Device Register Revisited

As I’ve mentioned before, talking to our devices can be tricky. That’s a relatively recent development; an older problem is how our devices talk to us. Siri and Alexa actually sound normal enough to not be jarring—if they correctly interpret our request—and I would expect that to continue to steadily improve.

But there’s still a need for getting printed communication right, and it seems there will “always” need to be a human to look closely at text rendered on the screen to help. A classic example for those of us of a certain age is the dreaded “syntax error” message of early computer systems. Frustrating and opaque, there is almost no information contained therein: the programmer may as well have left off the word “syntax” and just had the single word “error” displayed instead. It might have been less infuriating.

I spotted a somewhat more amusing example at a grocery store recently. At the self-checkout, the display admonished me to “Come on, Scan and Bag it!” This could be taken two different ways. One would be as enthusiastic encouragement to participate in the grocery shopping process; to get into the spirit of commerce and joyfully ring up those eggs, asparagus and bell peppers and get them into a bag so I can go home and make a frittata! Oh, boy!

But another reading might feel like you’ve been transported to a busier, perhaps slightly grittier grocery store, where even the machines are in a hurry. One might feel like the message to “Come on, Scan and Bag it!” could easily be followed by “I don’t have all day!” or “There’s other people in line, pal!”

The difference, of course, is tone, which isn’t something that our devices have a lot of flexibility with yet when they are speaking. (That will be an entirely different and possibly much more amusing chapter in voice assistant development.) But it’s also something that requires much more care any time that text is displayed that you might ordinarily simply say out loud. At a time that everyone is forming text messages exactly the same way they would speak, the mistaken assumption is that you know how I sound in my head just by the words I’m typing. Perhaps that’s why we lean on emoji to ensure the other person knows what we mean. ;)

Rough Translations

Foreign films have the challenge of being understood in more than one language. Subtitles are problematic gateways to these films because 1) translation is just plain tricky, and 2) they are, of course, written—they’re not quite speech from another language. I won’t dwell too much on the first point, but suffice it to say that translation isn’t just matching up words from one language to another. Languages say things differently—sometimes much differently—than others.

The ears are working overtime to attempt to understand what’s being said—or how it’s being said—whether we want them to or not. You can hear the actors’ inflections and tone, but you can’t count on really understanding the shades of meaning they convey because culturally and linguistically they don’t always match up with what we expect. Luckily, context helps and we can watch the actor’s behavior, too. Maybe that’s part of why we sometimes gesture when we speak. Gestures may have been our first language and now they’ve become a supplement. Or maybe gestures are a side effect of speaking?…

The eyes are working overtime to take in the whole scene, going back and forth between the action and the subtitles. Luckily, the average adult can read around 250 words per minute and the recommended talking speed for comprehension is about 150 words per minute, so there’s usually enough time to keep up with the subtitles. (A different, but fascinating phenomena is that human languages all cluster around nearly the same rate of spoken information per minute: other languages just sound a lot faster.)

During the film festival that just wrapped up we watched a comedy about a couple who did voiceovers of films when they lived in Russia, before they emigrated to Israel. The film is in both Hebrew and Russian, but the subtitles gave no indication which language was being used, so it was sometimes hard to tell when something was funny because the couple wasn’t fluent in Hebrew or if it was just funny. There was an Australian film that I wish we had subtitles for because it was often hard to understand the dialog. In fact, we’ll often turn subtitles on for British and Australian films and TV shows—just so we don’t miss any words or phrases that we aren’t familiar with. (And, to be honest, sometimes we even turn subtitles on for American films…)

Tangentially, it’s interesting to note that Shakespeare’s plays are deceptively challenging for modern English speakers because they are in a distinctly different dialect, if not quite a different language. The biggest problem being that the meanings of many words have drifted since then: the way that real people use certain words in everyday life have changed so much that we fool ourselves a little into believing we understand Shakespeare’s characters, even when we don’t quite. Word-for-word subtitles would not solve this problem nearly as well as a modern translation.

But even reading the text of my native language being spoken isn’t the same as listening to it being spoken. And reading the subtitles while listening to the actor’s rhythm and intonation and trying to re-assemble them in real time is work—and not always satisfying. It just doesn’t land the same way that spoken language does.

But all of the complexities and difficulties aside, subtitles are still enormously powerful. They give us a way to understand stories and ideas we might not otherwise have access to. They can entice us to pay a little more attention, tricking us into more carefully considering things we think we recognize. They emphasize a perennial question of life as well as film: what is going on in this scene?

Dinner Conversations

Two areas in which humans have advanced far beyond any other species are food and communication. We prepare food like no other animal not only by cooking, but by using an elaborate set of tools and techniques. And, of course, our language abilities have developed to an almost inconceivable level when you consider the range of complex and abstract concepts we can efficiently discuss; like what to make for dinner tonight.

We are not unique in either eating communal meals or our ability to communicate. We do stand out, however, near the edge of a certain precipice with another trait: the danger we face by talking during dinner.

The anatomy of most animals’ throats are such that an extremely important flap of tissue—the epiglottis—is directly on top of the windpipe (larynx) and closes it off completely from the rest of the throat throughout the entire act of swallowing. We humans, however, have a longer-than-usual passage above the epiglottis and the air that we breathe can carry with it the food and drink that we’re swallowing. Where animals have a railroad-type switching apparatus to keep food moving along the right track, we have more of an on-ramp/merge/off-ramp highway traffic arrangement; there are chances for both collisions and missing an exit, so to speak. In other words, if we attempt to both inhale and swallow simultaneously we may actually succeed at both—and immediately choke on our food.

Humans develop this “low larynx” after we are born, which creates a physiological hazard but also enormous acoustic and linguistic advantages. Other primates and human babies simply cannot make the same kinds of robust vowel sounds that adult humans can; vowel sounds which are found in almost all languages and are thought to be important for being clearly understood: the /i/ in “heed”, /u/ in “mood” and the /a/ in “palm”. Despite the fact that even today in the United States, choking is the fourth leading cause of unintentional death, evolution has apparently gathered more than enough benefit from human speech to offset the losses.

The simple, yet profound pleasures of cooking and sharing a meal with friends and family—talking and laughing in the kitchen and around the table—are quintessentially human. And, uniquely, they encourage a certain discipline: the subtle refinement in behavior of making a separation between taking a bite and taking a breath to compensate for an anatomy that no longer does.

Striking a Chord

I’ve noticed that over time it’s become less common to use the term ‘vocal cords’ and much more common to use ‘vocal folds’ instead. This seems fair, since I used to envision the apparatus as an array (or perhaps a bundle?) of thin ropes when it was ‘vocal cords’ and now I think of them as… well, something else. They’re nearly impossible to describe, and even when you look at several pictures of them they are difficult to comprehend at first. But ‘folds’ doesn’t immediately come to mind and they certainly don’t look like cords. I guess when I hear the word ‘fold’, I think of laundry and something that’s actually folded over, not just flaps of material. In any case, I think they kinda look like lips. (I don’t think there will ever be a movement to use the term ‘vocal lips’, but the description makes more sense to me, despite the somewhat awkward imagery and potential for confusion. Nonetheless…)

I also used to think that the misconception was simply caused by homophones: ‘cords’ and ‘chords’ sound exactly the same. A cord can be considered, among other things, to be a rope except not quite as stout. The vocal folds are like lips (to me), and so they are decidedly not like ropes (to me). However, a chord in the field of geometry is a straight line that joins any two points on a circle, the diameter being a very special case. If you consider the lips on the inside of your throat (yup, still a little weird) that can open and close, you might come to appreciate that each of the vocal folds is attached to the side of the throat, forming a chord instead of a cord.

Never mind that no one writes ‘vocal chords’, and never mind that ‘chord’ comes from the Latin chorda (rope) which brings us neatly back around to the wrong idea about what vocal cords actually look like.