statement was all about was– Descartes, he was a troubled man and had lots of doubts. In fact, he doubted everything. And his philosophy was about trying to determine what does he not have to doubt. How can he be sure that something really is true?
And after years of philosophising and doubting, he finally came to the conclusion that the one thing that he could know for certain was that he existed because he thought, because he had experiences. “I think, therefore I am” refers to the fact that the one thing that Descartes can be sure of is that he exists, and the evidence for that is he is busy having experience. Now, think of that. Descartes were saying the only thing that we can be sure of that it definitely exists is our subjective experience. And yet, that is what behaviourism sought to exclude from the science of the mind in order to be more objective.

In order to perform scientific experiments on the mind, we need an objective criterion for determining whether something else has a mind or not.

Mathematician, Alan Turing, designed the Turing Test for this very purpose. Alan Turing is sometimes referred to as the father of Computer Science and Artificial Intelligence because of the work he did on describing the types of problems a machine could theoretically solve. Such a machine is similar to a present day computer. Alan Turing also imagined intelligent machines. His test involved a human judge engaging in conversations with both a person and an ‘intelligent’ machine. The conversation would be conducted via a computer keyboard and screen so the judge remains unaware of the machine’s inability to speak. If the judge cannot reliably tell the machine from the human, the machine is said to have passed the test. The test does not focus on the correctness of answers to questions, but rather the judge is considering how closely each answer resembles the answer a human might have given.

In this course we are considering related questions, but drawing on more recent understandings of the mind. Since, as I argued, the starting point of a mind is that it has subjectivity, a flaw of the Turing Test would be that it excludes this defining property of the mind. It also excludes non-verbal communication and physical appearance, which are important for determining whether something has subjectivity. Alan Turing was rooted in a behaviourist way of thinking that was widely held at the time.

In recent demonstrations of advanced computers’ ability to be competitive in quiz shows, chess and other human activities, there was no need to hide which participant was the machine. Given that the Turing Test does not consider subjectivity, how might one determine whether something else has subjective experiences?

This article is from the free online

What is a Mind?

Created by
FutureLearn - Learning For Life

Reach your personal and professional goals

Unlock access to hundreds of expert online courses and degrees from top universities and educators to gain accredited qualifications and professional CV-building certificates.

Join over 18 million learners to launch, switch or build upon your career, all at your own pace, across a wide range of topic areas.

Start Learning now