On 22 May 2013, programmer Steve Wilhite was honoured with a Lifetime Achievement Award at the 17th Annual Webby Awards in New York. Wilhite received this award chiefly in recognition of the Graphics Interchange Format, popularly known as GIF, which was developed by him and his team at CompuServe in the 1980s. This format is best known for the short animations that it enables, which have become a staple of the Internet, especially since platforms such as Facebook and WhatsApp started providing support for the format in their messaging services.
While historically interesting, it may, at first glance, not seem to be of particular concern to people interested in the English language. There is, however, one memorable aspect to this story: Wilhite's acceptance speech. As is customary for such speeches at the Webby Awards, it contained only five words. True to form, Wilhite's made his speech in the form of a GIF, which read ‘It's pronounced “JIF” not “GIF”.’ With his ostensibly heartfelt speech, Wilhite made instant linguistic history. He seems to be the first ever coiner of a word (or acronym, to be more specific) who gave usage advice about his own creation. This case presents a unique possibility for students of prescriptivism, as we can study what the impact of such a statement is. Did users respect Wilhite's authority and subsequently accept his preferred pronunciation? In what follows, I will address this question. I will first discuss this particular usage itemFootnote 2, after which I will describe the data and the method I used for my analysis. Then, I will discuss the verdicts and arguments found in the dataset.
2. GIF or JIF
Wilhite's statement acknowledged that there was, in fact, variation in the pronunciation of the acronym GIF – otherwise, no usage advice would have been necessary – and this is indeed the case. The two pronunciations that Wilhite referred to are [gɪf], with what is commonly referred to as a ‘hard’ g, and [dʒɪf], which has a ‘soft’ g. Prior to Wilhite's speech, few people apparently payed attention to this usage item. For example, searching database LexisNexis (2019), which contains hundreds of newspapers, website publications and other written material, for ‘pronunciation of GIF’ only yields four hits prior to Wilhite's speech, all published between 2001 and 2012. After May 22nd, this number is 27. Additionally, some dictionaries did mention the word, such as the OED (s.v. GIF, n.), which has an entry dated 2006. But in usage guides the pronunciation of the acronym seems to be absent. The reason for this is unclear: It could have been a symptom of the long-time prescriptive focus on ‘old chestnuts’ (Weiner, Reference Weiner, Stanley and Hoad1988: 173), or the variation could have flown under the radar as a result of the relatively narrow (and young) user base.
The silence was shattered, however, due to Wilhite's speech. Either people suddenly started caring about the ‘proper’ pronunciation of this acronym (a scenario that does not necessarily seem that unlikely, based on the regular behaviour of language pundits), or they had already formed an opinion but failed to share it. Perhaps speakers did not want to share their views because they lacked an authority to refer to, which for language users, as opposed to prescriptivist writers, is often quite important (cf. Peters & Young, Reference Peters and Young1997; Peters, Reference Peters, Aarts and McMahon2006). Either way, it seems entirely feasible to think that Wilhite's statement would be picked up as this much-needed authority to which people could refer, and that it would have an effect on prescriptive comments, both because of his status as the inventor, and because of the previous lack of authority.
In order to investigate the role of authority in the case of GIF, I collected online statements about the pronunciation of GIF. For this, I used a variety of online sources, including Reddit, YouTube, websites, Facebook and various newspaper articles. In the case of the articles and YouTube videos, I looked both at the texts themselves and at the comments. In certain cases, especially on Reddit (one of the most popular online fora), threads consisted of several thousand comments. In these cases I took a random sample of the complete body of comments.
Statements and texts about the issue had to conform to certain parameters. Of course, they had to include at least one assertion about the proper pronunciation of the acronym GIF. Next, they had to be available online, and they had to have been uttered after Wilhite's speech in May 2013. Parodies, while interesting in their own right, were excluded, as were statements made by speakers who explicitly stated that they were non-native speakers. This led to a total of 899 texts (see Table 1). These texts varied in length from 12-word comments to newspaper articles of over 800 words long.
These texts were subsequently annotated for two parameters: verdict and argumentation. Verdict simply entails what the writer thought was the proper pronunciation. I distinguished between [gɪf], [dʒɪf], both or none. In this final case, only an argument was presented, usually against one of the variants, without the writer taking an explicit stance as to what was the correct pronunciation, such as NomyourfaceDinosaur does: ‘By that logic, the “s” in “scuba” would be pronounced like in “softly” since scuba is short for “self-contained underwater breathing apparatus” and “ROM” would be pronounced like “Rome”’ (2013).Footnote 3
Next, I tagged the arguments that were used in the statements. I distinguished between arguments in favour of the writer's verdict and those that were used against other opinions. I built upon earlier annotation schema developed for tackling argumentation in prescriptivism (cf. Kostadinova, Van der Meulen & Karsdorp, Reference Kostadinova, Meulen and Karsdorp2016; Van der Meulen, Reference Meulen, de Bruyn and Berns2018). However, as many of the arguments used in these schema's did not apply here (nobody would say, for example, that GIF had to be pronounced in a certain way because it was a Gallicism), the schema was adapted for this particular purpose. I distinguished between four main categories, Authority, Quality, System and Use, each of which contained several more specific options (see Table 2). Authority and Use are closely related, but the crucial difference between the two categories is that in the former the source of authority makes an explicit and metalinguistic judgement about the preference for a certain form. In the category Use, reference is made to linguistic behaviour of some speaker or speakers. System entails all arguments that are made to some intralinguistic phenomenon, such as a grammatical rule. Finally, the arguments found in the group Quality all display some evaluation of the use of the particular pronunciation.
The shape that the argumentation surrounding the pronunciation of GIF could take is nicely exemplified in the popular TV series The Big Bang Theory. In episode 20 of season 8, The Fortification Implementation (2015), the following dialogue plays out:
Howard: Settle this. Those little animated pictures on the Internet, are they called ‘gifs’ or ‘jifs’?
Leonard: Well, the G stands for ‘graphics’. That's a hard G, so I'd say ‘gif’.
Raj: What? The guy who invented it says it's ‘jif’.
Howard: I'm sorry, do you mean the guy or the juy?
The three characters use three different arguments: Leonard uses the System_Acronym argument, for Raj Authority_Invent is decisive, whereas Howard goes for System_Analogy. As for verdict, both Leonard (explicitly) and Howard (implicitly) favour [gɪf], while Raj prefers [dʒɪf]. As I will show below, these combinations of verdict and argument are actually quite typical for the population at large.
4. The verdict
Out of the 899 texts, 28 gave no verdict, but only one or more arguments. Of the remaining 871 verdicts, 498 favoured [gɪf] (57.2%), while 277 others (31.8%) deferred to Wilhite's variant (even though they may not mention him explicitly), and preferred [dʒɪf]. As such, there is a clear preference for the [gɪf] pronunciation, even though it is definitely not the only acceptable one. This qualifies the statement made by The Economist (June 29, 2017) that ‘it seems that Mr Wilhite's entreaties are falling on deaf ears.’
The abovementioned verdicts, however, only make up 89% of all verdicts. Two other standpoints were found. The first was somewhat surprising: in some texts, speakers confuse acronyms and abbreviations, which leads them to the (erroneous) argument that with GIF, the letters in the word should be pronounced separately: ‘Technically it's an acronym so we should be saying ‘G.I.F’ anyways so we're all wrong.?’ (RaviPatel, 2016). However, with 24 occurrences (2.8%), this verdict is found only marginally.
The second standpoint is more interesting from a theoretical point of view. While 91.8% of writers prefer only one pronunciation, whether it be [gɪf], [dʒɪf] or spelling it out, there are 72 writers (8.2%) who say using both is possible: ‘Both are accepted, with the soft g given precedence’ (dchaosblade4, 2015). This seemingly contradicts one of the fundamental notions of prescriptivism, the suppression of optional variability. This notion, which was coined by James and Leslie Milroy in their seminal Authority in Language (Reference Milroy and Milroy1985), entails that when variation exists, prescriptivism will choose one of the variants and will strive to eradicate any others. For the pronunciation of GIF this largely holds true, but the fact that this minority position exists is interesting. It could be a characteristic of lay people prescriptivism, rather than published prescriptivism. It could also, however, be a sign of modern prescriptivism, as some evidence shows that there is a move towards allowing more variation in prescriptivism in general (Van der Meulen, Reference Meulen, de Bruyn and Berns2018; Reference Meulen, Chapman and Rawlinsforthcoming 2019), thereby increasingly moving towards a more descriptive standpoint.
The 899 statements contained 1393 arguments in total. While all arguments in the four groups were attested, there was a clear preference for the groups System and Authority (see Figure 1).
The other two groups of arguments, Effect and Use, played a smaller but not insignificant role in arguing the correct pronunciation, as they make up 13.4 and 9.2% of all arguments respectively. Within these groups, the argument that occurs the most is Use_General. This argument is used 40 times to support a [gɪf] pronunciation. Conversely, the argument that is used most to support [dʒɪf] is Use_Inventor, which is found 30 times, for example when Mathieu QL says that ‘90% of people are bent on saying it with a hard “G” and with logical reasoning to back it up’ (Reference Mathieu2014).
Even though [dʒɪf] was not the overall preferred option, Authority_Inventor was, with 210 occurrences, the second most used argument. Wilhite's statement, then, does seem to have had a definite impact, at least to such an extent that people wanted to reflect on the occurrence. Unsurprisingly, the argument is used almost exclusively in relation to [dʒif]. For the most part (65.2% of cases) it is used to argue in favour of this pronunciation, as TQ White II does when he says that ‘he [Wilhite] invented it. Inventing it includes naming rights on my planet. Show some respect’ (2013). However, a fair amount of cases (34.8%) also address the point when they disagree with Wilhite, such as ComradeUncleJoe: ‘Language doesn't work like that. Just because you name it doesn't mean your pronunciation is correct.’ (2015)
A host of other authorities is mentioned, chief amongst which are dictionaries, such as the Longman Pronunciation Dictionary. Other authorities include Siri, Wikipedia and Google Translate. Surprisingly, and perhaps also uniquely within the prescriptive canon, quite a few writers mention President Obama as the authority in deciding the correct pronunciation. Commentator Ali Diaz for example says that ‘Yall are so late this argument BEEN settled when obama said gif?’ (Reference Diaz2015). This seems to be based on a Tumblr post by the official White House account on 10 June 2014, and a Tumblr Q&A on the same day, in which Obama said that [gɪf] was his ‘official position’ (2014).
As expected, the presence of authority as an argument is rather different from written usage guides. For example, a random sample of 254 entries from the Hyper Usage Guide of English (HUGE) Database (Straaijer, Reference Straaijer2014) only contain nine references to authority, and these relate exclusively to writers such as Shakespeare and Chaucer. Of course, in none of the ‘traditional’ usage items included in HUGE do we know who the inventor was (if there even was one). Still, it shows just how rare reference to an authority is in prescriptivism. More importantly, however, the preference in the present data set differs from the patterns that Severin (Reference Severinto appear) found in her research on Reddit. There, deference to the dictionary as an authority was the most common.
5.2 Linguistically sophisticated
The arguments used to debate the correct pronunciation of GIF show another remarkable trend: they display a level of linguistic sophistication that is, arguably, not often found in traditional usage guides. For example, the most-used argument (260 times) is the fact that GIF is an acronym. Advocates of [gɪf] point to the fact that the first letter of GIF stands for ‘graphic’, and that the acronym should be pronounced the same way as this actual word is pronounced. ItzFish, for example, states that ‘It's an acronym. Graphics interchange format. As you can see, this would be pronounced as Jraphics Interchange Format’ (2013). Proponents of the [dʒif] pronunciation, on the other hand, point to acronyms like SCUBA, in which the first vowel is always pronounced as [uː], even though the corresponding word, ‘underwater’, is pronounced with an [ʌ]: ‘It's not Jraphics Interchange Format by that logic, NASA would be pronounced Nay-sa and SCUBA would be pronounced Scuh-buh’ (DHCKris, 2015).
Several other linguistic arguments are encountered in the dataset. For example, some writers use analogy to support their claim (‘GIF is like “gift” and should be pronounced the same’), but homophony is also used, and some writers even employ a comparative Germanic perspective. In most of these cases, a single example is used. In a few rare cases, however, the position of GIF is discussed within a larger linguistic system. Blogger Tahj Mayes, for example, does not use just one analogous example, but points towards systematic rules: when followed by front vowels, the g is pronounced as a ‘soft’ g, when the g is followed by any other letter, a ‘hard’ g is used (Reference Mayes2016). YouTube user ThioJoe (2015) goes even further:
There's about one million words in the English language and statistically about 1.59 per cent of them begin with a [g]. That's 19,500 words that begin with a [dʒ]. On Wikipedia there's an article that's literally a list of words where a hard [g] is used instead of a soft [g]. (…) There are 61 words on that list. Let's even round it up to a 100. One hundred out of 19,500 is 0.5 percent, so you would rather argue that [dʒif] is the exception to a rule that's followed 99.5 percent of the time, instead of admitting that you never looked up the grammar rule in the first place.
All in all, such linguistic arguments, as combined in the category System, represent 53.4% of all arguments used to argue the pronunciation of GIF. The arguments are not always relevant or even right, but their scope is far removed from the ‘clumsy, awkward or having a peculiar propriety’ of more ‘traditional’ usage advice’ (Anderwald, Reference Anderwald2012: 28).
Contrary to Steve Wilhite's intention, his speech did not codify the pronunciation of the acronym GIF, as the majority of speakers still prefer [gif] over [dʒif]. However, his announcement did seem to have given an impetus to (or may have even started) the discussion about this particular usage item. Moreover, many people mentioned his judgment as the deciding argument in choosing which side they were on. As there are few or no verdicts available from before Wilhite's speech, it is uncertain whether people actually changed their minds. Either way, a certain authority is in this case definitely awarded to the coiner. Additionally, speakers seem to show a notable preference for using linguistically informed arguments to strengthen their cause in the case of GIF, such as analogy and explanations of acronyms. In this sense, the GIF debate seems to differ not only from traditional prescriptivism, but also from other prescriptive debates online.
In general, this article makes an argument for both the fine-grained study of prescriptive arguments, as well as a more detailed study of prescriptivism online. Various authors (e.g. Beal, Reference Beal2010; Lukač, Reference Lukač2016; Vriesendorp, Reference Vriesendorp2016; Lukač, Reference Lukač and Tieken–Boon van Ostade2018) have looked at the way in which well known usage items are treated in the digital age. Less attention, however, has been devoted to possible new usage items that have originated online, or that have arisen in the digital era. Such usage items may not only be interesting in their own right, they can also give us a more detailed understanding of the rise and life of usage items.
MARTEN VAN DER MEULEN obtained his MPhil in Linguistics in 2015 from Leiden University, the Netherlands. He is currently pursuing a PhD at the Radboud Universiteit Centre for Linguistics (Nijmegen, the Netherlands), in which he looks at the interplay between language use and prescriptivism in Dutch. In addition to his research, he is an avid blogger, and has (co-)authored several popular scientific books about language, most recently on cursing. Email: M.vanderMeulen@let.ru.nl