Shakespeare’s Othello, the Turing Test for Artificial Intelligence, and the indeterminacy of radical translation

It took close to 350 years for mathematics and philosophy to catch up with the ideas about the human mind explored in Shakespeare’s classic tragedy.  From computer science to the study of language, the opacity of other minds remains at the forefront of our understanding of each other and at the center of Iago’s scheming to take down Othello.

In 1950, the mathematician and computer science pioneer, Alan Turing, proposed a simple test to determine whether or not a computer should be considered intelligent.  The test, known fittingly today simply as the Turing Test, was conceived as a conversation between man and machine. This conversation would be conducted through a text-interface or some other channel to prevent a human evaluator from being aware they were interacting with a machine.  The evaluator would ask the computer questions to probe whether or not the responses displayed the full range of characteristics associated with human intelligence.  The goal was not for the machine to display a command of specific facts, or be able to answer any questions we might pose like a modern search engine.  Instead, it was to test how closely it matched the responses a human would give to the same or similar questions.  The machine would be considered truly intelligent if the evaluator couldn’t say for sure whether they were interviewing a human or a machine at the end of the test.  Mr. Turing referred to this as the “imitation game,” a phrase that provided the title for the award-winning film about his exploits to crack German code in World War II.

Fundamental to the Turing Test is our inability to look inside another mind and actually see what they are thinking, a fact often known as the opacity of other minds.  We can only access the outward show, believing others are intelligent and self-aware because we are and, within reason, others respond in a similar manner to a similar stimulus.  Mr. Turing himself was aware of this when he first proposed the test in a paper titled “Computing Machinery and Artificial Intelligence.”  He began with a simple goal, though one with almost unending underlying complexity, “I propose to consider the question ‘Can machines think?’”  The challenge is that both “thinking” and “machine” are notoriously difficult to define, except to say that we ourselves are thinking machines in the physical world.  Mr. Turing proposed to answer that question in a different way, “Can machines do what we (as thinking entities) can do?”  Therefore, if a person cannot determine whether they are dealing with another person or a machine, the machine should be considered intelligent.

Mr. Turing’s proposal and resulting conclusion was considered controversial at the time and remains so today.  In 1966 and 1977, computer programs were created that some have claimed pass the Turing Test without actually doing anything remotely “intelligent.”  Instead, these programs scanned the question for keywords and then returned generic phrases using those words or referenced an earlier portion of the conversation.  These programs were able to fool at least some people into thinking they were talking to another human being.  In modern parlance, we would refer to such programs as “chatbots” and we interact with them all the time.  For basic queries like the status of an order or shipment, it can be difficult to tell whether a human or robot is on the other end of the line, but Mr. Turing himself placed no limits on the extent of the evaluator’s questions.  It is to be supposed that he imagined a much more in depth “interrogation” than simply chatting about the weather, something much closer to the test for a replicant in the science fiction classic, Blade Runner.  We should also assume he imagined these evaluations to be wide ranging, from asking the computer what they think about Hamlet to describing the feeling of something.

In my opinion, the most serious challenge comes from the philosopher John Searle who proposed a thought experiment in 1980 known as the Chinese Room.  He asked us to consider a computer program that “understands” Chinese.  Similar to modern chatbot programs mentioned above, the machine takes Chinese characters as input and responds in Chinese.  It does this well enough to pass the Turing Test, or at least a limited version of it.  The question then becomes, does this computer actually understand Chinese or is it merely simulating an understanding?  Mr. Searle asked us to consider what would happen if a non-Chinese person played the role of the computer.  This person receives the input with no knowledge of Chinese, follows the instructions of the computer program in English or another language, and then produces the output the same as a computer.  In principle, if you excluded the time it would take to perform these tasks, the person could execute the program in the exact same way (another theory of Mr. Turing’s when he conceived of modern day computers), but no one would claim the person executing these tasks actually understood Chinese.  From there, Mr. Searle concluded that the Turing Test was fundamentally flawed because behavior can be simulated without understanding or intentionality.  This no doubt true, but the conclusion has been misinterpreted in my opinion. Rather than undermining the Turing Test, this actually proves Mr. Turing’s original point:  We do not have access to the Chinese rooms inside our own minds, much less those of others.  The operations our brains use to understand language and extract meaning are entirely unconscious.  We cannot peer into them to verify what is actually happening.  Instead, we can only assume understanding based on the outward show.

Mr. Turing wasn’t the only mid-20th century thinker building new theories about how we understand each other and the world based on the opacity of other minds.  William Van Orman Quine proposed a breakthrough theory in our understanding of language in 1960, the indeterminacy of radical translation.  He asked us to imagine being charged with translating a never before encountered language, what he called a radical translation, meaning the translator doesn’t know a single word of the language they are asked to translate and the native speakers do not know a single word of the language spoken by the translator.  Mr. Quine reasoned that the translator had no means with which to peer inside the speaker’s mind to determine what a word specifically meant.  All they had to work with was the outward show, that is what would prompt a native speaker to utter a word.  He called this “stimulus” meaning.  For example, a native speaker says the word “rabbit.”  The stimulus meaning would be all the circumstances that prompted them to use the word, from actually seeing a rabbit to cooking a rabbit in a stew, even the signs a hunter might see while tracking a rabbit.  The translator would then assemble these different stimulus meanings into his translation manual.  Mr. Quine asked what would happen if another translator arrived and performed the same process.  This second translator would not be able to recreate the exact same circumstances to prompt the exact same utterances.  Their list of stimulus meanings would necessarily differ at least somewhat, and their final translation manual must do so as well.  Therefore, the two translation manuals would have different outputs for the exact same language, and would not translate them the same.

This logic, however, can also be applied to native speakers using the same language.  You cannot look inside my head and “read” my personal definition of anything, from a simple rabbit to a more complex subject like love.  The only thing you have to go on is our shared stimulus meanings and opacity of other minds is once again at play.  We take it for granted that we mean the same thing when we speak the same language because our list of stimulus meanings will be almost exactly alike, but they will not be precisely the same, nor do we have any way of knowing how similar they are in advance.  Perhaps needless to say, Mr. Quine’s conception is not without critics.  Similar to Mr. Turing, these critics generally cling to the concept that under the outward show there is some way to determine inner meaning without truly demonstrating how that would work in a world where we are literally trapped in our own minds.  Ultimately, no one has proposed how to pierce that veil, nor is it clear such a thing is even possible in the first place when we know that qualia, the feeling of what happens, are unique to each individual.  In other words, your mind is not mine.  It’s similar in a lot of ways, but also different, meaning shared experiences are just that, shared, and not united on some fundamental level that we can call universal meaning, especially when minds cannot access each other directly, only through speech, words, and non-verbal cues.

The concepts described above are relatively modern conceptions, 20th century philosophy and ultimately computer science, probing questions about what makes us human, what it means to be intelligent and self-aware, and how we can recognize those traits in other beings.   Shakespeare, however, was putting the concept of the opacity of other minds to use hundreds of years earlier.  It is in fact the driving force behind his tragic masterpiece, Othello, which was first performed in 1604.  Othello is often described as a play about jealousy, what the text itself describes as the “green eyed monster.”  The villain, Iago, uses Othello and Rodrigo’s jealous nature and feelings for Desdemona to bring about Othello’s destruction, while he is motivated himself because he is jealous of Cassio’s promotion over him.  This conception, however, is both too broad and too specific at the same time.  Broad because Iago takes advantage of a certain kind of jealousy in his manipulations of Othello.  We all experience envy at things others have that we want, from a nice house to a fancy car to an attractive partner, but these are things that are not in our possession.  The jealousy Iago wields against Othello is subtly different however, and far more destructive.  Othello is married to Desdemona and has no real reason to doubt her faithfulness, yet once Iago convinces him she is having an affair with Rodrigo, he does and it burns inside him like a poison, “Thou hast set me on the rack.  I swear ’tis better to be much abused Than but to know ’t a little.”

At the same time, the conception is also too specific because the kind of jealousy Iago takes advantage of is ultimately a product of the opacity of other minds.  Othello has no direct means to determine whether or not Desdemona’s feelings for him are true, allowing Iago to exploit this weakness to a tragic end.  While laying his trap, Iago questions Othello specifically about honesty in general.  First, he asks what role Cassio played while Othello was courting his wife, a conversation that starts innocently enough.  “Did Michael Cassio, when you wooed my lady, Know of your love?”  Othello replies, “He did, from first to last.  Why dost thou ask?”  “But for a satisfaction of my thought, No further harm.”  This, of course, plants the initial seed, prompting Othello to ask Iago whether or not Cassio is honest.  “Indeed? Ay, indeed! Discern’st thou aught in that?  Is he not honest?”  The precise opening that Iago expected:

IAGO
Honest, my lord!

OTHELLO
Honest! ay, honest.

IAGO
My lord, for aught I know.

OTHELLO
What dost thou think?

IAGO
Think, my lord!

OTHELLO
Think, my lord!
By heaven, he echoes me,
As if there were some monster in his thought
Too hideous to be shown. Thou dost mean something:
I heard thee say even now, thou likedst not that,
When Cassio left my wife: what didst not like?

Honesty, of course, is another manifestation of the opacity of other minds.  We trust on faith and experience, lacking access to a direct means to ascertain the worthiness of the person.  Iago is a masterful liar and manipulator.  Cassio is an honest man.  Othello cannot know this for sure, and thus Iago is able to convince him of the opposite, often by warning him of the consequences of allowing him to speak his mind and pretending that he’d rather stay silent.  He tells Othello he thinks Cassio is honest while leading him to believe he isn’t, “For Michael Cassio, I dare be sworn I think that he is honest.”  He also tells him that  “Men should be what they seem;  Or those that be not, would they might seem none!”  He even warns Othello about jealousy in particular, telling him “O, beware, my lord, of jealousy;  It is the green-eyed monster which doth mock The meat it feeds on.”   While it might seem easy to blame Othello for his failings, there is barely a single character in the play that doesn’t fall victim to Iago’s schemes.  He manipulates Rodrigo against Cassio by convincing him that he can have Desdemona for himself.  He manipulates Cassio by first setting him up to take part in a drunken brawl, and then convincing him that he can help restore him to Othello’s good graces.

In this sense, scholars have often claimed that Iago can be seen as the author of the play.  There is barely a scene that occurs without his contrivance.  He both plants the seeds in the other characters’ minds, and positions them as though they were pieces on a chess board.  The play is most definitely his, but that can only be so because of the opacity of other minds.  In fact, the audience is the only participant in the drama sufficiently aware of Iago’s true nature and motivations until it’s too late.  He lays out his entire plan very early in the play:

I hate the Moor:
And it is thought abroad, that ‘twixt my sheets
He has done my office: I know not if’t be true;
But I, for mere suspicion in that kind,
Will do as if for surety. He holds me well;
The better shall my purpose work on him.
Cassio’s a proper man: let me see now:
To get his place and to plume up my will
In double knavery…

Iago even adds that Othello “is of a free and open nature, That thinks men honest that but seem to be so, And will as tenderly be led by the nose As asses are.”  Ironically, Iago would have succeeded in his diabolical plan save for his failure to know his own wife’s mind about Desdemona and her affection for the woman above even her own and her husband’s interests.  He is ultimately undone by knowledge he should have had himself, of the person he should’ve been closest to, meaning he himself falls victim to the opacity of other minds.  Shakespeare himself seems to make this last point clear, both for Iago and all of us, with Iago’s final words in the play, when he refuses to speak further:

Demand me nothing. What you know, you know.
From this time forth I never will speak word.

“What you know, you know.”  It might be impossible to better encapsulate the opacity of other minds in a single statement, much less one that uses only one-syllable words.  From the driving force to one of the greatest tragedies ever written, to cutting edge theories on the nature of intelligence and language, the opacity of mind remains as important today as it was in Shakespeare’s day, a barrier that might never be breached, perhaps nor should it.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s