“It’s you’re, not your”: Exploring misspelled words in YouTube comments

Kathleen Harris, University of Helsinki
Turo Hiltunen, University of Helsinki

Cite this article Open article PDF

Abstract

1. Introduction

The Internet is currently causing massive changes in how people both use and perceive language. Online communication has become a well-known source of numerous linguistic innovations that have gained attention both in and out of academics. Undoubtedly the most well-known of these are, for example, emoticons (such as smiley faces), acronyms (LOL and OMG), and textual representations of conversational prosody such as typing in capital letters to indicate shouting. Other examples that are associated with online communication include various types of deviations from spelling and punctuation norms, such as multiple punctuation, the use of asterisks for emphasis, and eccentric spelling (Danet 2001).

These characteristics of casual online communication have also garnered much attention in popular culture and the media. It is a common view among the general public that these features are signs of impoverished rather than innovative language, and are causing a decline in literacy and “proper” use of the English language. [1] Journalists and the media have played a major role in promoting the idea that the English language is in decay because of the Internet by consistently producing articles that claim evidence of widespread decline in standard usage.

Research concerning current Internet-related language change remains inconclusive for the most part, however. For example, in Language and the Internet (2001), David Crystal investigates the use of “Netspeak” in various areas (e.g. e-mails and chatgroups), concluding that it is an enrichment rather than a threat:

I don’t see the Internet being the death of languages, but the reverse... There is no indication, in any of the areas I have examined, of Netspeak replacing or threatening already existing varieties. On the contrary, the arrival of new, informal, even bizarre forms of language extends the range of our sensivity to linguistic contrasts. (Crystal 2001: 241–242)

Other CMC studies on text speak, or textisms – language used in mobile text messages and online instant messaging (IM) – have discovered conflicting correlations between the use of nonstandard language and standard literacy, ranging from a neutral effect to possible negative effects (Drouin & Davis 2009; Kemp 2010; Rosen et al. 2010; Varnhagen et al. 2009). On the other hand, both Bushnell et al. (2011) and Plester & Wood (2009) have even suggested a positive connection between so-called text speak and above-average literacy in adolescents.

In addition to conflicting claims of decline in literacy, the frequency of use of nonstandard forms in some genres of CMC is actually much lower than the general public and media appear to believe. For example, Tagliamonte & Denis (2008) have shown that less than 3% of Instant Messages sent by teenagers are of a nonstandard form. On the other hand, studies estimate that anywhere from 5-20% of SMS messages sent by teenagers and young adults contain nonstandard forms, and suggest that the estimates may be even higher for younger children (De Jonge & Kemp 2010; Ling & Baron 2007; Plester et al. 2008; Thurlow & Brown 2003). In any case, claims about a decline in literacy are often contextually ambiguous or even fabricated, and according to Thurlow (2006) serve only to stir up moral panic that has no concrete basis.

Unlike the studies referred to above which focus on IM and SMS, this study concentrates on a different genre of CMC. Comment sections, which can be found on a variety of websites such as image sharing sites, social media, and online newspapers, provide the data for our research. Little previous research has focused on comment sections, and analysis could provide insights into how people use language in this specific online setting. The aim of this study is to explore the prevalence of the use of a particular English nonstandard linguistic variant that is commonly encountered within online comment sections: the possessive your being used instead of the standard contraction you’re. Example (1) demonstrates a comment containing the use of your, taken from our YouTube corpus. The video that the comment refers to features a young boy that gives an inspirational and motivational pep talk to the world. The video went viral in February of 2013 and has gained nearly 30 million views and over 80,000 social media shares worldwide, with a hugely positive response by commenters and viewers. [2] The video is available below as well.

(1)	your awesome (YouTube: A Pep Talk from Kid President to You)

By current English language standards, this comment would be considered grammatically incorrect due to the use of the nonstandard your variant rather than the standard you’re. However, it cannot be definitively understood why the commenter chose to use the nonstandard variant. The lack of capitalization and punctuation suggests that the commenter chose to use your in order to economize time spent typing, but it is possible that there are other factors involved. In fact, it may not have been a choice at all, and is indeed a genuine error.

Our impression, based on both CMC user experience and the results of our analysis, is that nonstandard your is somewhat different from other nonstandard variants that occur within CMC. Not only is it extremely common within YouTube comments, but it is also salient to the extent that it could be considered to epitomize the public view of CMC by exemplifying the perceived decline in literacy. The aim of this article is to investigate whether this hypothesis can be supported by data from the online comments of two popular content-sharing sites, YouTube and 9GAG.

The idea of nonstandard your being salient can often be seen in online interactions and user reactions to its use. It is often cited by Internet users as being one of the most irritating errors that another user can make. Many people feel strongly that the nonstandard variant should be something a person learns to avoid, and many are quick to correct a person that uses it. Even so, you’re and your are frequently confused. Mahdawi (2012) reports on a Twitter feed @YourInAmerica, which collects examples of nonstandard your in tweets where people complain about other people’s language:

(2)	If your in America and can’t speak English then you shouldn’t own a donut shop. (@YourInAmerica)

This shows that the forms are indeed easily confused, even by people who see themselves competent to pass judgment on other people’s language use. Correcting somebody’s grammar and spelling is, of course, extremely common in online interactions (see Section 2.1).

An interesting aspect of nonstandard your is that it may also be used intentionally, both in CMC and non-CMC settings. The clothing company Wet Seal, whose target demographic is teenage girls, provides us with an example of the latter situation. In 2011, the company produced a T-shirt that had the words If your single, so am I printed across the front. Many customers complained about the misspelling, to which the company replied on Twitter with the following sentence:

(3)	It’s a fashion statement ... I am jealous for you’re [sic] keen eye for grammatical errors though (Goldwert 2011).

The company’s response indicated that the misspelling was indeed intentional, and the perceived error was not fixed. The deliberateness of the misspellings both in the T-shirt and the company’s response demonstrate how this particular variant can be used to get a reaction from people. In that sense, the salience of your is demonstrated.

In this article, we present the results of a corpus-based study into the use of nonstandard your within comment sections. Our data includes comments from two popular video- and image-sharing sites, YouTube and 9GAG. As will be seen, each of these sites have their own characteristics, defined by their user demographics, typical forms of interaction, and distinct online reputations. Our research questions are as follows:

1. How common is the use of your in place of you’re within the YouTube and 9GAG comment sections?
2. When compared with other linguistic variables, is the use of your distinct?
3. What are some possible explanations for the use of your?

To answer these questions, we compare the frequency of nonstandard your within the online comments of the two websites. In addition to you’re/your, we also study a number of other linguistic variables that are often associated with informal CMC: I’m/Im, he’s/hes, don’t/dont, that’s/thats, I/i, are/R, you/U, for/4, it’s/its, too/to. Like you’re/your, these variables often draw criticism and have been used as examples of literacy decline. Our analysis reveals an interesting finding-- not only is the your variant used quite commonly within YouTube comments, it is actually preferred over the standard you’re. This result is not mirrored within the 9GAG comments, which demonstrate a distinct preference for the standard variant. Based on this finding, we discuss a number of possible factors and influences such as genuine error, economy, anonymity and community identity, specific pragmatic functions, and adolescent language.

This article is structured as follows. Section 2 provides the theoretical background on user-generated comments in online discussion forums. Section 3 describes the materials and methods used in this study, and Section 4 presents the results. Section 5 considers the implications of these results.

2. Background

2.1 Norms of spelling in the Internet age

Milroy and Milroy (1991: 56) write that “twentieth century English spelling is almost absolutely invariant”, meaning that there is little room for variation in spelling in most mediums. In other words, excluding regional variation, spelling is a binary choice in which there is either a “correct” or “incorrect” choice. Making the wrong choice can be highly stigmatized, and often results in an individual being harshly judged as being uneducated, unintelligent, or low class (Milroy & Milroy 1991: 2).

Recently, however, fixed conventions of spelling have been challenged. With the emergence of CMC as a new medium, the rigid rules that govern spelling are giving way to a more laissez-faire attitude towards standard language. One manifestation of this development is the frequent use of playful non-standard spellings (e.g. abbreviations, eye dialect and colloquial contractions) in SMS text messages, which are generally seen as contextually appropriate (Tagg 2009: 127). For Shortis (2009), the whole idea of invariant spelling is becoming outdated:

The spelling practices associated with the age of print appear to be giving way to a looser, more permeable sense of what counts as spelling. Spelling is becoming a deployment of choices from a range of options including the standard English form among others (2009: 17).

This view is shared by Baron, who uses the term linguistic whateverism to refer to the younger generations’ “marked indifference to the need for consistency in linguistic usage” (2008: 169). However, Baron points out that the relaxed outlook on prescriptive rules is not merely due to the use of computers, but are rather caused by changes in education, the social agenda, and an increase in the speed in which we compose texts (2008: 170). From this perspective, it is clear that orthography takes on a more important role in computer-mediated communication than it does in written language in general (cf. Crystal 2011: 67).

While attitudes towards linguistic innovations have clearly become more relaxed, spelling is still a somewhat divisive topic within Internet culture. Many Internet users still remain critical of nonstandard spelling, even despite the fact that many online participants are likely communicating in a language which is not their native language. Many websites, including YouTube and 9GAG, use English as a lingua franca, with many of the users enforcing Anglo-centric language standards by presupposing that other users should uphold native levels of fluency. Genuine spelling errors – mistakes that are not considered to be a typographic error or innovation – often incite strong emotional reactions from other users. It is common to see users often referred to as grammar nazis correcting the spelling or grammar of other users online, with the corrector’s intention ranging from polite correction to hostile ridicule. [3] Consequently, it is also common for users to use misspellings as a tool in order to garner corrections and responses, incite anger, or to make a humorous statement. The phenomenon of language correction has become so common online that there are a number of informal Internet “laws” that have been declared concerning corrections. For example, Skitt’s Law states that “any post correcting an error in another post will contain at least one error itself”. [4] On the other hand, it is possible that non-native speakers of English are in turn influencing online language standards by contributing to relaxed attitudes towards nonstandard use.

2.2 Computer-mediated communication (CMC)

The current study builds on previous work within computer-mediated communication (CMC), an expanding area of study at the interface of linguistics and communication studies. In this study, we follow Herring’s (2007) definition of CMC: “predominantly text-based human-human interaction mediated by networked computers or mobile telephony”. However, it could be noted that scholars are far from unanimous about terminology; other terms with roughly the same meaning include Internet linguistics (Crystal 2011), digitally-mediated communication (Crystal 2011), digital networked writing (Androutsopolous 2011), electronically-mediated communication (Baron 2008), keyboard-to-screen communication (Jucker & Dürscheid 2012), electronic communication (Herring 2012) and Internet-mediated communication (Yus 2011).

Technically, CMC represents written language produced as a result of typing, and it is therefore possible to edit and revise the message before it is sent to the receiver, as in most other forms of written language. Interestingly, however, CMC is often seen as separate from the two traditional modes of communication, or a hybrid of both spoken and written language (Androutsopoulos 2011, Baron 2008, Crystal 2011, Herring 1996). [5] As previously mentioned, CMC often lacks the invariance and stability of written language. From another perspective, CMC is similar to speech in that it can be informal and spontaneous, and yet it lacks extralinguistic cues found in face-to-face such as intonation and volume. In order to replace characteristics of both written and spoken language that are lacking, CMC shows a remarkable propensity for adaptation.

2.3 Online comments as a form of CMC

Throughout this study, we shall use the term online comments to refer to user-generated comments, posted on a website as a reaction to another text, video, or image (see also Sairio, this volume, for a study of online comments on Amazon.com ). Such a text could be a simple piece of writing, such as a typical blog post, but here we focus on comments that are made in response to a multimodal prompt consisting of sound, images, and video. [6]

There are some basic characteristics that the majority of online comments have in common. They tend to be fairly short, between one and twenty words. In terms of page layout, they are grouped under the video or image, and they commonly refer to some aspect of it. Some comments may be specifically addressed to another commenter, but all comments can be read by anyone who accesses the page. The language used is often colloquial, and the comments may contain typos and other errors, suggesting that they may not have been edited before being posted.

The status of online comments as a genre is currently an open question. Genres are typically defined in terms of a shared communicative purpose and conventions (Swales 1990; Herring and Paolillo 2006). Using these criteria, it is certainly possible to see online comments as a distinct genre or a subgenre within CMC, given that users clearly recognize the basic conventions of online comments and follow them. On the other hand, the genre system of the Internet has been shown to be complex and dynamic, given that different genres are easily mixed within one web page (Santini 2007).

Much of online communication can be classified as either synchronous, which requires users to be online simultaneously, or asynchronous, which does not (see Herring 2001). However, these labels are somewhat problematic in the context of online comments, which may serve both of these purposes. A user may have a real-time discussion with other users (which is often the case) or a user may post a comment without any expectation of real-time interaction.

In addition, much variety exists within comment sections. Different websites and communities tend to evolve their own system of commenting, etiquette, and linguistic characteristics that vary according to the demographics of the users, which often results in each website having its own unique identity, reputation, or dialect. For instance, FutisForum, a Finland-based internet football forum, features passionate debates about football where users draw on a variety of linguistic resources, including multilingualism (Kytölä 2013, this volume). On the other hand, on the online auction site eBay, buyers and sellers give feedback about each other after a successful transactions has taken place, and this feedback will be visible on their respective profile pages (Wood and Smith 2005: 123).

Community identity and reputation are particularly relevant to YouTube, one of the two websites which we focus on in this study (see Section 2.4 below). YouTube users as a whole also have the very distinct online reputation of producing supposedly illiterate comments that are full of spelling and grammar errors. In fact, YouTube comments are often referred to by Internet users as being the epitome of nonsensical online comments, both in content and structure. Take for example, the following user-submitted definitions (Examples 4 and 5) for YouTube comments on the popular website Urban Dictionary: [7]

(4)	Such a cesspool of illiteracy, misogyny, racism, idiocy, hatred, violence, barbarism, intolerance and general lack of humanity has not been seen since the middle ages.
(5)	The ultimate example of human stupidity.

Another example of this perception is Randall Munroe’s webcomic xkcd (Munroe 2006). A comic entitled YouTube opens with the following text: The Internet has always had loud dumb people, but I’ve never seen anything quite as bad as the people who comment on YouTube Videos (Figure 1).

YouTube’s commenting reputation is so prolific that a number of steps have recently been taken in an attempt to remedy the public’s perception. For example, in the past, Google has made an attempt to solve the problem by strongly encouraging users to use their real name (Hill 2012). [8] Even more recently, changes to the entire commenting system have been implemented that are expected to assist in filtering out unwanted comments as well as require a higher level of nonanonymity (Rose 2013). [9] While these changes are meant to focus mainly on content rather than language use, it may also affect a user’s decision to use a nonstandard form. [10]

The previous examples demonstrate the common belief within Internet culture that the YouTube comment section is the epitome of the simplistic and “wrong” nature of language used online, in which people often attempt to draw a correlation between the use of text speak and lack of intelligence.

2.4 YouTube

The first website included in the analysis, YouTube, is a video sharing website that was created in 2005 and has since become popular worldwide. [11] With over 800 million unique users visiting the site each month, YouTube is the third most popular website in the world (YouTube, Alexa.com), after Facebook and Google. Users originate from many different countries, with 30% of traffic originating from the United States. The site also reports that it has a broad demographic, with users typically being anywhere from 18-54 years old (Statistics, Youtube.com).

The main function of YouTube is to allow “billions of people to discover, watch and share originally-created videos”. [12] Users can upload a video to their personal channel on the site, where other users can view, rate, and post comments about the video. It is these user-generated comments that are the focus of this study.

Along with being a platform for uploading and watching videos, YouTube is also a popular online social network, allowing users to create a personal profile page and interact with other users (e.g. Lange 2012, Wattenhofer et al. 2012 and the references therein). At the same time, it is clearly not a prototypical online community, which according to Jones (1997) is characterised by four qualities: a minimum level of interactivity, a variety of communicators, a common public space, and a minimum level of sustained membership. For one, the majority of users of the service merely watch videos without ever commenting on them (YouTube videos are also frequently embedded on other websites), and in this situation we can hardly talk about a sustained participation in a community. Having said that, it is not uncommon that individual videos attract users who post comments regularly and engage in a debate with other users, creating a virtual social network within the larger community. Additionally, it is not uncommon for specific users to become well-known on the site, attracting large numbers of followers and fans. Posting comments enables users to express feelings towards these video makers, establishing new social connections and maintaining existing ones (see further Lange 2012). In addition, the fact that YouTube comments seem to have a distinct reputation, as previously described (section 2.3), supports the idea of YouTube as a virtual community also on a global level. Wattenhofer et al. describe YouTube as a "content-driven network" (2012:356), where users interact through uploaded video content rather than through direct messages, which sets it apart from traditional online social networks.

A user’s comment can be “voted up” by clicking a thumbs up icon or “voted down” by clicking a thumbs down icon. The comment with the most up-votes is displayed directly below the video. Comments that receive an undetermined amount of down-votes are hidden from the comment section in an attempt to keep the site clean from unpopular statements that are often abusive or vulgar. It is also possible for the uploader of a video to disable the commenting option on a particular video and to delete specific comments.

The level of anonymity on YouTube can vary. YouTube is a public website, meaning that all comments are viewable by unregistered users. However, to upload a video or post a comment an account is required. User accounts require a user name, but allow for anonymity. A user may use an account that is linked to their Google account, which is likely to use their real name rather than a user name. A user may also use an account that has no identifiable associations with their real life, but that maintains a certain reputation or persona within the online community. Wood and Smith (2005: 64) use the term “pseudonymity” to refer to the mid-position at the continuum between one’s real-life identity and complete anonymity. A user may also use an account that they perceive to be completely anonymous and unidentifiable by others. The two latter levels of anonymity are respectively known as technical and social anonymity (Hayne & Rice 1997). It is most likely that YouTube hosts a mixture of these two types of anonymity in which the sheer number of accounts causes the majority to fall into the socially anonymous category.

2.5 9GAG

The second website included in the analysis is 9GAG. [13] Like YouTube, 9GAG is a social media site containing user-generated comments, but is also different from it in several important respects. Most importantly, user comments on 9GAG.com are normally not anonymous. While 9GAG.com is a public page allowing users to view images and comments without logging into a registered on-site account, users must have a Facebook profile in order to post comments. It is of course possible and likely that some users create a Facebook profile that uses a fake name and photo, because they might not feel comfortable posting on sites other than Facebook with an account that is attached to their real life. [14] In addition, some comments may originate from a so-called troll account, an account on Facebook or any social site that is used specifically for trolling. However, our impression is that the majority of users encountered on 9GAG appear to be using real Facebook accounts with suspiciously troll-like profiles only appearing very rarely. [15] Therefore, even though both technical and social anonymity are possible on the site, the comments posted on 9GAG.com can reasonably be treated as non-anonymous.

Like YouTube, content on 9GAG is user-uploaded, but rather than a video, it consists of an image or an image with superimposed text. The images usually focus on humor, memes, and Internet culture in general. The site has a voting system in which users vote for their favorite images, and a comment section beneath each image. 9GAG is currently ranked as being the 324th most visited site in the world (9GAG, Alexa.com), which is high for the type of site that it is. 9GAG has about 67 million unique visitors per month and more than 2 billion page views (Gannes 2012). The audience is mostly 35 and under, with the majority being between the ages of 18-24. Gender is roughly equal, and the site is most often accessed from universities. The site is popular worldwide, but the top three countries with the most visitors are the United States, Brazil, and Germany (9GAG, Alexa.com).

3. Material and methods

3.1 Data

The materials used for this study consist of two corpora, containing user comments collected from two websites described in the previous section, YouTube and 9GAG. Table 1 provides a summary of these corpora.

The YouTube corpus consists of 105,193 words that were manually collected from the user comment sections of 86 different YouTube videos (see the Appendix for a full list of the videos.) The data were collected in two parts, with the first between October and November of 2011 and the second between January and March of 2013. In order to avoid bias in the selection process, the videos that provided the comments were accessed directly from the main page of the website. The main page typically contains 19 videos spanning a variety of topics that are currently popular and are likely the most-viewed videos of the moment. By viewing the page without logging into a YouTube account, clearing all cookies and browser cache, and adjusting the page setting to Worldwide, a default main page that was uninfluenced by previously watched videos or other factors could be achieved.

Approximately 1,000 to 1,500 words were manually collected from each video by copying and pasting the comments into a .txt file. [16] The comments were accessed by clicking the “All comments” link beneath the video, which displays every comment posted. The default view for the page sorts the comments by the time that they were posted. However, it is optional to sort comments as a thread, in which replies are nested beneath the comment that they are aimed at. The thread option was used for this study in order to have a clearer view of context. The comments were then collected consecutively, with the following exceptions: (i) non-English comments, (ii) spam comments such as advertisements, and (iii) posts by users urging others to view their videos. YouTube receives massive amounts of spam, but is somewhat efficient at removing the comments with the help of other users that “flag” a comment as spam by clicking an icon. The average length of YouTube comments within our data is 10 words and 56 characters.

The 9GAG corpus consists of 56,719 words collected from the comment sections of 84 uploaded images. Because the number of comments as a whole tends to be much lower than on YouTube, no attempt was made to collect a certain amount of words from each image. The majority of images did not have anywhere near 1,000 words, and the word counts varied depending on the popularity of the image (as opposed to YouTube, which tends to have high word counts across the board on popular videos). Therefore, all available comments were collected from an image. The images that were included in the study were taken from the main page of the site, which displays the most popular images of the moment. The main page is updated frequently, with images being added as they gain popularity. By clicking an image, a new tab opens in the browser displaying the image and the comments below it. Comments are arranged in a thread format, displaying replies nested beneath the comment that is being addressed. As with the collection of YouTube comments, all comments were included except for non-English comments and spam. The average length of 9GAG comments within our data is 11 words and 58 characters, which is similar to the average length of YouTube comments.

3.2 Method of analysis

The purpose of the quantitative analysis is to compare the relative frequency of the nonstandard your in 9GAG and YouTube, as illustrated in example (6) below. In other words, we are interested in how frequently online commenters use your as a variant of you’re, illustrated in (7), and whether there are differences between these two image and video sharing websites.

(6)	[…] No wonder your so angry. (YouTube: A Pep Talk from Kid President to You)
(7)	[…] And yes, never stop going after your dream. just make sure you’re not dreaming somebody else’s dream. ;) (YouTube: A Pep Talk from Kid President to You)

To this end, we used AntConc 3.2.4 to retrieve all instances of your and you’re from both corpora. [17] After manually removing instances of the possessive your (e.g. What’s your point?) from our database, we counted the relative frequencies of these two variants. [18] We then used the χ² test to investigate the null hypothesis that the proportions of the variants are the same in the two populations that the corpora represent, with the significance level set at 0.05. [19]

To better understand why users would choose a particular variant, we included four types of additional linguistic variables in the analysis: apostrophe omission, capitalization, replacement, and commonly confused expressions. Apostrophe omission includes four types of choices: I’m/Im (examples 8–9), he’s/hes (10–11), don’t/dont (12–13), and that’s/thats (14–15).

(8)	Damn it I’m 16 and my hairline is pulling back haha. This songs the shit (YouTube: Bald Guy)
(9)	Im 16 and loosing my hair : / this video makes me feel 10x better. you are the bomb Bald Peter! (YouTube: Bald Guy)
(10)	Two days ago this had 50 views... yesterday it had 50,000 views... an hour later it had 300,000 views.. NOW IT HAS 1,000,000! This guy deserves it! He’s hilarious! (YouTube: Bread & Milk)
(11)	hes in hysterics lol (YouTube: Bread & Milk)
(12)	I don’t understand how this is so popular and funny (YouTube: Leonard Cooper)
(13)	Normandy is where a major battle of WWll took place...sad that you dont know that (YouTube: Leonard Cooper)
(14)	That’s awesome! Nice video (YouTube: Red Hot Nickel Ball)
(15)	thats cool (YouTube: Red Hot Nickel Ball)

Under capitalization, we consider the capitalized first person pronoun I and the lowercase nonstandard variant (16–17) (cf. Tagliamonte & Denis 2008). [20]

(16)	I tried opening a plastic package with a can opener and it didnt work. Too bad because I really thought it would help me out :( (YouTube: Awesome Solutions)
(17)	once i tried to shave with a cheese grater. that sucked (YouTube: Awesome Solutions)

We use the term replacement to mean the practice of replacing a word or part of a word with a single letter or number. Three types were included in the study: are/R, you/U, and for/4 (examples 18–23). [21]

(18)	Tattoos are very dangerous the ink is made of mercury and rat poison and you can die in a matter of 10 years depending how many you get!!!! Just a warning!!! (YouTube: Cat Woman)
(19)	You both r goin to be wonderful parents one day:) (YouTube: Cat Woman)
(20)	I laughed because you are so proud of a dumb dance haha! (YouTube: Harlem Shake- Black Edition)
(21)	Nasty unemployed people in this vid they probably got like $10 to be on tv. Don’t bash the harlem shake because it funn y and everyone loves it. U just hate that d world don’t give a shit about u (YouTube: Harlem Shake- Black Edition)
(22)	Ain’t nobody got time for yo attitude. (YouTube: Sweet Brown)
(23)	Haha Lord Jesus, ain’t nobody got time 4 that. (YouTube: Sweet Brown)

Finally, our term commonly confused expressions subsumes two easily confused words: to in place of too and the use of the possessive its in place of contraction it’s (examples 24–27). [22]

(24)	Great, now she’s going to think it’s funny to rip up her homework when she goes to school.... :p (YouTube: Baby Laughing)
(25)	its crazy how believable this looks lmao (YouTube: Inaugeration 2013)
(26)	yeah i saw it to. when i paused just before. The other guy is much much bigger than Beckham (YouTube: Short Film)
(27)	Futurama. i can name animated shows too! (YouTube: Ultimate Robot Takeover)

These four additional variable categories are similar to the main variable in that they are binary and consist of a standard and a nonstandard variant. This being the case, we can compare the relative frequencies of the nonstandard variants of different variables in order to see whether the main variable stands out in some way. In other words, if users choose to use your in order to economize their typing, we would expect to find similar frequencies in the linguistic variables that are indicative of economy, such as apostrophe omission. However, if the relative frequency of nonstandard your turns out to be markedly different from that of the other nonstandard variants, it is likely that other factors account for its use.

4. Results

This section presents the results of the quantitative analysis, and their relevance to the research questions of this article will be discussed in Section 5, together with examples from the data.

4.1. You’re/your

Our quantitative analysis shows that, compared to YouTube, the users of 9GAG show a much higher preference for the standard variant in all cases. In addition, it appears that you’re/your is indeed different from the other variables: your occurs far more frequently than other nonstandard forms. In fact, it is the only variable in which the nonstandard form is actually preferred over the standard form in the YouTube data. The use of your in place of you’re occurs 103 times in the YouTube, while the use of the standard you’re occurs 90 times. In other words, given the binary choice between standard and the nonstandard form, the latter is preferred by YouTube commenters in 53.3% of the cases (see Table 2). Within the 9GAG data, however, the variant your only occurs eight times, accounting for a mere 7.3% of the total uses.

The observed difference in the percentages appears to be considerable, as can be observed in Figure 4, which shows the distribution as a mosaic plot. Nonetheless, a χ²-test was performed to determine whether the two variants were equally preferred by users of YouTube and 9GAG. The test shows that the distributional difference between the corpora is statistically significant (χ²=63.5, df=1, p<0.001). In other words, users of YouTube are much more likely to choose the nonstandard variant your than 9GAG users.

4.2. Apostrophe omission

Omitting apostrophes from contractions is a common practice in CMC that is often used as an economizing tool. Our analysis considers the omission of apostrophes in four contexts (I’m/Im, he’s/hes, don’t/dont, and that’s/thats). The numbers of occurrences for each pair are found in Tables 3–6.

As shown in Figure 5, three of the four variables are similar to each other in that the relative frequency of the nonstandard variant is much higher in YouTube. The exception is the variable I’m/Im (the upper left panel), where it is the 9GAG users who show a preference for the nonstandard spelling without the apostrophe.

4.3 Capitalization

Table 7 shows that both YouTube and 9GAG users employ the capitalized form I the majority of the time. However, as illustrated in Table 7 below, the incidence of the nonstandard form is again significantly higher in the YouTube data, with over one fourth of the instances being spelled with a lowercase letter, as opposed to 14 % in 9GAG.com comments (see the upper left panel of Figure 6).

4.4 Replacement

Compared to the other variables, replacement is relatively rare. In other words, the vast majority of users in both data sets prefer to use the standard form. The distribution of the forms is shown in Tables 8–10, and visualized in Figure 6.

Although the frequency of nonstandard forms is low, their proportion is always higher in YouTube, which is consistent with the other variables studied. The nonstandard variant of the variable for/4 is particularly rare, occurring merely 8 times in our data (it is the only variable of the study that does not yield a statistically significant result). The most common nonstandard variant involving replacement is U, which accounts for 10% of the instances in of you/U in YouTube and 3% in 9GAG.

4.5 Commonly confused expressions

Finally, the two commonly confused expressions (too and it’s) display a distribution similar to the variables analyzed previously: the standard form is more common in both corpora, but nonstandard forms are comparatively more common in YouTube. The relevant distributions are shown in Tables 11–12, and as mosaic plots in Figure 7.

5. Discussion of findings

To sum up the results of the quantitative part of our study, the analysis of corpus data confirms the impression held by many Internet users: nonstandard your is definitely a common occurrence. This form is particularly frequent in YouTube comments, where it is used more frequently than its standard counterpart you’re. Compared to the 9GAG corpus, the frequency of nonstandard your is strikingly high.

The analysis of the additional linguistic variables (capitalization, apostrophe omission replacement, commonly confused expressions) enables us to look for explanations for the observed distribution of our main variable of interest. The analysis revealed a consistent difference between the two corpora: YouTube comments consistently have a higher proportion of nonstandard variants, indicating that the commenters have a very relaxed attitude towards linguistic norms, including orthography. [23]

We can establish with the help of quantitative analysis that there is a significant difference between the ways in which users of YouTube and 9GAG use nonstandard forms, however we cannot tell what has caused it. To do this, we need to explore the data in qualitative terms. In order to suggest answers for our second and third research questions, we will now discuss six possible factors ‒ genuine error, economy, adolescent language, anonymity, community identity, and specific pragmatic functions ‒ that could be influences on the use of your. While it is possible that there are other factors involved, we have decided to focus on these in particular as we believe they are likely to be the most influential.

5.1 Genuine error

The first possible explanation is that the commenters using the form your are making a genuine error. This idea has some initial appeal: the distinction between your and you’re is often problematic for both native and non-native English speakers, both on and offline. Moreover, unrevised electronic texts (such as e-mails) frequently contain misspellings, which are acceptable as long as they do not seriously interfere with the message (Crystal 2001: 111). It is therefore reasonable to assume that at least some of the occurrences of the nonstandard variant are unintentional. YouTube is used worldwide, and the users’ proficiency level varies considerably, as the following examples demonstrate.

(28)	This is pretty cool but why can’t they sell tickets with #’s on them or just have a ticket with your seat # on it.. Who cares if your first or last as long as you get your seat and don’t miss the game (YouTube: Japan Culture Shock)
(29)	your right about that there all one big joke Herman Cain Perry and newt too (YouTube: Now is the Time)

In example (28), it is difficult to determine whether your might be a genuine error. The commenter uses standard language for the most part and there is no indication that it is a mistake. The fact that the commenter also uses possessive your correctly is interesting.

The comment quoted as example (29), which refers to American politicians, is likely to be written by an American native speaker of English, because of the content of the message. In this case, the use of your is probably a genuine error, given that the comment also contains another error, the use of there in the place of their.

However, the fact that in YouTube your is even more frequent than the standard variant is unlikely to have been caused only by users making genuine mistakes. And even if this is the case, then it seems plausible that your would have an occurrence rate similar to the other nonstandard forms that were examined. Our data shows, however, that compared to your the other nonstandard forms are much less frequent. The second most frequent nonstandard form is its (in place of it’s), yet even that form only has an occurrence rate of 39.7%, much lower than that of your (see Table 12 above).

In addition to the disparity in occurrence rates across variables, the results from the 9GAG corpus suggest that something other than genuine error is influencing the use of your in YouTube comments. In 9GAG, nonstandard your occurs at a rate that is roughly similar to the other variables. If the use of your were indeed due to genuine error, we could expect to see similar frequencies in both corpora. However, as both YouTube and 9GAG have similar demographics in that both sites report users from a wide variety of ages and linguistic backgrounds, it is unlikely that the high relative frequency of your would be due to genuine error.

5.2 Economy

Due to the speech-like quality of CMC, users often seek to economize their typing by shortening words and omitting letters, punctuation, capitalization, and apostrophes (Crystal 2001; Androutsopolous 2011). This being the case, it is likely that economy is responsible for some of the occurrences of your. Take for instance Examples (30) and (31):

(30)	This was not the best half time show ever, unless your a big fan of her. I thought it was alright and she’s a great singer so no need for death threats guys. (YouTube: Beyoncé Super Bowl Halftime Show Rehearsal: Day 1)
(31)	thats because your a fucking tool (YouTube: Harlem Reacts to ‘Harlem Shake’ Videos)

Example (30) does not demonstrate any indications that your would be used in order to economize the typing process. It begins with a capital letter, uses standard punctuation and commas, capitalizes I, includes an apostrophe in she’s, and uses no shortened words or replacements. Conversely, Example (31) does not begin with a capital letter, uses no punctuation, and does not include an apostrophe within thats. Unlike Example (30), Example (31) could therefore be interpreted as being indicative of economical typing.

Our findings suggest that economy is definitely not the only factor responsible for the use of nonstandard your, possible not even the main one. With three of the variables investigated ‒ apostrophe omission, lack of capitalization, and replacement –the use of nonstandard forms is characteristically driven by economy, yet the rates of the nonstandard forms were far below the occurrence rate of your in each case. For example, the four pairs of words that were examined in regards to apostrophe omission had an average rate of 27.2% for the nonstandard forms within the YouTube corpus, while your occurred 52.8% of the time (see Tables 2–6). If users were preferring to use your solely as a means to expedite their typing, we would expect it to occur at a similar rate as the phenomena examined within the three categories. In the case of the 9GAG corpus, however, the results are comparable: nonstandard forms in the apostrophe omission category had an average rate of 6.5%, which is quite near the 7.3% rate for your.

Also of interest is a comparison of the frequency of the I/i and you/U variables with data from Tagliamonte and Denis (2008) on instant messaging (IM) data. While the use of U occurs at roughly similar rates throughout the YouTube, 9GAG, and IM corpora, the use of nonstandard i demonstrated drastic differences: it is much more common in Tagliamonte and Denis’s IM data, where it accounts for c. 75 percent of the instances. While Tagliamonte and Denis did not study the use of your in their IM data, the results indicate that perhaps we would expect higher rates of occurrence if your use is common due to a desire to economize typing.

5.3 Anonymity and community identity

People who engage in online communication can often choose how they want to present themselves, and in that way have some control over how they are perceived by others (Wood and Smith 2005: 63). Anonymity in CMC and its resulting disinhibition (Suler 2004) is thought to be a source for a number of interesting online phenomena such as trolling, flaming, and cyberbullying.

It seems possible that anonymity could also affect a user’s choice to use standard or nonstandard language, such that there would be less pressure for using standard language on sites that allow anonymity, as the users know that they will not lose face for misspelling. This is likely to be the case in Example (32); we believe that the commenter would not have produced this kind of comment if they had not been able to write anonymously. This comment demonstrates a stereotypical YouTube comment through its negative content, nonstandard spelling, and use of caps lock. The fact that the commenter makes a point of writing that they think they spelled a word wrongly while simultaneously using several nonstandard variants leads us to believe that this comment could be considered to be an example of trolling as well.

(32)

okay im with you. (im christain. I think i spelled that wrong) shit. The whole youtube ganged up on 09ast or whatever the fuck his username is. All i can say is, FUCK YOU PIECE OF SHIT! ASIANS ARE SAVAGES? YOUR THE FUCKING SAVAGE THATS SCUM. I HOPE GOD SENDS SOME FEAR TO KILL YOU IN YOUR SLEEP, THEN YOU GO TO HELL AND THE DEVIL TORTURES YOU TILL THE END OF TIME. IF GOD DOESNT SEND SOMEONE TO KILL YOU, IM GOING OVER THERE AND PUSH A PILLOW OVER YOUR FACE AT NIGHT TILL YOU STOP

The two websites included in our study each offer a different level of anonymity. As previously mentioned, YouTube offers a high level of anonymity, whereas 9GAG requires users to post comments using their Facebook account attached to their actual identity. If the hypothesis is correct, then we would expect to see more instances of nonstandard your in the YouTube compared to 9GAG. This is precisely what goes on in our data; as shown in Table 2 and Figure 4, the frequency of your is significantly higher in YouTube than in 9GAG, which provides strong evidence for the hypothesis that your is linked to the level of anonymity.

That said, we believe that the level of anonymity is not the only relevant explanatory factor. In particular, we suspect that awareness of and identification with the virtual YouTube community are important factors along with anonymity, possibly even more important. As previously discussed (Section 2.3), YouTube commenters have a reputation for nonstandard language and misspellings. For this reason, it could be that users of YouTube not only feel more free to ignore standard rules, but also feel encouraged to do so. In tandem with anonymity, users may incorporate the YouTube community identity and play with language in a way that users elsewhere do not.

5.4 Specific pragmatic function

Next, we shall consider the hypothesis that nonstandard your would be intentionally used for expressing specific pragmatic meanings commonly associated with CMC and Internet culture in general, for example trolling, memes, or flaming. In this scenario, users would intentionally use your to express conventionalized meanings, which the readers would recognize and be able to decode. If this is the case, the prominence of these functions could account for the high frequency of your we observed in our YouTube data.

There is evidence that intentional errors have acquired specific pragmatic functions in online communication. For example, Internet memes such as Lolcats are rich in nonstandard spelling and grammar, which are used mainly as a way to produce a satirical or humorous effect (See Figure 8). [24]

Another example is the definite article the,which is occasionally spelled teh for a variety of reasons: to add emphasis to the noun that it precedes, (example 33), to mock other users (example 34), or to add humor to the statement.

(33)	that is teh lame (Know Your Meme: s.v. teh)
(34)	OMG j00 ar3 teh PWNED lol!!111 (Urban Dictionary, s.v. teh)

The usage illustrated in Example (33), which could be paraphrased to “that is the lamest”, is similar to very or the best. One can think of it as being similar to the phenomenon in speech in which a speaker that wishes to add a superlative effect to a noun will pronounce the (normally pronounced as /ðə/) as /ði/. Example (34), by contrast, uses teh to satirize those who type too quickly in order to make a point and consequently make numerous typographical errors, to mock bad spellers, or even to mock users that use a large amount of Internet slang. Example (34) demonstrates excessive use of Internet slang, with the intent to mock someone.

Our data does not offer direct support the hypothesis that your would be an Internet meme. [26] The website Know Your Meme, which is meticulous in identifying and tracking current and past memes, does not contain a page for your. [27] And if the use of your was in fact a meme, we might expect to find it being used often in the 9GAG corpus, whose content often focuses on memes.

Nonstandard your could also be associated with another type of online act, namely trolling. [28] An Internet troll is “a CMC user who constructs the identity of sincerely wishing to be part of the group in question, including professing, or conveying pseudo-sincere intentions, but whose real intention(s) is/are to cause disruption and/or to trigger or exacerbate conflict for the purposes of their own amusement” (Hardaker 2010: 237). One commonly known tactic is to use misspelled words within troll posts. If a troll seeks attention, including an error is a surefire way of provoking another user into replying. The inclusion of errors by a troll can also work in their favor by feigning lack of intelligence. Many trolls will post an unpopular opinion that appears ignorant to other users as a way of luring them to reply. Trolls might also be aware of the strong emotional reactions that many users have toward misspellings, and especially towards the use of your.

Our data contains some examples where your clearly serves this purpose. Example (35) demonstrates one such instance.

(35)	(A)	You’re a faggot
	(B)	>your
	(C)	0/10 [29]
	(D)	Actually, he is correct seeing as how you’re is you are.
	(E)	Ha you actually are a faggot !!
	(F)	Go back to school. You’re - You are?
	(G)	you sir deserve a medal
	(H)	BUUUURRNNN
	(A)	As in, you are a faggot.

The first comment (A), You’re a faggot, is followed by a grammar correction that is incorrect (B) and the subsequent reaction from various users. Comment (B) is corrected by (D) and (F), chastised for inept trolling by (C), and applauded for good trolling by (G) and (H). This example clearly demonstrates how people use your for a specific function, in this case antagonizing the other users.

Example (35) is unlikely to be a genuine error, as trolling is indicated several times. However, cases like this are few in our data; the vast majority contain no overt indication that your would have been used intentionally for trolling. It therefore follows that in contrast to teh, which has a number of specific pragmatic functions, there is no evidence of your being commonly used to perform any one specific function. Even so, individual instances are often used for a variety of functions such as humor or sarcasm.

5.5 Adolescent language

Lastly, we consider the idea that your would simply be another manifestation of teenspeak, the language of adolescent users of these media-sharing websites, on a par with such forms as LOL, hehe and TTYL (cf. Tagliamonte and Denis 2008: 11). The Wet Seal T-shirt discussed in our Introduction may lend support to this idea, as it had been manufactured for sale to teenage girls. Indeed, comments to one of the YouTube videos included in the analysis had a higher your rate than the others. The video in question, Naughty Omegle Fun, contained 8 occurrences of nonstandard your as opposed to 4 standard forms. This particular vlog has a target audience of teen girls, which would offer some support for the idea that your would be linked to the commenters’ age.

(36)	your handsome! Love your videos (YouTube: Naughty Omegle Fun)

Having said that, if the use of nonstandard your was due to teenspeak, we would expect to see a higher rate of your use within the 9GAG corpus, which is mainly visited by teenagers and college students; YouTube has a much wider demographic. As we have previously observed, this hypothesis is clearly not borne out by our data. Unfortunately, our research design does not allow us to investigate this link systematically, as reliable demographic information about individual commenters is not available due to the anonymous nature of our data. Pursuing this idea is therefore left for further research.

To sum up the qualitative part, each factor that we have discussed likely accounts for some portion of your uses in YouTube comments, but the data do not indicate any one distinct factor that could be proposed as being the major influence behind the nonstandard your. Rather, it is clear that a combination of factors work together to produce a community that is especially prone to your use.

Genuine error, economy, and adolescent language all likely contribute to your use. However, it seems more likely that the high level of anonymity in combination with a user’s self-identification with a community in which nonstandard language use is acceptable and even expected could be the driving force behind your use within our study. These two factors may then be exploited to a lesser degree by certain users to perform specific pragmatic acts in which nonstandard your can be used as a tool to gain replies, attention, and emotional responses.

6. Conclusion

In this study we examined the frequency of the use of the possessive your in place of the contraction you’re in addition to a variety of other nonstandard variants within YouTube comments. The use of your was determined to be very frequent, and even accounted for more than half of the combined standard and nonstandard forms. When compared to comments from 9GAG, the frequency of use becomes even more distinctive. Within the 9GAG corpus, users showed a higher preference for the standard you’re, with the use of your being rare.

The discrepancy between the frequencies within the two corpora could have a number of causes. However, we believe that two factors account for the majority of uses: the prevalence of a YouTube community identity, wherein users feel free to use nonstandard language, and that users do not feel the need to make a distinction between the standard and nonstandard form in anonymous settings. The lack of your uses on 9GAG does suggest that it is either a YouTube-specific practice or a practice that is specific to sites that allow a high level of anonymity. This interpretation is further supported by the finding that your accounts for a much higher proportion of instances within the YouTube corpus than the other nonstandard variants. Within the 9GAG corpus, however, the use of your is comparable to the other nonstandard forms, suggesting that it is related to genuine error, economy, or a lax attitude towards using standard forms.

While the scope of the study provided sufficient data for quantitative analysis, future studies would do well to use a larger corpus to find confirmation for the observed trends. Such a corpus would also allow for more in-depth and reliable quantitative analysis. The scope of the present study was limited to two websites. By including a wider range of websites in the analysis, it would also be possible to explore the extent to which nonstandard your is determined by the characteristics of specific virtual communities and their norms of interaction. However, even a large corpus will be limited without demographic information about users, and given the anonymity of this type of CMC data, this is obviously a major challenge for corpus compilation. Future corpus-based research on your would therefore clearly benefit from other methodological approaches, in particular from ethnographic analysis of virtual communities.

Notes

[1] On language complaints and standard language in general, see Milroy and Milroy (1991). [Go back up]

[2] A viral video is a video clip, usually uploaded to YouTube, that has become popular through sharing online and has gained millions of views. [Go back up]

[3] For more information, see e.g. The Urban Dictionary (s.v. Grammar Nazi), which provides user-generated definitions of this concept (http://www.urbandictionary.com/define.php?term=Grammar%20Nazi). [Go back up]

[4] http://rationalwiki.org/wiki/Internet_law#Skitt.27s_Law. Skitt’s law is also commonly known online as Muphry’s Law and Hartman’s Law of Prescriptive Retaliation. [Go back up]

[5] It should be noted that the characteristics that are associated with spoken or written language are not definite or absolute, but should be viewed as generalizations that often vary (Biber 1988: 24). [Go back up]

[6] Of course, blog posts are often multimodal, too. [Go back up]

[7] http://www.urbandictionary.com, s. v. YouTube comments. [Go back up]

[8] http://www.forbes.com/sites/kashmirhill/2012/07/23/googles-secret-weapon-to-fix-youtubes-awful-comments-your-real-identity/ [Go back up]

[9] http://gizmodo.com/youtube-comments-will-soon-be-less-racist-homophobic-1377565582 [Go back up]

[10] All data used in this study were collected prior to the comment system changes. [Go back up]

[11] http://www.youtube.com [Go back up]

[12] http://www.youtube.com/yt/about/en-GB/ [Go back up]

[13] http://9gag.com [Go back up]

[14] For example, unless a commenter remembers to uncheck an option box when posting a comment on 9GAG, both their comment and a link to the image they have commented on will appear in their Facebook feed, which some users might find to be invasive. [Go back up]

[15] It is not possible to know precisely which profiles are using a fake name, but several factors on a profile such as friends, photos, and timeline history can help to identify a troll account. For example, a profile with only a few friends and all of which use a cartoon as a profile photo is very likely to be a troll account. [Go back up]

[16] The variation in number is due to some comments being very long. Rather than include a fragmented comment once 1,000 words were collected, the entire comment was included in the corpus. [Go back up]

[17] The searches included only the variants your and you’re, excluding any other types of error, such as typographical errors (yuor) or mistakes in spacing (yourawesome). We did not consider other possible respellings of these forms, e.g. ur and yr, attested in SMS text messaging, cf. Tagg 2009. [Go back up]

[18] The manual examination was independently performed by two researchers to control for errors. After examination, only the instances that both researchers agreed that pertain to the study were included. [Go back up]

[19] See Myslín and Gries (2010: 93–95) for another way of evaluating the ratios of spelling variant in online data. We decided to employ the the χ²-test because the indepence of data points is not a major concern in our data, given that comments are short and rarely contribute more than one instance of a variant. [Go back up]

[20] When conducting the search for standard and nonstandard uses of I, only I as an independent word was included in the results. Accordingly, instances in which a user typed i’m were not included in these counts. [Go back up]

[21] As before, the context of each instance was examined to only include instances where the search word is being used exactly as its standard counterpart. For example, searches for U would also return the word U-boat, and searches for 4 would return hits such as 2:45 and 4,000, which have to be excluded from the analysis. [Go back up]

[22] While the variable its/it’s could have been included in the apostrophe omission category, it was included in this category because each variant within the variable has a standard use. Note also that we only consider instances where its is used instead of it’s, but not the other way around (i.e. the use of it’s in a context where the correct form would have been its). We would expect the latter error to be very infrequent in online comments. At least in text messaging, which is a similar genre in terms of formality, Tagg (2009: 148) found that the form its is used exclusively as a contraction of it is. [Go back up]

[23] The only exception is the variable I/i; we currently have no explanation for this anomaly. [Go back up]

[24] See Wikipedia, s.v. Lolcat and Know Your Meme, s.v. Lolcat. [Go back up]

[25] http://en.wikipedia.org/wiki/File:Cat_crying_%28Lolcat%29.jpg [Go back up]

[26] For more on Internet memes, see Knobel & Lankshear (2006). [Go back up]

[27] http://www.knowyourmeme.com [Go back up]

[28] For more on trolling, see Donath (1999) and Herring et al. (2002). [Go back up]

[29] Refers to Troll Points, wherein an unsuccessful trolling attempt will receive 0 out 10 points. [Go back up]

References

Androutsopoulos, Jannis. 2011. “Language change and digital media: A review of conceptions and evidence”. Standard languages and language standards in a changing Europe, ed. by Nik Coupland & Tore Kristiansen. Oslo: Novus. http://jannisandroutsopoulos.files.wordpress.com/2011/11/language-change-and-digital-media-preprint.pdf

Baron, Naomi. 2008. Always On: Language in an Online and Mobile World. New York: Oxford University Press.

Biber, Douglas. 1988. Variations across speech and writing. Cambridge: Cambridge University Press.

Bushnell, Catherine, Nenagh Kemp & Frances Heritage Martin. 2011. “Text-messaging practices and links to general spelling skill: A study of Australian children”. Australian Journal of Educational & Developmental Psychology 11: 27–38.

Crystal, David. 2011. Internet Linguistics: A Student Guide. London: Routledge.

Danet, Brenda. 2001. Cyberpl@y: Communicating Online. Oxford: Berg.

De Jonge, Sarah & Nenagh Kemp. 2010. “Text-message abbreviations and language skills in high school and university students”. Journal of Research in Reading 35(1): 49–68. doi:10.1111/j.1467-9817.2010.01466.x

Donath, Judith. 1999. “Identity and deception in the virtual community”. Communities in Cyberspace, ed. by Marc Smith & Peter Kollock, 27–58. London: Routledge.

Drouin, Michelle & C. Davis. 2009. “R u txting? Is the use of text speak hurting your literacy?”. Journal of Literacy Research, 41(1): 46–67. doi:10.1080/10862960802695131

Gannes, Liz. 2012. “Meet 9GAG, the community comedy site that’s growing like crazy”. All Things Digital. http://allthingsd.com/20120412/meet-9gag-the-community-comedy-site-thats-growing-like-crazy/

Goldwert, Lindsay. 2011. “Wet Seal hawking grammatically incorrect ‘Your Single’ t-shirts; Error appears to be intentional”. NY Daily News. http://www.nydailynews.com/life-style/fashion/wet-seal-hawking-grammatically-incorrect-single-t-shirts-error-appears-intentional-article-1.150365

Hardaker, Claire. 2010. “Trolling in asynchronous computer-mediated communication: From user discussions to academic definitions”. Journal of Politeness Research: Language, Behavior, Culture 6(2): 215–242. doi:10.1515/JPLR.2010.011

Hayne, Stephen & Ronald Rice. 1997. “Attribution accuracy when using anonymity in group support systems”. International Journal of Human-Computer Studies. 47(3): 429–452. doi:10.1006/ijhc.1997.0134

Herring, Susan C., ed. 1996. Computer-Mediated Communication. Amsterdam: John Benjamins.

Herring, Susan C. 2001. “Computer-mediated Discourse“. The Handbook of Discourse Analysis, ed. by Deborah Schiffrin, Deborah Tannen & Heidi E. Hamilton, 612–633. Oxford: Blackwell Publishing.

Herring, Susan C. 2007. “A faceted classification scheme for computer-mediated discourse”. Language@Internet 4. http://www.languageatinternet.org/articles/2007/761/

Herring, Susan C. 2012. “Grammar and Electronic Communication”. Encyclopedia of Applied Linguistics, ed. by Carol Chapelle. Hoboken, NJ: Wiley-Blackwell. doi:10.1002/9781405198431.wbeal0466

Herring, Susan C., Kirk Job-Sluder, Rebecca Scheckler & Sasha Barab. 2002. “Searching for safety online: Managing ‘trolling’ in a feminist forum”. The Information Society 18: 371–384. doi:10.1080/01972240290108186

Herring, Susan C. & J. C. Paolillo. 2006. “Gender and genre variation in weblogs”. Journal of Sociolinguistics 10(4): 439–459. doi:10.1111/j.1467-9841.2006.00287.x

Hill, Kashmir. 2012. “Google”s secret weapon to fix YouTube’s awful comments: Your real identity”. Forbes.com. www.forbes.com/sites/kashmirhill/2012/07/23/googles-secret-weapon-to-fix-youtubes-awful-comments-your-real-identity/

Jones, Quentin. 1997. “Virtual-communities, virtual settlements & cyber-archeology: A theoretical outline”. Journal of Computer-Mediated Communication 3(3). doi:10.1111/j.1083-6101.1997.tb00075.x

Jucker, Andreas H. & Christa Dürscheid. 2012. “The linguistics of keyboard-to-screen communication. A new terminological framework”. Linguistik Online 56: 1–26. 10.5167/uzh-67310

Kemp, Nenagh. 2010. “Texting vs. txting: Reading and writing text messages, and links with other linguistic skills”. Writing Systems Research 2(1): 53–71. doi:10.1093/wsr/wsq002

Knobel, Michele & Colin Lankshear. 2006. “Online Memes, Affinities, and Cultural Production”. A New Literacies Sampler, ed. by Michele Knobel & Colin Lankshear, 199–228. New York: Peter Lang.

Kytölä, Samu. 2013. “Multilingual language use and metapragmatic reflexivity in Finnish internet football forums: A study in the sociolinguistics of globalization”. Jyväskylä studies in humanities. Jyväskylä: University of Jyväskylä. http://urn.fi/URN:ISBN:978-951-39-5132-0

Lange, Patricia. 2007. “Publicly private and privately public: Social networking on YouTube”. Journal of Computer-Mediated Communication 13(1): 361–380. doi:10.1111/j.1083-6101.2007.00400.x

Ling, Rich & Naomi Baron. 2007. “Text messaging and IM: Linguistic comparison of American college data”. Journal of Language and Social Psychology 26(3): 291–298. doi:10.1177/0261927X06303480

Mahdawi, Arwa. 2012. “‘Your in America’ sets grammar fascists against fascists”. The Guardian. http://www.theguardian.com/commentisfree/2012/nov/28/your-in-america-grammar-fascists

Milroy, James & Leslie Milroy. 1991. Authority in Language. London: Routledge.

Munroe, Randall. 2006. xkcd: YouTube. http://xkcd.com/202/.

Myslín, Mark & Stefan Th. Gries. 2010. “k dixez? A corpus study of Spanish Internet orthography”. Literary and Linguistic Computing 25(1): 85–104. 10.1093/llc/fqp037

Plester, Beverly, Clare Wood & Victoria Bell. 2008. “Txt msg n school literacy: Does texting and knowledge of text abbreviations adversely affect children’s literacy attainment?”. Literacy 42(3): 137–144. doi:10.1111/j.1741-4369.2008.00489.x

Plester, Beverly & Clare Wood. 2009. “Exploring relationships between traditional and new media literacies: British preteen texters at school”. Journal of Computer-Mediated Communication 14(4): 1108–1129. doi:10.1111/j.1083-6101.2009.01483.x

Rose, Brent. 2013. “YouTube Comments Will Soon Be Less Racist, Homophobic, and Confusing”. Gizmodo.com. http://gizmodo.com/youtube-comments-will-soon-be-less-racist-homophobic-1377565582/

Rosen, Larry, Jennifer Chang, Lynne Erwin, L. Mark Carrier & Nancy Cheever. 2010. “The Relationship Between “Textisms” and Formal and Informal Writing Among Young Adults”. Communication Research 37(3): 420–440. doi:10.1177/0093650210362465

Sairio, Anni. 2014. “‘No other reviews, no purchase, no wish list’: Book reviews and community norms on Amazon.com”. Texts and Discourses in New Media (Studies in Variation, Contacts and Change in English 15), ed. by Jukka Tyrkkö & Sirpa Leppänen. Helsinki: VARIENG. http://www.helsinki.fi/varieng/series/volumes/15/nm_fifteen/sairio/

Santini, Beatrice. 2007. “Characterizing genres of web pages: Genre hybridism and individualization”. Proceedings of the 40th Annual Hawaii International Conference on System Sciences. Washington: IEEE Computer Society. doi:10.1109/HICSS.2007.124

Shortis, Tim. 2009. “Revoicing txt: Spelling, vernacular orthography and ‘unregimented writing’”. The Texture of Internet: Netlinguistics in Progress, ed. by Santiago Posteguillo, María José Esteve & M. Lluïsa Gea Valor, 2–23. Cambridge: Cambridge Scholar Press.

Suler, John. 2004. “The Online Disinhibition Effect”. CyberPsychology & Behavior 7(3): 321–326. doi:10.1089/1094931041291295

Swales, John. 1990. Genre Analysis. Cambridge: CUP

Tagg, Caroline. 2009. A Corpus Linguistics Study of SMS Text Messaging. PhD dissertation, University of Birmingham.

Tagliamonte, Sali & Derek Denis. 2008. “LINGUISTIC RUIN? LOL! INSTANT MESSAGING AND TEEN LANGUAGE”. American Speech 83(1): 3–34. doi:10.1215/00031283-2008-001

Thurlow, Crispin. 2006. “From statistical panic to moral panic: The metadiscursive construction and popular exaggeration of new media language in the print media”. Journal of Computer-Mediated Communication 11(3): 667–701. doi:10.1111/j.1083-6101.2006.00031.x

Thurlow, Crispin & Alex Brown. 2003. “Generation txt? The sociolinguistics of young people’s text-messaging”. Discourse Analysis Online 1(1). http://extra.shu.ac.uk/daol/articles/v1/n1/a3/thurlow2002003-t.html

Varnhagen, Connie, G. Peggy McFall, Nicole Pugh, Lisa Routledge, Heather Sumida-McDonald & Trudy Kwong. 2009. “lol: new language and spelling in instant messaging”. Journal of Reading and Writing 23: 719–733. doi:10.1007/s11145-009-9181-y

Wattenhofer, Mirjam, Roger Wattenhofer & Zack Zhu. 2012. “The YouTube social network”. Proceedings of the Sixth International AAAI Conference on Weblogs and Social Media Online. http://www.aaai.org/ocs/index.php/ICWSM/ICWSM12/paper/download/4581%26lt%3B/5003

Wood, Andrew F. & Matthew J. Smith. 2005. Online Communication. Linking Technology, Identity & Culture (Second Edition). Mahwah, NJ: Lawrence Erlbaum Associates.

Yus Ramos, Francisco. 2011. Cyberpragmatics: Internet-mediated Communication in Context. Amsterdam: John Benjamins.

Web sources

9gag.com Statistics, Alexa.com http://www.alexa.com/siteinfo/9gag.com

About YouTube, Youtube.com. 2013. http://www.youtube.com/t/about_youtube

“Grammar Nazi”, Urbandictionary.com. http://www.urbandictionary.com/define.php?term=Grammar%20Nazi

“LOLcats”, Knowyourmeme.com. http://knowyourmeme.com/memes/lolcats

“Lolcat”, Wikipedia.org. http://en.wikipedia.org/wiki/Lolcat

“Skitt’s Law”, Rationalwiki.org. http://en.wikipedia.org/wiki/Lolcat

Statistics, Youtube.com. 2013. http://www.youtube.com/yt/press/statistics.html

“Teh”, Knowyourmeme.com. http://knowyourmeme.com/memes/teh

“Teh”, Urbandictionary.com. http://www.urbandictionary.com/define.php?term=teh

YouTube comments, Urbandictionary.com. http://www.urbandictionary.com/define.php?term=youtube+comments

Youtube.com Statistics, Alexa.com. 2013. http://www.alexa.com/siteinfo/youtube.com

Appendix 1. Materials used in the YouTube corpus

Appendix 2. Materials used in the 9GAG corpus