survey – Languaged Life

Long-Term Implications of Accent Representation in Children’s Media

Roni Grushkevich, Claire Lim, Kendall Vanderwouw, Daniel Zhou

Who is the most memorable villain you remember from your childhood era? We hypothesize that most individuals will remember a villain portrayed with a heavy accent. This is due to the phenomenon of othering and the idea that children will have a hard time connecting with a character that sounds different from them and the standard variety. We will use the childhood show, Phineas and Ferb, to see if this is true. Through the conduction of a survey, analyzing voice recordings in Praat, and doing sound analysis from an episode of Phineas and Ferb we will be able to see the phenomenon of othering. In Praat, we proved this phenomenon by showing that Dr. Doofenshmirtz, the antagonist, has a lower /æ/ F1 formant than Phineas and a native American English speaker. Additionally, analyzing the Hail Doofania episode, we were able to prove that Doofenshmirtz pronounced 6 sounds differently from a native American English speaker. All this proves the idea that villains are portrayed differently with negative attributes on children’s TV shows.

[expander_maker id=”1″ more=”Read more” less=”Read less”]

Introduction/Background

Our group proposes to do research on the effects of villainization in children’s shows, which has possible long-term implications on people’s perceptions of membership in their ingroup. This membership perception can be encapsulated by the phenomenon of othering. Othering is a process in which a more powerful and/or influential group reinforces differences and authority over another (Williams & Korn, 2017). Hence, our project aims to look at whether adoption of othering by children’s shows’ producers for villainous characters influences the audience’s perceptions of who is “good” (ingroup) and who is “bad” (outgroup).

Previous studies conducted by sociolinguists Calvin Gidney and Julie Dobrow (1998) indicated most foreign accented villains in children’s shows were exclusively European such as German, British, and Russian. These accents in particular are often characterized with high arrogance, intelligence, and socioeconomic status while low on friendliness, pleasantness, and honesty (Shah, 2019). These sentiments towards European accents may be reflective of World War II, Revolutionary War, and Cold War hostilities that dominated media/propaganda during the time.

We choose to look at the children’s show Phineas and Ferb, where the main antagonist, Dr. Doofenshmirtz, is portrayed with an exaggerated foreign accent. Heroes typically speak in the standard dialect and embody idealistic standards (Dobrow & Gidney, 1998), which holds true for the show’s protagonists, Phineas and Ferb. Through these differences in speech patterns, nonstandard accents become associated with “otherness” and antagonism (Constable, 2021).

Specifically, we will examine the phonetic production of vowels by accented English-speaking villains. When villains “mispronounce” English vowels, it is an obvious symbol of their communicative differences. That said, Dr. Doofenshmirtz is a great character to analyze as he does not speak the standard variety like most viewers. In addition, we plan to explore another linguistic variable that distinguishes Doofenshmirtz’s speech from the American standard.

Survey

To supplement the studies we will be citing, we also found it important to explore the existence of accented villains within the media we consumed as children and the types of accents they adopt. To do this, we constructed a preliminary survey prompting college-aged students (aged 18 to 24) to name “the most memorable villain from the movies/shows [they] watched as a child (age 2-10).” In the second iteration of the survey, participants were asked if they distinctly recall the villain expressing an accent coupled with any overall impressions they had of the villain. The survey was shared amongst UCLA students and other college friends, as long as all participants fell within the age range.

The results of the survey demonstrated a strong prevalence of foreign accents in the villains chosen, evidenced by Figure 1. Out of 38 total responses, 23 responses were accented villains. Not only was a clear statistical significance demonstrated of the trend, but actually a majority of the villain choices were associated with foreign accents. Given the breadth of types of respondents coming from California to Uganda to Singapore accompanied by the complete autonomy of selection of the villain, the survey results highlight the abundance of accented villains in popular media.

In addition, the values from Figure 2 present an interesting observation regarding the variety and uniqueness of villain selections. In total there were only 24 unique villains chosen. Of those 24 villains, only 10 were actually accented. The disparity between the number of accented villains and number of accented responses is explained by the incredibly high number of repeat selections of villains that are accented. For instance, Ursula (who adopts a condescending British accent) from The Little Mermaid was listed five times and Dr. Doofenshmirtz was repeated four times. In fact, of the 14 duplicate responses, 12 contained accented villains. This informs on the consolidation of popular children’s media; most children consume much of the same television shows and movies (Disney).

Another interesting note revolves around the traits that stand out or make the villain memorable. The comments on the impressions of the villain fell into two camps: “scary”/“intimidating” or “cool”/“respectable”. Not a single comment mentioned how the villain sounds or their voice. The “othering” of the villains through accented speech takes an implicit or passive effect on the viewer rather than being the focal point. The viewer then subconsciously categorizes the character in the out group. These observations and key takeaways from the survey helped propel the main research portion of the study and guided the choice of Dr. Heinz Doofenshmirtz as the character to analyze.

Methods

Our main goal in collecting data was to show that Doofenshmirtz follows German pronunciation patterns when speaking English. His nonstandard speech is meant to “other” him from the protagonists, determining his characterization as “bad.” Therefore, our data took a two-pronged approach: first, we used the spectrogram software Praat to analyze sound files of subjects speaking. This way, we could look at Doofenshmirtz’s articulatory properties and point to which features distinguish him as German-sounding. Then, we sent speakers of English to watch an episode of the show and comment on which non-standard features they heard him using. Between these two methods, we gained an understanding of both how Doofenshmirtz speaks and how English-speakers perceive his speech.

When researching in Praat, we looked at the vowels /i/ and /æ/. /i/ exists in both English and German, so we expected no statistical difference in the articulation of a German versus English speaker. On the other hand, the /æ/ sound exists in English but not German. In English, German-speakers tend to replace this sound with the higher vowel /ɛ/, which is easier to produce based on its presence in the German language.

We used samples from four subjects in Praat, two German-accented speakers and two “standard” American speakers. Our “baseline” German accent came from a 50-year-old native speaker who grew up in Düsseldorf but now lives in California. We were particularly interested to see how his speech compared to the German-accented Doofenshmirtz (who, intriguingly enough, was voiced by an American actor). We also took a sample from an 18-year-old native Californian, who we determined to have “standard” vowel production. Finally, we compared her speech to the pronunciation of the show’s protagonist, Phinneas (who was also voiced by an American actor, but was intended to have “standard” production). We wanted to ensure that our samples were comparable, so our first three subjects all produced Doofenshmirtz’s catchphrase: “Curse you, Perry the Platypus!” (/i/ is found in the word /ˈperi /, while /æ/ is prominent in /ˈplætəpəs/). Although no samples existed of Phinneas saying this sentence, we located a similar sentence where he still said both words in proximity.

Our other group of data came from observing Doofenshmirtz’s speech throughout an episode. Inherently, the sounds of the German language differ from those of English. Some sounds feel tricky or unnatural to non-native speakers. Speakers often replace an English sound with a similar foreign counterpart they know how to produce, just as how a German speaker might replace /æ/ with the /ɛ/ sound of their native language. Often, it is this sound substitution that creates the phenomenon of an accent. For our project, we wanted to see how American individuals perceived Doofenshmirtz’s so-called accent. We asked two college-aged subjects to watch Season 1, Episode 26 – “Out of Toon / Hail Doofania” and explain how his speech compares to the substitutions we’d expect of a German-speaker.

Results/Analysis

Praat

In Praat, we used the “View & Edit” feature to analyze something called formants (lines on a spectrogram that indicate a concentration of energy). For this project, we were mainly concerned with the first formant (F1), which shows vowel height. We found that all of our subjects had similar F1 formants for the vowel /i/. Since this sound exists in both languages, non-native speakers made no substitutions when speaking English. Our Hz values ranged from 376.7 to 460, meaning there was less than a 100-Hz difference between how the two most differently accented speakers produced the vowel. Therefore, the data showed no trends in production variations for Germans and Americans in /i/.

On the other hand, the data showed interesting patterns when it came to /æ/. We predicted that German speakers would have a lower F1 value than English-speakers (due to the fact that /ɛ/ has a lower F1 than /æ/). Overall, we found this to be true. Our authentic German subject had the lowest F1 of the group at 714.1. Likewise, our highest formant value came from “standard” speaker Phineas at 1239. Interestingly, the dubbed voices had a much higher average F1 formant, which we hypothesized had something to do with the nature of the audio files. If we separate our samples by this type of audio file (real speakers vs dubbed speakers), we find that German speakers’ formants are 100-150 Hz lower than those of American speakers. This indicates that they do, in fact, produce ash more similar to /ɛ/.

Sound Analysis

We paid attention to foreign-sounding allophones, compared to a Californian English speaker’s pronunciation, in Hail Doofania. We only analyzed Dr. Doofenshmirtz’s speech and ignored any pronunciation deviations in his song.

In his speech, we identified 6 sounds that were pronounced differently from how a Californian English speaker would pronounce the sounds. The most frequently occurring deviations were his pronunciation of the alveolar approximant ([ɹ]) and voiced postalveolar affricate ([dʒ]). 23 of the 34 deviations were [ɹ] and 6 deviations were [dʒ]. We identified pronunciation deviations just from hearing and our personal judgments given the time and resource constraints of this project.

We looked at the German International Phonetic Alphabet (IPA) sounds to see if the allophones used by Dr. Doofenshmirtz were from the German sound repository. We proposed that his mispronunciation of English allophones was because he was borrowing German sounds to pronounce English sounds. But, the sounds that Dr. Doofenshmirtz used were actually not found in the German IPA

As the allophones used by German speakers for the different analyzed sounds were not what Dr. Doofenshmirtz used, we came up with 2 possible explanations to account for Dr. Doofenshmirtz’s different pronunciations.

First, we suggest that the producers might have wanted to distance Dr. Doofenshmirtz from a German identity. Perhaps they wanted Dr. Doofenshmirtz to have some accent aspects which alluded to a German identity but not have the character be too strongly associated with a German identity. Podesva (2011) posits that the creation of an identity or a “persona” utilizes several linguistic features, and different components of a linguistic feature are adopted depending on the context the speaker is in. This suggests that a bricolage of linguistic features are used to construct an identity. For Dr. Doofenshmirtz, this means that his identity as an individual from Drusselstein (a fictional East European country) was not solely reliant on his accent but on other features too. Hence, it is possible that the show’s producers did not use the German allophones to pronounce the different Californian English sounds to index Dr. Doofenshmirtz’s heritage, but utilized other linguistic and non-linguistic features to do so.

Second, it could be because Dr. Doofenshmirtz is voiced by Dan Povenmire, a native Californian English speaker. Thus, Dr. Doofenshmirtz’s mispronunciation could stem from how a Californian English speaker perceives a German accent to sound like and not how an actual German accent sounds like.

Interestingly, we found that our German speaker pronounced the sounds we studied, in the same way that Californian English speakers do. We attribute this to his exposure to the Californian English accent and was able to distinguish the different allophones and use them according to the context he was in.

Discussion and Conclusions

We looked at the amount of time Dr. Doofenshmirtz spoke in Hail Doofania to gauge the significance of his accent’s contribution to his identity. Dr. Doofenshmirtz had 112 seconds of speech time in the 9 minute and 11 second episode (551 seconds). This equates to about 20% of the episode. As such, Doofenshmirtz has only 20% of the episode to leave an impression on the audience, making it crucial for the producers to adopt several indices to create Doofenshmirtz’s character as an “evil scientist”. The identified indices were: his accent, his clothing (lab coat), the content of his speech (his plans to take over the Tri-state Area) and actions (fighting with Perry the Platypus).

We felt that some younger audience members might be too young to comprehend the sinister nature of Dr. Doofenshmirtz’s plans and might instead rely on auditory and visual cues (his accent and attire & fight scenes, respectively) to index Dr. Doofenshmirtz as a villain. By extension, we think that it is possible for these auditory and visual cues to lay the groundwork for associations that these child audiences make with someone who is “bad” and in an outgroup.

Given that it is more likely for children to come across speakers of different accents than people dressed in lab coats in their everyday lives, we believe that it is possible for them to consequently (because of cartoon villains having accents) associate people with accents as members of an outgroup. This demonstrates how accented villains are part of the phenomenon of othering and can influence children’s perception of “good” and “bad”.

We believe that the results we obtained through Praat were very useful. They supported our findings and proved that accented speakers sound different from the standard. We wish we had more time to obtain and analyze more data. On the other hand, the survey results were not very helpful. We were not able to come to any conclusions through the survey results, but we definitely feel that the survey pushed us in the right direction.

One of the most important takeaways that we believe show producers should know is that they must be mindful of the stereotypes they create or reinforce in their audience. These shows have long term implications that heavily influence individuals.

References

Constable, E. (2021, March 17). Dear Disney, stop teaching kids that foreign accents are evil. LHS Epic. https://lhsepic.com/9602/opinion/dear-disney-stop-teaching-kids-that-foreign-accents-are-evil/

Dobrow, J. R., & Gidney, C. L. (1998). The Good, the Bad, and the Foreign: The Use of Dialect in Children’s Animated Television. The Annals of the American Academy of Political and Social Science, 557, 105-119. JSTOR. https://www.jstor.org/stable/1049446

Flege, J. E., Bohn, O.-S., & Jang, S. (1997). Effects of experience on non-native speakers’ production and perception of English vowels. Journal of Phonetics, 24(4), 437-470. ScienceDirect. https://doi.org/10.1006/jpho.1997.0052

Paquette-Smith, M., Buckler, H., White, K. S., Choi, J., & Johnson, E. K. (2019). The Effect of Accent Exposure on Children’s Sociolinguistic Evaluation of Peers. Developmental Psychology, 5(4), 809-822. American Psychological Association. https://psycnet.apa.org/doi/10.1037/dev0000659

Phineas and Ferb Wiki. (n.d.). Drusselstein | Phineas and Ferb Wiki | Fandom. Phineas and Ferb Wiki. https://phineasandferb.fandom.com/wiki/Drusselstein

Shah, A. P. (2019). Why are Certain Accents Judged the Way they are? Decoding Qualitative Patterns of Accent Bias. Advances in Language and Literary Studies, 10(3), 128-139. Australian International Academic Centre PTY.LTD. https://doi.org/10.7575/aiac.alls.v.10n.3p.128

Sharma, M. (2016). CHAPTER SEVEN: Disney and the Ethnic Other: A Semiotic Analysis of American Identity. In Teaching with Disney (pp. 95-107). Peter Lang AG. Williams, M. G., & Korn, J. (2017). Othering and Fear: Cultural Values and Hiro’s Race in Thomas & Friends’ Hero of the Rails. Journal of Communication Inquiry, 41(1), 22-41. SAGE. 10.1177/0196859916656836

[/expander_maker]

How LOL got between X and Z

Michelle Johnson, Kayla Sasser, Lucy (Chenyi) Wang, Grace Shoemaker, and Lien Joy Campbell

Figure 1. An example conversation between Gen X and Gen Z showing possible generational gap in the usage of humor markers – emojis in this case.

Even though the sad emojis in that exchange were used in a sad context, many people might laugh or find that inappropriate. Whether you are one of those people or someone likely to use emojis just like “Mom”, read on. As texting has grown to be a more popular form of regular communication, it may seem as if connecting with people has only become easier – but with ubiquity comes complexity. And if you are not among those at the vanguard of these complexities (the youth), you could be missing out. This brings us to the question: does expressing humor over text vary by generation? In this study we focused on Generation X and Generation Z’s use of emojis, emoticons, and other ways they chose to convey humor and tone in texts. In focusing on humor we were able to analyze the frequency of humor makers and their meanings in context. Based on our data, we found that there were definite differences in how the generations use and react to text language. Keep reading to learn what these key differences were and how we studied them (and maybe how to finally make that teenager in your life laugh).

[expander_maker id=”1″ more=”Read more” less=”Read less”]

Intro and Background

Generation Z (those born between 1997 and 2012) grew up and learned how to communicate post-advent of the invention of instant messaging. Their texting style and speaking styles are intertwined and take inspiration from each other. On the other hand, those of Generation X (born between 1965 and 1980) had to transfer previously-established styles of humor and communication to the new technological medium (Downs, 2019). This accounts for the disconnect between considering texting to be a form of writing (like an email or letter) and considering it simply as talking put onto a screen. The term “written speech,” coined by John McWhorter, gives a name to the adaptation of texting to account for all the complexities of face-to-face communication that change how the content of a message is received: emotion, formality, humor, tone, body language and facial expression. Across platforms from iMessage to TikTok, young texters use unspoken and quickly changing combinations of punctuation, capitalization, and symbols to directly translate trends and slang into the digital world.

Methods

Following our belief that Gen X and Gen Z would communicate humor over text in significantly different ways and a path laid out by a study conducted by Sánchez-Moya and Cruz-Moya (2015), we chose to create a survey that focused on responder’s opinions on texting and their texting habits. We specifically targeted people’s habits in the use of humor markers like emojis and typed laughter by asking them to choose the most appropriate option to represent a feeling or as a response to various tonal and emotional contexts. Once we had our responses, we organized our data in terms of type of marker (emoji/emoticon/capitalization) and focused on whether the marker was used literally or creatively in relation to each generation. We expected a wider range of responses in Gen Z and more similar, literal responses from Gen X.

Results and Analysis

We began our analysis by categorizing the responses we received in terms of whether they were literal or not, and we additionally compared the use of emoticons and capitalization. Further, we then also analyzed the frequency of answers we received for each question. We will present examples of each of these analyses and the contexts in which they were applied.

Beginning with our analysis of literal vs. non-literal use of markers, figure 2 presents a strong difference in Gen X and Gen Z’s preferences for literal and non-literal emojis.

The above graph illustrates a strong example of a general trend we found in our data: that overall, Gen X preferred to utilize emoji and other humor makers literally. In comparison, Gen Z showed a preference for less literal uses. Also, specifically for this question, within the categories of literal and non-literal, Gen X preferred a laughing emoji (😂) to show that they were laughing in 55% of their responses whereas Gen Z preferred a crying emoji (😭)–the exact opposite–to show that they were laughing in 42% of their responses. This was an even stronger non-literal response than expected suggesting a much higher degree of irony in Gen Z’s texting than in Gen X’s.

Moreover Figure 2.1 presents another strong case for Gen X’s preference for literal marker use and Gen Z’s preference for non-literal outside of just emojis. Figure 2.2 presents the response options as well as their categorization as either literal or non-literal.

Not only was Gen X’s preference for literal answers and Gen Z’s preference for non-literal answers illustrated in their selection of emojis but also in their preference for other answer types too. In the above example we took the unmarked and expected literal responses to the presented situation to be congratulatory, positive, and generally aligned with the topic of the context, whereas the non-literal responses demonstrate an indirect type of response by focusing on a non-topicalized part of the context (i.e. the bathroom). Again, we observed a strong preference from Gen X for a literal or positive response and a strong preference from Gen Z for a non-literal or indirect response.

In addition to studying the differences in how humor markers were used to convey literal and non-literal meaning we also wanted to provide insight into the different variations in the types of markers commonly used. Generally, we expected to see a more diverse use of these markers and variations, not just emojis, in Gen Z’s texting, leading us to figure 3. Figure 3 illustrates the overall differences in emoji (😂,😩) and emoticon ( :(, 🙂 ) use according to generation.

We chose to study emoji vs. emoticon use specifically as we believed that there would be a strong difference between the generations. However, both Gen X and Gen Z tended to prefer emojis. Unexpectedly, Gen X overwhelmingly preferred to use emojis over emoticons. We had thought that due to their longer history and generally less ambiguous and more established static meaning that Gen X would favor emoticons (Bai et al., 2019). This was not the case. Interestingly too, Gen Z actually tended to use more emoticons than Gen X. This result however supports our belief that Gen Z would demonstrate a broader range of humor marker use, splitting their results more evenly between emoji and emoticon. This could also demonstrate that Gen Z is exhibiting more creativity or nuanced flexibility in how they use these markers and what they take them to mean.

Next, we chose to study another variational marker, capitalization (OR SHOUTING). We chose to study capitalization in addition to emoji/emoticon differences as it is a unique action in texting that specifically denotes tone (McCulloch, 2019). Figure 4 below illustrates our comparison of capitalization use according to generation.

As shown above, Gen Z favored the use of capitalization while Gen X preferred messages that mixed capitalization and lowercase. This illustrates a stronger preference in Gen Z for using messages that convey a stronger or louder tone and demonstrates McWhorter’s idea of “written speech” in the younger generations (2017). Interestingly too, the younger generations’ relatively strong preference for “shouting” over text could indicate a recent change in what all-caps texting “means” and illustrate a higher level of comfort with the nuance of tone that all capitalized text creates. In comparison, Gen X may still interpret it as simply yelling at someone and therefore use it more sparingly. However, to corroborate those claims more testing would need to be conducted.

Finally, we got even more specific with our analyses–we categorized and analyzed the answers to each question on the survey to measure the frequency of each response per question for each generation. We did this to examine the specific texting behavior of the generations on a smaller, context-dependent scale. Figure 5 is a particularly interesting example that demonstrates the general trend in the generational behavior we observed.

We discovered, as in the above example, that Gen Z’s responses were more evenly spread out across the response options creating a much more dispersed answer graph (seen in red). Meanwhile, Gen X tended to answer more similarly to one another, strongly favoring one answer, ‘terrible 😔,’ as can be seen by the single tall blue bar. This pattern was relatively consistent across all of our data and was in line not only with our prediction that Gen Z would show a wider range of responses, but also with our prediction that Gen X would tend to use more literal responses. Moreover, another point to note in figure 5 particularly is that Gen X answered ‘terrible 😭’ 20.43% of the time, which in this context was interpreted as a literal use of a negative emoji. However, given our results in figure 1, this could also denote a more sarcastic or ironic tone. Such an analysis could also be in line with the other trend illustrated above as Gen Z showed a greater preference for answers that denote a less literal more ironic tone (terrible 🙃, terrible 😁).

Overall, our data demonstrated that Gen X tended to use emojis more literally and more consistently while Gen Z preferred to use them less literally and showed a wider range in their use. In terms of emoticons, both Gen X and Gen Z preferred emojis with Gen X showing a much stronger preference, and in cases of capitalization, Gen Z used messages in all caps much more than Gen X.

With all that said, we would like to address some possible confounds that could affect our data and analyses. Firstly, there was quite a disparity in the number of responses we received from each generation, heavily skewing toward Gen Z. This may have been because this survey was distributed by us (members of Gen Z), which also brings to light another possible issue: our own generational biases in both the analysis of the data and the creation of the survey. Additionally, this study’s construction as a multiple-choice survey poses the possibility that the choice of answers may have directed people’s responses. Moreover, we did not notice a significant effect of gender in our study. However, it could be a very interesting avenue to pursue in future research.

Discussion and Conclusion

As seen in our data analysis, we found that there is in fact a gap in the usage of humor markers between the two generations, which supported our initial predictions. More specifically, based on the overwhelming choice by Gen X to use literal meanings, it could be suggested that they tend to use them (especially emojis) at a surface level. Meanwhile, Gen Z’s varied usage of all four markers looks to be a bit more nuanced. Their choices reflect that they use ironic and non-literal meanings frequently in humorous contexts. The variation in their responses also suggests that each marker of humor could have its own unique function or meaning depending on the context. Such variety among Gen Z could be the result of their community of practice, which frequently takes part in internet culture and has therefore been able to develop their own unique understandings of humorous texts. It also reinforces McWhorter’s earlier suggestion that texting can involve more than words–it conveys natural human conversational gestures as well. Overall, it does therefore seem fair to say that there is more at play in Gen Z’s usage.

After conducting our study, we identified limitations in our methods that leave room for improvement. As mentioned earlier, the survey was created entirely by members of Gen Z. This could prove to be problematic because the response options are potentially more biased toward a typical Gen Z response and not adequately represent typical Gen X responses. Including the input of Gen X members could have created a more balanced selection of responses. Another less obvious limitation of our study is that we did not account for phone differences. We realize that the appearance of Android and iOS emojis differ and that this difference may procure different emotional responses and therefore be used in a different context than our survey initially accounted for.

So, what are the next steps? Our findings and conclusions tell us that there is definitely room for further research on intergenerational communication over text. Improving upon this study’s weaknesses and widening its scale could provide more insights into the big differences that lie between the text-language of Gen Z and Gen X. Some new topics of interest include: different attitudes towards the appearances of emojis (i.e. Android vs. iOS), the evolution and idiosyncrasies of Gen Z’s online language, and analyses of textual gaps that may occur on the basis of factors other than generation.

References:

Bai, Q., Dan, Q., Mu, Z., & Yang, M. (2019). A Systematic Review of Emoji: Current Research and Future Perspectives. Frontiers in psychology, 10, 2221. https://doi.org/10.3389/fpsyg.2019.02221

Downs, H. (2019). Bridging the Gap: How the Generations Communicate. Concordia Journal of Communication Research, 6. https://doi.org/10.54416/SEZY7453

McCulloch, G. (2019). Because Internet. Penguin Adult HC/TR & Riverhead Books.

McWhorter, J. H. (2017). Words on the move: Why English won’t- and can’t- sit still (like, literally). Picador, Henry Holt and Company.

Sánchez-Moya, A. & Cruz-Moya, O. (2015). Whatsapp, Textese, and Moral Panics: Discourse Features and Habits Across Two Generations. Procedia – Social and Behavioral Sciences, 173, pp. 300-306. https://doi.org/10.1016/j.sbspro.2015.02.069

[/expander_maker]