A 1967 episode in the first Star Trek series, entitled “Arena,” featured Captain Kirk in a deadly competition with the captain of a starship of reptilian creatures known as the Gorn. The Gorn and the humanoids (earth people and Vulcans) were about to enter a new region of the universe. To determine which of the two races of intelligent creatures would be allowed to venture into a new area of the universe, a race of even more intelligent extraterrestrials, annoyed at the violent propensities of the two civilizations about to enter their territory, whisked Captain Kirk and the Gorn captain to a desolate planet, resembling parts of Arizona, to decide the question by requiring them fight to the death.
The even more advanced race of extraterrestrials, whom viewers never see but only hear, had the foresight to realize that the Gorn and Captain Kirk would need to speak to each other. Each captain was provided with a device that would instantly and accurately interpret each captain’s utterances to the other. The device was small and handheld, and resembled something between a cellphone and a microphone. It might not be an exaggeration to assert that this single episode is to blame for the naïve assumption nurtured by many people who grew up watching the series (and by the “next generation”!) that technology of that type is just around the corner. The fact is, it continues to be just around the corner.
So, while the answer to this question is simply and definitively “no,” it deserves an explanation because, in the popular media, one often encounters glowing prospects for the field of artificial intelligence (A.I.). Depending on the application or purpose for which A.I. is designed, some of these are reasonable claims. For instance, checking in luggage at the airport or paying for groceries at a supermarket involve many processes that A.I. could do and which, if implemented, would result in the elimination or major reassignments of employees in the travel industry or your friendly cashiers at a local supermarket.
It probably is true then that jobs involving repetitive, predictive tasks now performed by people may soon be replaced or greatly modified by the intervention of A.I. However, claims about the capacity of A.I. to mediate human communication involving any conversation beyond the rudimentary transactional conversations between speakers of distinct languages remain in the realm of science fiction and can be dismissed as hyperbole.
It is quite reasonable that anyone curious about what technical experts have accomplished in the field of computer-assisted translation (CAT) should ask the question “Will computers replace Analytic Linguists?” After all, the media often reports projections about how many jobs that exist now will be taken over within a few short years. People with bilingual talents may fear that they could made obsolete by technology. There is nothing to fear unless decision makers, ignorant about the nature of language while knowing so much about other things, may act on this belief in various ways.
One situation I am familiar with occurred in a university whose administrators saw a glowing article about hand-held interpreting devices and urged the language faculty to revamp their curriculum in response to this eminent technological advance. The administrators admonished some language faculty by saying that failure to keep up with the times might result in the elimination the foreign-language program. Technology, they asserted, was about to solve all the problems of translation once and for all. I mentioned the Gorn, the hand-held device and the fact that Star Trek is set in the 23rd century, but administrative enthusiasm for this coming age was childlike. They seemed confident that they knew all about this industry.
From another ivory tower, one voice of reason and measured enthusiasm was heard in an article in New York Times (Nov. 5, 2018). Dr. Melanie Mitchell, a professor of computer science at Portland State University, wrote an article entitled “Artificial Intelligence Hits the Barrier of Meaning,” in which she sets the record straight about the limits of A.I. by getting to the heart of the issue.
In a nutshell, Dr. Mitchell reminds her readers that computers don’t think and they cannot analyze the vast volume of sensory data that humans make sense of, understand and can communicate to each other in an instant. She says “Machine learning algorithms don’t yet understand things the way humans do — with sometimes disastrous consequences.” I am looking forward to reading her book “Artificial Intelligence: A Guide for Thinking Humans,” which will be published in 2019 by Farrar, Straus, and Giroux.
It also may be illuminating to analyze the term “Artificial Intelligence” from a semantic perspective. The term is an oxymoron: intelligence cannot be artificial. The etymology of “artificial” reveals that it means that something is “made by means of art” – something produced by people and not occurring naturally, such as a Michelangelo’s David as opposed to the sandstone arches in southern Utah. The intelligence (or lack thereof) resides and remains in the maker, in the techie. The techie, not the machine, thinks and is capable of meta-cognition. The techie is aware of his or her own knowledge and consciousness. A computer can only mimic the responses that we associate with consciousness, but the computer does not consciously respond. Humans programmed them.
Epistemological and ontological arguments about the nature of consciousness and free will, the mind-body connection and so forth are interesting. These also have a time-honored place in intellectual discourse since the days of Plato. Pragmatically speaking, for translators and interpreters, and for those who need them, those discussions are irrelevant when and where the boots hit the ground. Let’s see what “boots on the ground” means in terms of the tasks Analytic Linguists perform hours on end and in changing contexts.
Working in support of criminal investigations, Analytic Linguists must be able to recognize and distinguish subtleties of human voices. This is to properly attribute utterances to the ones making them. Computer programs using voice recognition software only can graphically represent patterns. But the human ear already knows just by attentive listening.
Analytic Linguists need to transcribe (write down) a verbatim record of real human speech – colloquial speech, often of uneducated speakers. Computer programs cannot type from voice input the vast varieties and anomalies of multiple speakers engaged in overlapping and colloquial speech – not even the highly useful program Dragon Speak (which an individual user has to “train” to understand the voice of the owner).
Analytic Linguists must monitor conversations in real time and extract what Dr. Mitchell rightly insists is fundamental to human communication – the meaning of a conversation (which often is stated in ways that are intended to deceive or mislead listeners). Computer programs designed for translation or interpreting are incapable of making inferences about meaning based on irony, humor, sarcasm or any number of rhetorical strategies people use to communicate not just a message, but their attitudes about their message.
Lastly, Analytic Linguists must produce an accurate translation, preserving indications of register and the personalities of the original speakers. Computer programs cannot translate a joke or figurative language of any kind (from source-language text to target-language text). Consider that even if a computer program wrote poetry, it would not be able to explain its own work. A poem by a computer is a parlor trick wherein the programmer is the Great and Terrible Oz standing behind the curtain.
In case you were wondering, in the Star Trek episode “Arena,” Captain Kirk spared the Gorn star ship captain, just when Captain Kirk could have killed him. The super intelligent race expressed surprise at Captain Kirk’s sudden display of compassion. They said his behavior gave them hope for the future of humanity.
As for the reliability of any hand-held interpreting device or CAT program, we should keep in mind that years later, and in movie depicting events in a galaxy far-far away, Darth Vader reminded the captain of the Imperial Star Cruiser not to place too much faith in the technological marvels they had created.