emphasis mine.

How To Write A Name

There are several ways to refer to a single film. As an example, Da xiang xidi erzuo is a film title I don’t recognize. 大象席地而坐 is a title I’d recognize, but not exactly be able to read. An Elephant Sitting Still is the way I’ve referred to this movie when recommending it to friends. Region, language, and choice of script all can have an impact on the title we know of film by.

Last year, as I was reading through Senses of Cinema’s World Poll 2019, I saw that first transcription, followed by the English localized title. I’d come across other non-English titles in the list prior to that point, and without thinking had assumed they were the original title. For this particular entry I knew that wasn’t the case. How does a publication make this kind of decision, and why would one make this specific choice?

A Matter Of Style Guides

Senses of Cinema’s style guide is pretty clear about their handling of film titles:

For non-English language films, list the original title first, followed by English title in parentheses if there is one. For example: Il deserto rosso (Red Desert, Michelangelo Antonioni, 1964). Thereafter use the English title.

This is a great stance! I’d landed somewhere similar when designing my film diary as well as when writing Flags of Dead Nations. The original title of a film, when it exists, is as much a part of the work as the shots within. Even the translated titles are (see A Touch of Sin), but the original carries special significance.

The choice to use a film’s original title doesn’t address the usage of pinyin, a romanization system. Romanization is the process of transforming a writing system into the Latin script. Using the original script would mean writing 大象席地而坐. The example from the style guide, Il Deserto Rosso isn’t helpful here, as Italian and English both already use Latin characters. I’d have to look elsewhere for an explanation for the choice to romanize.

Single Script Rationales

In May, I placed a hold on Heroes Shed No Tears 英雄無淚 at the library. When it was available I received an email that included this section:

   Ying xiong wu lei [videorecording]
   CALL NO: CHINESE BLU F YING
   BARCODE: 31223135872605
   LOCATION: MAIN - 1st Floor AV Cente
   PICKUP AT: MAIN LIBRAR BY: 05-29-21

The library had made an interesting choice to refer to the movie exclusively with pinyin0. This doesn’t quite satisfy the “original title” rule that Senses put forward, since it’s a Cantonese language film, and Cantonese has its own romanization systems. However, the choice to only use Latin script here makes sense. Two lines in this text are truncated (“Cente” and “LIBRAR”). Whatever cataloguing system this library uses has limited widths. That certainly restricts them from using both English and non-English titles simultaneously, and doesn’t inspire confidence in the ability to handle non-Latin characters.

Senses of Cinema likely faced a similar dilemma. The site began in 1999, only 8 years after the debut of the World Wide Web. In 2001, about 80% of sites in Google’s index were in Latin script (or a subset of it). In that era, it would have been highly unlikely for an English language website to even be capable of presenting multiple scripts simultaneously. This is likely enough to explain how their style guide was interpreted to lead to Da xiang xi di zuo1.

Today, we don’t have that same challenge, given the widespread adoption of Unicode. Unicode is an encoding system that largely tries to represent all human languages, allowing characters in different writing systems to easily sit side by side. What would a style guide for today’s capabilities look like?

Against Romanization

At its best, romanization can give a reader the opportunity to make sounds that approximate those of the original language. This is a useful tool in language learning to get students speaking before they’re able to write (the typical order when learning our first language). In a strictly written format, the benefit of making sounds in one’s head while reading is smaller, and it invites several new problems with it.

House and Foreignization

Romanizing film titles, while ostensibly done to make it more accessible, can sometimes imbue that title with a wholly invented foreignness.

A prominent example of this phenomenon is the 1977 Japanese film House. This film was introduced to me as “Hausu”, and still so much English language writing about the film uses that rendering. The opening title card, posters for the theatrical release of the film, as well as interviews with the director all indicate that the name is House, and that having an English title was intentional. When referred to as “Hausu”, the speaker/writer is choosing to emphasize the “Japaneseness” of the film in a way that it itself doesn’t invite.

A more recent example of this is the film We Are Little Zombies, which IMDB chooses to write as “Wî â Ritoru Zonbîzu”. This isn’t strictly wrong from a romanization perspective, but again accentuates differences that don’t meaningfully exist. “Wî” would be pronounced basically identically to the English word “we”, and the “n” in “Zonbîzu” would be pronounced as an “m” by Japanese speakers due to the presence of the “b”.

In the posters for these films, the Japanese rendering of the title is typically quite small, an indication that it exists as an annotation rather than as the “original” title. They serve as a reading aid for those who might not be able to fully read the English title. Choosing to elevate those reading guides positions the publication as the authority of “original title” over the creators themselves.

Ambiguity and Reversibility

Not all romanization schemes are, or can be, reversible in such a way that a reader who knows the language in question can determine what the non-romanized expression is. In the World Poll mentioned above, there’s at least one entry for Mirai-no-Mirai. Unless one is already familiar with the movie and its original title, it isn’t possible to intuit how that name would be written in Japanese, and thus what the name means.

Romanization inevitably means that words that are homophones (sharing the same sound) will become homographs (written the same). Context can sometimes provide the clues necessary for reversibility, as in 寝ても覚めても (Netemo Sametemo), but that’s not reliable. We lose real information in dropping the native script.

A Better Gateway

Most immediately, the choice to romanize impedes access to materials written in the film’s own language. Using the example of An Elephant Sitting Still from above, if I want to find Chinese language writing about the film, the romanized form won’t get me far.

Librarians and archivists have been discussing this issue for decades. In 1977, Charles Sumner Spalding, the former assistant director of cataloging at the Library of Congress, published Romanization Revisited. Previously a supporter of romanization in cataloging, in that piece he outlined the various challenges that romanization continued to pose, and instead suggested having language/script specific catalogs. What use was romanized Greek to a monolingual English reader? What impediments did romanization inflict upon the bilingual Greek & English reader, in not allowing that ability to be used?

The most common search operations we perform today are on unstructured search engines like Google. In those cases, the romanized title is usually only going to produce other usages of the romanized title. It doesn’t even provide access to the native language materials that the librarians and archivists were debating in the 70s.

Presumption of Audience

Ultimately, my motivation against romanization is a personal one. I’m not a monolingual English speaker. And I won’t presume my audience is. (In many cases, my audience is primarily me anyway). And even for those that are, the normalization of a mixed script world is good in and of itself.

Challenges

Despite the presence of good arguments against romanization, eschewing it introduces its own problems.

Mixed Reading Direction

One commonality between English and Italian, as in our previous Il Deserto Rosso example, that can’t be taken for granted is reading direction. Both English and Italian are read left-to-right. Arabic, on the other hand, is read right-to-left, potentially introducing a burden when spliced into an English sentence.

Since most documents on the internet use only one script, authors don’t often consider how to handle bidirectional text, and it produces a number of unexpected artifacts.

A style guide that encourages using the native script of the original language of a title will have to consider the potential impact of including a title with a reading direction opposite that of its surrounding text.

Han Unification

In an attempt to keep the character count of Unicode below a certain size, the Unicode Consortium decided to reuse certain code points of characters in Chinese, Japanese, and Korean Hanja that referred to the same general character, even if they’re written differently. This choice is akin to having “δ” and “d” be identical characters that only differed when using a specifically Greek font.

The result of this is that Chinese titles in my film diary (which doesn’t yet specify language per-film) will often use the wrong character, because it finds the Japanese equivalent first.

While this is a solvable problem, it’s an additional consideration and source of possible error. It also means that copy and pasting text from one location into another might result in it being apparently altered.

My Style From Here On

Prior to researching and writing this, I didn’t have a deliberate style. That inconsistency ate at me. Why did I sometimes give certain titles one treatment, and then others a different one?

Now, having gone through this whole exercise, I feel good about the following approach:

In prose

Priority should be given to the flow of the sentence. As a result, English language title comes first with the original title in parentheses. In subsequent appearances, only use the English title.

Example:

Last night I watched The Mirror (آینه). The Mirror is a…

In listings

In a list-like format there’s no sentence flow to consider, so give priority to the original title, with the English title in the secondary position using whatever stylization or separation makes sense. For all subsequent appearances, continue to use both titles in the same order.

Example:

Tomorrow’s double feature:

  • 用心棒, Yojimbo
  • 椿三十郎, Sanjuro

Open Questions

This only touched on titles of artworks, but there are other proper names that are worthy of the same exploration. How should one refer to an author? How should one refer to a place? What does “original name” mean for either of these things?

I’m not sure how to approach those just yet, and I’ll likely fumble around, trying different things for a while before landing somewhere concrete. A not too dissimilar approach to the journey that led me here now.

0 It’s worth mentioning that pinyin would typically include tone marks, which none of the examples here use.

1 Senses of Cinema has a persistent typo in the title where one character, 而/er, is missing.