By Jason Harrow
The editors of Oxford Dictionaries recently named the word of the year in 2015.  No, that is not a strange typo: The word of the year really was an emoji. But which one was it? While the title of the article displayed the emoji using Apple’s character set, the first sentence of the explanatory blog post used a different character set and printed that is the word of the year.
Of course, the two renderings were of the same emoji. The characters looked different because they were in different emoji “fonts”: The headline emoji, which appears rounder and with smaller, brighter tears, came from Apple Color Emoji, but the in-text version that looks flatter and has bluer tears is from Twitter’s character set.
Oxford’s use of two different emoji character sets gives rise to a question that has become increasingly relevant as emoji use has exploded around the globe: Could either of those versions of “face with tears of joy” be subjected to copyright protection? And, if so, does that mean that a publisher or media outlet might be committing copyright infringement when it chooses to display messages or text containing particular emoji characters?
This is a difficult question, as there is little relevant case law or congressional guidance regarding whether emojis are protected under the Copyright Act. But this article argues that emojis may well not be copyrightable because emoji sets are typefaces, and typefaces have long been held to be uncopyrightable. Nonetheless, the copyrightability of emojis is unsettled, so there would be some risk in treating an emoji set as not subject to copyright.
Typefaces “as typeface” are not copyrightable.
Before discussing whether emojis are subject to copyright, it is important to understand that, according to the Copyright Office and the federal courts to have examined the issue, typefaces are not copyrightable.
Congress actually considered the issue of typeface copyright when it passed the 1976 Copyright Act, though nothing about typefaces appears in Title 17 of the U.S. Code. Instead, the Committee on the Judiciary stated in the House Report that accompanied the Act that it had “considered, but chosen to defer, the possibility of protecting the design of typefaces.” The Committee stated that it “does not regard the design of typeface . . . to be a copyrightable ‘pictorial, graphic, or sculptural work’” within the meaning of the Copyright Act. The committee also defined typeface “as a set of letters, numbers, or other symbolic characters, whose forms are related by repeating design elements consistently applied in a notational system and are intended to be embodied in article whose intrinsic utilitarian function is for use in composing text or other cognizable combinations of characters.”
As the House Report indicates, the view that typefaces are not copyrightable is longstanding. Indeed, the leading appellate case to have considered the issue, Eltra v. Ringer, noted that “typeface has never been considered entitled to copyright” under the 1909 Act. Accordingly, the Eltra court in 1978 affirmed the Copyright Office’s rejection of an application to register a then-new typeface called “Orion.” The court reasoned that typeface is not copyrightable because it “is an industrial design in which the design cannot exist independently and separately as a work of art.” The U.S. Court of Appeals for the Ninth Circuit has also noted that “typefaces are not afforded copyright protection.” The Copyright Office has followed the view of the House and the 4th Circuit and has promulgated a rule that applications for registration of “typeface as typeface” will “not be entertained.”
But this does not leave font designers in the digital age entirely without copyright protection for their work. As the Copyright Office explained in its rulemaking, type designers can register copyrights in the font software used to create and display their typefaces on computers. Font makers can then license that software, even though the actual letters and numbers produced by the software would be unprotected under copyright law. Type designers have additional ways to protect their design work, including by trademarking the name of the font.
The fact that typefaces are not copyrightable means that the wholesale copying of typefaces is not prohibited by copyright law, so long as the copier is able to draw the figures by hand or otherwise reproduce the letters and figures without copying any software code. And, although this seems like a difficult result for professional typeface designers, in fact one scholar reports that “many professional font designers are quite happy with this result.”  That is because having “the copyright in the code gives the company the necessary head start in the market, after which it is not necessarily a problem if a competitor recreates the resulting typeface design from scratch without copying the protected code.” Plus, “some degree of copying is necessary and accepted,” since a typeface can only be useful as a typeface if each figure is recognizable as the given letter, number, or symbol it represents.
Emoji sets could be considered uncopyrightable typefaces
When emojis were first created in Japan in the 1990s, they were a series of proprietary symbols that individual mobile companies created for their users. But now that emojis have evolved, it is becoming clear that emoji character sets implement a kind of universal pictorial language of our digital age. Accordingly, there is now a strong argument that emoji sets should be considered typefaces implementing the emoji “language” and therefore are not protectable under federal copyright laws.
Although the Copyright Act does not contain a definition of typeface, the legislative history does. The House Report defines typeface as “a set of letters, numbers, or other symbolic characters, whose forms are related by repeating design elements consistently applied in a notational system and are intended to be embodied in article whose intrinsic utilitarian function is for use in composing text or other cognizable combinations of characters.” Under that definition, emoji sets could be considered typeface. As the committee recognized, typeface need not be only letters or numbers, but can also be “symbolic characters.” Thus, emojis could potentially be considered typeface so long as they meet the other requirements.
The key to being considered a typeface, then, should be whether emojis serve a “utilitarian function” in a “notational system” with a “cognizable combination of characters.” It is becoming clear that emojis also meet this requirement. Emoji character sets no longer consist of whatever sets of symbols that the manufacturer may wish to include but now conform to standards set by Unicode, which is the international consortium that writes specifications to “enable people around the world to use computers in any language.”
These Unicode standards are the equivalent of the specification for transmitting and displaying character sets in any other language. For instance, just as the familiar basic Latin forms have been assigned standard hexadecimal codes — a lowercase “m” is 006D, for instance — so too does each individual emoji now have a code — “smiling face with sunglasses” is 1F60E.
From there, in just the way that a smartphone displays a lowercase “m” differently based on whatever font is selected, a modern smartphone also displays the “smiling face with sunglasses character” according to whatever emoji “font” set is enabled. Apple uses Apple Color Emoji, which displays that character like this: . The Android character set looks like this: . And both of those are variations of the Unicode “standard” for that character, which is rendered in black-and-white and looks like this: .
These emoji characters have important aesthetic differences among them. But just because the designer made aesthetic decisions — should the face be round or another shape? — does not mean the resulting emojis are pictorial or graphic works subject to copyright protection. Instead, the aesthetic choices likely constitute the industrial design of a set of utilitarian symbols whose purpose is to represent characters in the emerging “notational system” that is the emoji language. Viewed this way, emoji characters are the building blocks of a new, global digital language — and should be treated in the way that the building blocks of our pre-existing languages are treated. In fact, given the exploding popularity of emojis, some typeface designers have speculated that an emoji set (perhaps Apple Color Emoji) may well end up as the signature typeface of the 21st century.
The Copyright Office, too, recognizes that typeface can extend beyond minor variations of a standard alphabet. Its Compendium of U.S. Copyright Office Practices notes that it refuses typeface registration “regardless of how novel and creative the shape and form of the typeface characters may be.” Rather, a “typeface character cannot be analogized to a work of art, because the creative aspects of the character (if any) cannot be separated from the utilitarian nature of that character.” Here, too, where emoji sets implement the standard Unicode character set, it is hard to separate the unique creative aspects of one rendering of “face with tears of joy” from its ability to be used as a character in the emoji alphabet.
This concept of a set of symbols that convey complex meanings may seem strange to native English speakers, but phonemic alphabets (like the Latin alphabet that this article is primarily written in) are not the only way to construct a writing system. Rather, the use of logograms or pictograms in written language has both ample historical precedent, such as Ancient Egyptian Hieroglyphs, and continued usage, such as Chinese characters. Few would contend that Chinese character sets should be copyrightable as typeface even though typefaces depicting the Latin alphabet should not. The same principle could well be applied to emojis.
Admittedly, we may be too early in the life of emojis for a court to declare emoji sets uncopyrightable works of typeface design. There are some factors that point toward an argument that emoji characters should be copyrightable works of visual art.
One might argue that individual emojis can be appreciated for their individual aesthetic properties to a greater extent than individual letters can — elevating them to works of “pictorial” or “graphic” art. After all, unlike Latin characters, emojis are so closely representational of their underlying meaning: the smiling face with sunglass emoji looks like … well, a smiling face with sunglasses, as opposed to these jagged forms we call letters that call to mind a given image. But what makes emojis “artistic” in this way is also precisely what makes them so universally useful: Emojis constitute a kind of writing “whose fundamental unit is not the word” but, maybe, “something better.” Emojis thus still could be considered “useful articles” even though many of them are representational of the ideas they are meant to convey.
From a legal perspective, a search of Copyright Office records reveals that only one company has so far tried to register its emoji characters in a comprehensive way: Apple. That is, the Copyright Office has recently registered as “2-d artwork” at least some of Apple’s emoji characters as they existed in 2008 and 2011. But the fact Apple has several hundred registrations for individual emoji characters (plus a registration for its “typeface palette” in iOS 3.2) is far from conclusive proof that the characters are copyrightable. After all, the issuance of registrations from the Copyright Office is only some evidence that a work may be copyrightable; ultimately, that determination must be made by a court. If Apple were forced to defend the copyrightability of these individual characters, it could face some problems.
In particular, the title of the registrations undermines somewhat Apple’s case for copyrightability of emojis as “artwork.” Almost all of the registrations use the word “typeface” in the title: for instance, there are a number of registrations for “Apple Emoji 2008 typeface,” followed by a category signifier (“PPL” for people; “OBJ” for object), and then, frequently, a unique identifier that includes the Unicode hexadecimal code. More recent registrations are titled with the pattern “Apple Emoji 2011 typeface for OS X” followed by a descriptor, like “WOLF FACE,” “CLOSED BOOK,” or “SEEDLING.”
This titling is potentially problematic for a couple of reasons. First, the fact that “typeface” appears in the title of the works should have signaled to the Copyright Office that the submissions were part of an unregistrable typeface according to its own regulations; it is unclear if Apple addressed that issue in its submission. Second, the fact that many of the submissions are identified by their Unicode hexadecimal code also makes explicit that those symbols are implementations of a standardized set of symbols, not a set of proprietary symbols that Apple designed on its own. Thus, it is far from clear that all of these registrations will hold up in an infringement suit, at least if the suit is brought against someone who simply displays the typeface but who has not copied any of Apple’s software code.
There are only a few cases addressing the copyrightability of typefaces, and there is none discussing the copyrightability of emojis. Obviously, then, the IP status of emoji is new, untested legal ground. What should best practices be for media organizations that wish to use or display emojis created by Apple, Google, Samsung or any other company that may assert copyrights in the character sets they created?
One option is to treat emojis as you would other commercial typefaces, as proposed in this article. A more conservative option would be to treat particular emojis as works of visual art and ensure that each particular display of a given emoji is properly licensed. An organization may also wish to create its own emoji set based on the Unicode standard to avoid copyright issues. Organizations that choose this option should probably be careful not to copy Apple’s emoji set verbatim, but, just as close copying is common for Latin-based typefaces, close copying might also be acceptable in the realm of emojis.
In all events, however, media organizations should take care to make sure not to copy the software code for an emoji font set without authorization; this would be a clear copyright violation. They should also be careful not to use copyrighted code as the basis for creating a new emoji set.
It will also be important to monitor how emojis are used over the next several years. Languages evolve, and emojis will too. If emoji usage continues on its current path, we may see an explosion in both emoji “font” design and acceptable usage of emojis in new and surprising ways. If that is the case, the argument that emoji sets are typefaces for our new universal language will only grow stronger.
 A comparison of the way “Face With Tears of Joy” looks in those and other character sets can be found at: http://emojipedia.org/face-with-tears-of-joy/
 H.R. Rep. 94-1476 at 55.
 Eltra v. Ringer, 579 F.3d 294 (4th Cir. 1978); see also Eltra v. Ringer, 194 U.S.P.Q. 198 (E.D. Va. 1976).
 Monotype Corp. PLC v. International Typeface Corp., 43 F.3d 443, 446 (9th Cir. 1994).
 37 C.F.R. § 202.1.
 57 Fed. Reg. 6201–02 (1992).
 Adobe Sys. Inc. v. S. Software Inc., No. 95-cv-20710, 1998 WL 104303, at *3–4 (N.D. Cal. Feb. 2, 1998).
 See Monotype Corp. PLC v. Int’l Typeface Corp., 43 F.3d 443, 446 (9th Cir. 1994).
 Jacqueline D. Lipton, To © or Not to ©? Copyright and Innovation in the Digital Typeface Industry, 43 U.C. Davis L. Rev. 143, 174 (2008).
 Id. at 169.
 H.R. Rep. 94-1476 at 55 (emphases added).
 In Japan in the 1990s, each carrier had its own noncompatible emoji sets, resulting in a “jumble of different proprietary approaches.” http://www.theverge.com/2013/3/4/3966140/how-emoji-conquered-the-world.
 As one commentator recently argued, a given emoji character is “a discrete and combinable written element that conveys semantic . . . meaning.” http://www.vice.com/read/sam-kriss-laughing-and-crying?utm_source=vicetwitterus.
 U.S. Copyright Office, Compendium of U.S. Copyright Office Practices (3d Ed. 2014) at § 906.4. The Compendium notes that “simple emoticons, such as the typical smiley face” are unregistrable on the grounds that they fall into the category of “well-known and commonly used symbols.” Id. § 313.4(J). While it may be difficult to argue that an entire digital language composed of hundreds of emojis would all be unregistrable as “well-known symbols” — though perhaps emojis’ explosion in popularity will change that conclusion — certain emojis that do implement well-known symbols would not be subject to copyright on these independent grounds.
 The registrations can be found by searching Copyright Office records using the keyword “emoji.”
 See 17 U.S.C. § 410. Moreover, as a technical matter, it appears that some of Apple’s registrations may not come with a presumption of validity, because they were not filed within five years of first publication. See 17 U.S.C. § 410(c).
 17 U.S.C. § 410.
 Or, if emoji are treated as works of visual art, certain uses also may be permissible because they are fair. Oxford’s use of particular emoji characters in an article about emojis, for instance, is likely a fair use. Displaying a person texting using particular emojis may also be a fair use.
 The full list of emoji can be found at http://unicode.org/emoji/charts/full-emoji-list.html.