Perhaps “Word Nerding through Netflix”

Read the background and objections here, then delve in to my POV.
Please note I’ve used quotation marks rather than italicizing words as words (in captions/subtitles) with the aim of making a more accessible document.

Spoiler alert: you aren’t going to learn a language with the “Language Learning with Netflix” browser extension. You may confirm what you know, learn the odd word, or see something spelled that you’d only heard before, but you aren’t going to learn a language.

Now, I’m a language nerd, so I’m not knocking different modes of language acquisition or people’s desire to expand their worldview or personal skills. But to get viewers’ hopes up by presenting this tool thus is like saying you’ll learn to be a chef by working as a cashier at McDonald’s. You’ll learn stuff, it may be fun—it may even be “cool” as the linked article says—but you won’t be able to converse in the original language of the show. Especially based on most of the subtitles.

There are indeed some very cool functionalities to this tool. You can choose to see the automatic voice recognition–software’s subtitle translation or the human translation or both. Most useful is the ability to set the automatic pause on each text box. Unfortunately, the two versions of subtitles are so poorly handled that there’s no way in Hades you could learn much language from them.

I experimented with a film and language I was familiar with: “Incendies” by director Denis Villeneuve (2010) in French—Québécois to be exact, which is no mean feat for a French-from-France-translator to tackle. (I know because I went to translation school and, being in Canada, we dealt with Québécois French as much as European French.) (And sidebar, it’s a difficult film to watch but excellent. I recommend it.) I chose to watch both the “machine” translations (as Netflix calls the autocraptions) and the “human” translations simultaneously. In the following examples, when all three languages are shown, the original French is on top, machine in the middle, and human on the bottom. The Arabic is not subtitled. And by the way, I could have selected a film in any of the languages I know and found similar issues; this is just an illustration.
Let’s look at some examples.

Longshot man and woman approaching their car in a city , captioned J'ai la crisse de paix/I have the crease of peace/I feel so fucking peaceful
“J’ai la crisse de paix” is not about “the crease of peace” or even, dictionary-wise, “the crisis of peace”: it’s swearing with “Christ” and colloquially would be used as in the human subtitled “I feel so fucking peaceful.” So, that part is good! If it were France, it would likely have been some form of “putain,” but it looks like the translator asked someone who was familiar with swearing as you’d find it in Quebec or perhaps Maritime Canada (because as we’ll see below, the rest of the translation is problematic). But how did a machine supposedly translate “la crisse” to “the crease” if they’re using a corpus dictionary? Autocraptions 0, #NoMoreCraptions 1.

Young man at side of woman in hospital stretcher captioned Souffle haletant/Breathless breath
“Haleter” means “to pant,” “gasp,” or “puff” in French, but for the moment, let’s look at the autocraption “(Breathless breath)” which a human has not chosen to correct. That’s somewhat of a grammatical nullification in English, never mind a contradiction in meaning. In this scene, the young man is upset, stressed. A good subtitle would have replaced the machine one with something like “(sighs with stress)” or “(anxiously sighs).” This subtitle is used many times in the film, unfortunately.

Doctor examining woman in hospital, adult children looking on, captioned Elle est absente en general/She is absent in general/She's usually confused
Here the doctor is taking a history of the woman and asking her children questions about her health and behaviour of late. The machine subtitle is typically autogenerated: it just translated the line literally. The subtitler is just wrong. “Absente” and “désorientée” or “confuse” wouldn’t be synonymous here. In fact, here’s an argument for giving captioners and subtitlers reasonable work timelines instead of ridiculous demands of urgency. Had the subtitler watched the film first, they would have known that the woman has PTSD, which is unknown to her children, so her son just finds her emotionally unavailable and is very hurt and angry about that. Therefore, the subtitle must be “She’s always absent”: the English audience would understand that it doesn’t mean just physically but more so emotionally; the always would be more colloquial than “usually,” and it would be understood as “[not literally] always” but “[pretty much] always.” So if someone went to a French class and used what they’d learned here and said to the teacher, “Je suis absente” to indicate they needed further help, they’d be laughed at. Not what you want when learning a language.

Young woman looking from desk skeptically at offscreen woman, captioned J'etais meme pas nee/I was not even born/You're kidding right? I wasn't born
Here the “You’re kidding right?” is a hangover from the previous shot/subtitle and shouldn’t even be included again. But the young woman is being asked about something from thirty-five years ago and predictably responds with the French line as shown. The machine version is literally correct but not idiomatically. The human translation is incomplete and misses the mark, thus leaving the viewer in the dark. “I wasn’t born” is not the same as a snarky “I wasn’t even born yet” or “I wasn’t even alive then.” So let’s imagine someone (for some reason) wanted to learn how to say “I wasn’t born” in French: they would use a totally incorrect/inappropriate construction, confusing their listener. Part of learning a language is about clarity, so that there is no miscommunication.

Woman looking distraught in the front of a bus, captioned Cris de fillette/Cree of little girl
The translation by machine apparently went for an aural equivalent here; a human should have changed this to “(cries of little girl).” Unfortunately, this one is doubly problematic in Canada: “Cree” is the name of the Algonquian language of the indigenous Cree people. Confusion could reign supreme here, especially in a film so much about culture and place. Furthermore, knowing it was a Canadian film, viewers might see the subtitle briefly, wonder at it, and then lose track after it has passed by but still be pondering the meaning: audience immersion down the toilet. It certainly would detract from the cultural aspect of learning French.

Longshot of open-doored car in the countryside, with a man pointing the way to a woman on the road, captioned Stridulations d'insectes/Stridulations of insects
This is a good example of the need to understand diction and register in audiovisual translation. “Stridulations” can mean “chirps,” “chirring,” or “shrill sounds.” It refers to the sound crickets and other insects make by rubbing their legs, wings, etc. together. In English, “stridulations” would only be used within a scientific context, perhaps even only an academic one. Here, it’s just about the countryside setting, and we would say “(insects chirring)”—if at all. There’s an argument that the caption is not even necessary as it doesn’t advance the plot: we can see it’s empty and remote. In any event, a language student who then said on a beautiful summer night in Provence, “Oh listen to the stridulations of insects!” would be looked at like they had three heads…or too big a head. Subtitling and captioning is not about dictionary and thesaurus use. The audiovisual translator has to understand meaning, context, and changes in the target language. For the record, I don’t believe the audio has insects: I think it’s birds and the wind.

Arab older man captioned C'est la Femme qui chante./It's the woman singing./She's the Woman who Sings. Number 72.
The machine definitely blew this one with its literal translation. This is a key thematic and character-relevant phrase and is even a chapter title in the film. The human was closer but the “Number 72” is repeated in the next subtitle. Also, there is no understanding of capitalization conventions: as an epithet and important theme, “the Woman Who Sings” needs a capital on “who” in headline style; here it’s a mixture of headline and sentence. Probably the subtitler is working under the misapprehension that “little words” don’t get capitalized, a rule from the dinosaur age. All caps on the phrase would forewarn a language learner that this is not everyday usage.

Arab older man, captioned Inspiration/Inspiration
Here, “Inspiration” (and elsewhere “(Grande respiration)” as “(Great breath)”) is a total craption. Inspiration comes from the Holy Spirit or a muse or a lightbulb above your head, but its Latin root about breathing cannot be applied here. Furthermore, it’s hardly a notable or significant inhalation (unlike an example below) and could have been omitted. I don’t understand how a professional translator or QC person could have stetted this machine error.

Vista with woman at wall and car in midground, chapter title La Femme Qui Chante, captioned Un homme parle en arabe sur un haut-parleur/A man speakes in Arabic on a speaker/THE WOMAN WHO SINGS

Vista with woman at wall and car in midground, chapter title La Femme Qui Chante, captioned Gazouillis d'oiseaux/Bird chirping/THE WOMAN WHO SINGS
The errors in these subtitles are obvious in that the chapter title denies access to the viewer of the other subtitles if using the human version. “La Femme Qui Chante” should have been made a forced narrative, and the correct translations of the audio should have been “(man speaking over PA system)” and “(birds chirping),” despite the latter being insignificant. Most importantly, during the top shot, the young woman is sobbing (plot pertinent!) and that absolutely should be captioned, with the PA part placed on the next shot where that audio continues. No one’s learning any language here.

Middleeastern-dressed nurse speaking over an ill Middleeastern woman's bed, captioned Mme Mika?/Ms. Maika?/Mrs Manka?
As far as I know, Arab culture doesn’t espouse women’s lib, so the machine “Ms.” is a cultural #SubtitleFail. Then, it seems the translator is used to British conventions because in North America we use a period after “Mrs.” and the surname is misspelled. In any case, these inconsistencies would be confusing to a language learner without the knowledge of these cultural points.

Middleeastern-dressed nurse speaking over an ill Middleeastern woman's bed, captioned Elle a recueilli les enfants./She collected the children./She safeguarded the babies
Here the nurse is interpreting from the patient’s Arabic. “Safeguarded” is the wrong diction for this scene: it’s too formal and, in terms of babies, is a bit archaic. For the newborns, who are essentially refugees, “took in” is an appropriate choice. A student using this would sound like they were talking about a report by a board of governors rather than caring for little ones.

Young woman facing young man, her face expressing horror, captioned Inspiration/Inspiration
No spoilers, but here is another misrepresentation of “(Inspiration).” This is a gut-wrenching gasp of horror at the first of two climaxes in the film…

Closeup of young man and woman, captioned  Je vous aime/I like You/I love you.

Closeup of young man and woman, captioned Vore mere, Nawal/Your mother, Nawal./Your mother.

Closeup of young man and woman, captioned  Reniflements/Sniffles/Nawal
The problems with these three subsequent subtitles are obvious. Again, they take the viewer out of the narrative, disrupting their immersion in the poignant dénouement of the story, and teaching nothing about language.

These are just a few examples to illustrate how the notion of teaching a language is far more complex than throwing up some setting options and calling it language learning.
Yes, it’s great if you know some, say, Polish and want to check what a character said, or if you need to pause the subtitles for better comprehension. But to suggest that language lessons are being made available by a streaming service that is known for its problematic subtitles and its craptions is misleading. It’s just another way Netflix holds a monopoly on the international offerings of video-on-demand but is putting the cart way before the horse. They need to get serious about native target-language speakers as subtitle and caption editors and fix the timed text before they start misinforming the public about foreign languages. For now, I’d recommend using some language-learning software or apps, or—much better—taking accredited classes in the language you want to learn. You can’t learn how to drive an eighteen-wheeler on the highway by trying out a Segway.

Please follow and like us:

Book Review: The Routledge Handbook of Audiovisual Translation

Abstract watercolour spheres as decoration of textbook,  The Routledge Handbook of Audiovisual TranslationEdited by Luis Pérez-González as part of the Routledge Handbooks in Translation and Interpreting Studies series, this new book is textbook material but is still accessible to the nonacademic with an interest in audiovisual translation.

I spent my first two years of university studying translation and linguistics and, in hindsight, now regret not having stayed in that stream. While my work focuses on the end steps of the AVT process (whether subtitles or captions/SDH), I’m still interested in language and how it is not as discrete from the technical production process as most people think. Scholarly work in this area is being taken more seriously as the field has now been accepted as a bona fide academic discipline.

Because they were brought up by so many of the 32 leading scholars who contributed essay-chapters, I’d like to discuss the main themes I noted: changes in technology, obviously, but also inclusion, exclusion, and changes in quality standards (the latter being my favourite aspect, of course).

The book provides some history in terms of subtitles, captions, and translation in cinema and discusses some of the software options currently available. It’s interesting that where Alina Secară’s part (p.139, 141) mentions eyeglass development as a means of caption delivery, even that area is changing quickly as we saw in October 2018 with the National Theatre in London’s introduction of Smart Caption Glasses by Epson. There is also a return (for me) to some concepts I read about in books I reviewed and interviewed authors about, such as Nornes’s thoughts on abusive subtitling (p.460) and Dwyer’s on prosumers (p.442) and the politics of fansubbing.

There seems to be a tension between the inclusion and exclusion that can be found in AVT. As I understand it, inclusion involves the performativity (p.446) and widespread participation by various factions (p.419, 438, 442). Sometimes the work is done by collectives on Viki or Amara, for example, and sometimes by fewer contributors, such as individual YouTubers—whether it’s their own content or someone else’s. The idea of prosumerism is covered not only by Dwyer but also Díaz-Cintas (p.31), Pérez-González (p.31) and Jones (p.187). Dwyer introduced me to the element of play being part of the performativity (p.446), and it took me this second crack at the literature to understand the degree to which AVT not only involves various politics (e.g. participation) but also the economics of the social contracts that are understood in many unofficial or unsanctioned undertakings. Localization straddles the areas of inclusion and exclusion, both as an “act of homage” (p.446) but also a kind of bowdlerization, such as the de-anglicization of text in Harry Potter for an American audience (Guillot on Nornes’s corrupt domestication, p.38).

But all is not warm and fuzzy. There is exclusion that is perhaps inevitable with AVT. In her discussion of music-video fansubbing, Johnson (p.421) cites Pérez-González and the “widespread assumptions of the dominance of English in globalizing process.” Dwyer (p.441) talks about the “global language politics and hierarchies” by netizens or global citizens. In her chapter on AVT and activism, Baker notes that not only fansubbers but also most subtitlers and captioners are not credited, or at least work unappreciated, in anonymity or invisibility (Baker, p.456–57). In my own advocacy efforts, which call for subtitle and caption editing to be recognized by film awards as much as other technical contributions like sound editing, I will give shout outs to excellent translations for film (such as in Les Innocentes, 2016; I can’t find my original post praising the subtitler anymore, so if anyone knows their name, please contact me!). I don’t understand why title designers are front and centre, but the professionals who made the audience’s comprehension of the dialogue accessible aren’t considered worthy of a credit line. Secară (p.138) also quotes Rondin’s discussion of smart glasses as a solution “without interfering with the overall show.” Maybe this is just my politics, but it always sounds like providing caption users with the technology to take part in this cultural content is a pain in the ass and must not disturb the public, such as the public’s general distaste for open captioning, unfortunately supported by a deaf person in a recent piece. From what I hear in Deaf social circles and forums, the expectation isn’t perfection, just something that’s effective (not craptions, for example). Captioning excellence seems like it shouldn’t require advocacy for improvement. It’s not like we accept mediocrity in the latest smartphones. Anyway, that’s a jump I made in my thinking.

Of course, what I was most thrilled by were the chapters where AVT training and teaching are addressed and what the future of quality assurance will involve with legislation. For instance, here, the Accessible Canada Act (ACA) is forthcoming, and the AODA is in place, but my Twitter feed is full of justified complaints by people of all types of disabilities because standards on paper and actual, informed enforcement are not the same thing. Merchán’s chapter (29) about training and McLoughlin’s (30) about teaching and learning made me hopeful. I was thrilled to read about Ken Loach and his rejection of the traditional AVT-as-postproduction model because budgets don’t plan or allow for quality subtitling/captioning, and Liz Crow (p.506) seeing accessibility as integral to the production process rather than a lowly add-on. Pablo Romero-Fresco has a book coming out shortly, Accessible Filmmaking Guide (London, BFI), which I couldn’t be more excited about (and he’s graciously agreed to an interview with me once I’ve read it). Study of filmmaker/subtitler collaboration by the University of Roehampton and programs like the MA in filmmaking at Kingston University (London) addressing accessibility and AVT as par for the course also give me hope. I’m currently trying to impress upon colleges near me the importance of caption editing being taught as a foundational course and program requisite because all the ACAs and equivalents in the world aren’t going to eradicate the problem of craptions (as inaccessibility) if filmmakers aren’t taught the soft skills now. I can’t figure out why more postsecondary institutions aren’t scrambling to implement this, particularly when they advertise accessibility production as one of their training outcomes. Mohawk College’s Accessible Media Production is the only program where I can see the genesis of serious application to this in their curriculum.

I loved the quotation of Marleau from 1982 that Secară concludes her chapter with (p.142)—and here surtitling could easily be replaced by subtitling: “…surtitling and captioning services are not to be regarded as ‘un mal necessaire’ [sic] (‘a necessary evil’).” I’ve attempted to walk the walk in my rhetoric about this and have launched an award for excellence in captioning in the hope that we will raise more Loaches and Crows who will see captioning excellence as one of the foundational stones in the building of a film, and not as a requirement remembered just as the student is about to hit Send. The d/Deaf, hard of hearing, and many other types of caption users are not dismissible, and as I’ve written before, I’m not going to shut up about it. Fortunately, inquiries about the award from filmmakers are heartening: there is will—but also many barriers remain.

Pérez-González’s edited collection of essays by some of the top scholars in audiovisual translation today—for me—is summarized best in Romero-Fresco’s position that AVT services are an afterthought at best. He notes that the United Nations’ ITU Focus Group on Media Accessibility and filmmakers such as Tarantino and Iñárritu are trying to influence, respectively, the profession and the process by being involved in subtitling (p.510). I don’t see change being swift, but I hope that ten years from now we will see improvements in quality via subtitle and caption editing. Meanwhile, The Routledge Handbook of Audiovisual Translation  gives the student, academic, professional, and interested lay reader an excellent idea of the lay of the land in AVT. It will be interesting to see what has—and hasn’t—changed in education, standards, and enforcement by the time a second edition is published.


Please follow and like us: