Having closed captions is always better than not having them at all. But sloppy captions — that is, captions that are misspelled, ignore rules of capitalization, or are simply illegible in one way or another (low contrast, too small, all caps) — show a lack of respect for viewers who use them. Sloppy captions imply that closed captions really aren’t all that important. They’re an afterthought, a necessary evil, an add-on, a nuisance, a bother. While having closed captions is infinitely better than not having captions at all, we shouldn’t settle for captions of inferior quality.

Consider a recent episode of 30 Rock entitled “Future Husband.” It’s episode 14 from the 4th season and was originally broadcast on NBC on March 11, 2010. The full episode is still available as of today (4/18/10) on Hulu.com, which is where I watched it for the first time, complete with closed captions enabled.

These are some pretty sloppy captions. As a reader of this post has kindly pointed out to me (see comments below), it’s likely that the problems associated with these captions are due to the use of auto-caption technology to re-capitalize all-caps TV captions. The resulting caption file simply wasn’t corrected by a human. That would explain a number of the problems with this caption file, particularly the failure to capitalize proper nouns that also pass in other contexts as verbs (e.g. miss, will, don). But it doesn’t easily explain the use of capitalized common nouns in the middle of a sentence or the failure of the technology to capitalize proper nouns (e.g. city, state, and country names) that any speech recognition system should capitalize automatically.

Regardless, I’ve never seen captions on Hulu as bad as these — that’s the larger point here. In this post, I want to explore what effects very minor capitalization errors can have on the reception of captions.

To be fair, the captions could be much worse. They’re spelled correctly, adhere successfully to a verbatim style, and reflect a sensitivity to temporal precision. But they also exhibit an utter disregard for the basic rules of capitalization. Inconsistent or erroneous capitalization violates viewers’ expectations. When I see words that should be capitalized but aren’t, or words that shouldn’t be capitalized but are, I have to make adjustments in the way I process the captions and how I experience the video content. Such adjustments may disrupt the immersive experience of watching a TV show or movie because viewers must stop being fully engaged with the content and start trying to rationalize or mentally fix mis-capitalized words.

I counted and logged 79 examples of capitalization errors in this single caption file. Quite a large number, I’d say, for a 22-minute episode of network TV. Here are just a few of the many capitalization errors, grouped into a small number of loose, overlapping categories:

  • Proper nouns that aren’t capitalized
    When proper nouns aren’t capitalized, viewers’ expectations are challenged. When an uncapitalized proper noun can also be mistaken for a verb, it may cause even greater reception problems. Words like “don,” “miss,” “will,” “fed,” and “screech” are proper names or forms of address in this episode of 30 Rock. But they are also verbs and may be interpreted as such when they are not capitalized.

    The caption reads: Oh, miss Lemon, it's so romantic.The caption reads: You're will Smith, and I'm the dog.The caption reads: or when screech went to the masquerade ball in disguise.The caption reads: which is the song I sang to don at his promotion dinner.The caption reads: I'd have sex with K-fed again.The caption reads: My parents have kabletown down in pennsylvania.
    This is by far the largest category of mis-capitalizations. I logged 40 examples into this category, including a whopping 12 “dons” and 5 english/british examples. State names (pennsylvania, connecticut), country names (vietnam), city names (philadelphia) people’s names (igor, K-fed, don, miss, will Smith, jamakiah), movie titles (fatal attraction), and company names (kabletown) fall into this category.

  • Common nouns that are capitalized
    When common nouns are capitalized, viewers may wonder if they were capitalized on purpose to create a special class or perhaps to instill a term with special significance. As a viewer, I tend to assume that capitalized words are intentional and thus integral to the meaning of a film, movie, or TV show. I also assume that caption files will adhere to rules of capitalization. The rules for capitalization are simple, easy to follow, and rarely violated accidentally. So when common nouns are capitalized, I conclude (erroneously in the case of this TV episode) that it must have been done on purpose.

    In the first example below (“Almost Exclusively Women…”) I assumed on first viewing that a special class of people was being created for humorous or some other effect. Of course, I quickly came to understand that the captions were just plain sloppy.

    The caption reads: Almost Exclusively Women that look like me.The caption reads: About the Lindsey vonn Who won the gold medal for skiing.The caption reads: said Nicole Kidman Should get an Oscar for it.The caption reads: spoke about this, and that sale is Dead.The caption reads: Well, then you better think of Something else.The caption reads: Do you think it was easy for me to cut Ties.The caption reads: You must be my wallet Angel.The caption reads: Miss Lemon, how did the dentist's Office go?The caption reads: from Dr. Kaplan's Office

    This last example seems intentional, as though “Dr. Kaplan’s Office” signifies a proper noun, a formal way of representing the name of the office. And that’s precisely the problem. Viewers may try to rationalize mis-capitalized examples by making them fit their preconceived notions of what proper nouns do and how they work. (By the way, I logged 7 examples of “Office” in capitalized form.)

  • Acronyms that aren’t capitalized
    Uncapitalized acronyms don’t cause as much disruption to the viewing experience as common nouns that have been mis-capitalized. Uncapitalized acronyms are just sloppy. The caption reads: did I put a toaster waffle into my dvd player?The caption reads: Hey, cnbc, nightly news rules.The caption reads: They're selling nbc to a company called kabletown.The caption reads: I'll never be the ceo of the general.The caption reads: during the rca Deal.In the CEO example, both CEO and “The General” should be capitalized, because “The General” designates a nickname for General Electric. When capitalized, “The General” is more quickly understood by caption users. In the RCA example, “Deal” is a common noun that should not have been capitalized. When it’s capitalized, it takes on a special significance (cf. The New Deal) that probably isn’t warranted or intended here.
  • The “just plain sloppy” category

    File these under “what were they thinking?” The saving grace of these examples may be that the capitalization errors are so egregious, so obvious, that they can’t be mistaken as intentional. As a viewer, I know they’re wrong as soon as I see them, and I’m not likely to try to rationalize them as intentional (unlike some of the examples above).
    The caption reads: DON't worry. It's just a donkey spell.The caption reads: current g.E. Executive Jack Donaghy.The caption reads: That's the sound of me being erased from CONTACT lists.

Conclusion: Just a minor inconvenience or something more?

At the end of the day, capitalization errors are minor in comparison to the much more pressing problem that uncaptioned Web videos pose. Still, I worry about the message these captions send about the low value placed on accessibility. Why are such captions allowed to pass muster at Hulu? Caption users deserve higher quality captions. Even something as seemingly minor as a mis-capitalized word can cause reception problems for users, as I’ve tried to show.

I also worry that sloppy captions, should they become more widespread, can interfere with the efforts of young children and non-native speakers of English to use closed captions to increase their command over written English. A recurring argument among caption advocates is that captions can help all children — deaf, hearing, hard-of-hearing — to increase their control over and knowledge of written English. But that argument has always assumed, it seems to me, that captions by default will adhere strictly to proper usage guidelines. Sloppy captions could call the literacy argument into question.

Personally, I find sloppy captions annoying and insulting. I expect and deserve better.