Hyperrhiz 26

The 2020 Presidential Debate But Only the Parts Where Someone Breathes Loudly or Sighs

Zach Whalen
University of Mary Washington

Citation: Whalen, Zach. “The 2020 Presidential Debate But Only the Parts Where Someone Breathes Loudly or Sighs.” Hyperrhiz: New Media Cultures, no. 26, 2023. doi:10.20415/hyp/026.a02

Abstract: This essay is both an artist statement and a case study focusing on the surprising viewership of a video I created in 2020 unexpectedly became popular in April 2021. In the course of this reflection and analysis, I show how the meaning of this video involves the algorithms driving the YouTube platform, and the various and painful contexts of the year 2020. Although I view the video as a successful piece of art in its own way, I also argue that, as a meme, it must also be viewed with reference to its three salient contexts: the deformative meme ecology, YouTube as a platform, and the year 2020. Because the memetic logics demonstrated by my video bring together the participatory logics of appropriation with the capricious attention of the YouTube algorithm, my access to the videos’ audience analytics affords a unique perspective on the memetics of resonance and spread.  Ultimately, a close analysis of this video will shed light on the ecology of video memes on YouTube, the recommendation algorithm as an audience, and the way that those two structures come into view through the analytics data to offer insight on viewers’ responses that not typically available to artists. Ultimately, I conclude that the qualities of media encouraged by the analytics dashboard’s emphasis on engagement are not necessarily consistent with the goals of critical media art, but it is possible to consider how further attention to metrics could create compelling art with attention to algorithmic affordances.

Keywords: YouTube, video memes, deformance, analytics, algorithms, 2020, post-memes.

“This is the hard hitting debate that the government doesn't want you to see.”
– X Caines (7 months ago)

“This makes me feel like I can’t breathe.”
– echilcote423 (3 weeks ago (edited))

These two quotes represent comments (2,700+ as of Summer 2022) on a video I created in October 2020 that has so far been viewed around 300,000 views on YouTube. The video lasts for one minute and thirty-five seconds, and other than a very brief title, it consists of exactly what its name describes: a supercut of breathing and sighing from mostly Donald Trump and Joe Biden during one of their debates leading up to the 2020 presidential election. I created the edited video with a Python program called “videogrep” (Lavigne), and uploaded it to the YouTube channel that I otherwise use mostly for uploading lectures and tutorials for my classes. Here it languished in typical obscurity until some time in early April 2021 when – for reasons I don’t fully understand – YouTube began recommending it to certain viewers which caused the view count to tick up rapidly.

This essay is both an artist statement and a case study as I try to disentangle my ambivalence about the “success” of this video in the contexts of video memes, the YouTube platform, and everything that comes to mind about the year 2020. That is, even though this video is my most popular by far, the reasons people respond to it – as evidenced in the video’s analytics and as attested in comments – are sometimes inscrutable and, just occasionally, deeply offensive to me. I view the video as a successful piece of art that speaks on its own terms, but because this video is, as I argue, memetic content, it must be viewed with reference to its three salient contexts: the deformative meme ecology, YouTube as a platform, and the year 2020. As Ryan Milner argues, memes are a form of participatory media where – through their underlying logics of multimodality, reappropriation, resonance, collectivism, and spread – “[m]emetic media are the result of situated political and technological contexts. Drawing from these contexts, they depend on the social processes at the heart of social texts” (Milner 40–41). Therefore, because the memetic logics demonstrated by my video unintentionally conjoined the participatory logics of appropriation with the capricious attention of the YouTube algorithm, my access to the videos’ engagement analytics affords a unique perspective on the memetics of resonance and spread.  Ultimately, a close analysis of this video will shed light on the ecology of video memes on YouTube, the recommendation algorithm as an audience, and the way that those two structures come together in analytics to generate an insight on audience response not typically available in media art.

The Video Itself

The video is one minute and thirty seconds long. For 4 seconds, the fist part of the title appears; white text on a background: “The 2020 Presidential Debate.” That text fades to be replaced by the rest of the title, “But Only The Parts Where Someone Breathes Loudly or Sighs.” The black of that title fades up into the first of several very short clips (most are less than a second long) presented in very rapid succession. As the title promises, these clips are the breathing and sighing sounds emitted by Donald Trump, Joe Biden, and Chris Wallace. Some off-screen commentators also join as the video closes on another fade to black.

The visual element of each clip is exactly as it appeared in the PBS footage that will be familiar to anyone who watched the original debate on September 29, 2020. The shots are mostly straight on each candidate, hold each in side-by-side frames, or occasionally cut to the moderator.

As many commentators attest, watching the video can be a stressful experience. Since most of the loudest breathing sounds captured and collated into this supercut are the rapid inhalation just before a statement, the video is mostly a long series of sharp inhalations that (mostly Trump’s) increase in intensity as the video progresses. The final seconds of the video relieve that tension with the longer inhalations (or sighs?) from off-screen (most likely from PBS NewsHour host Judy Woodruff) as the camera shot pulls back on the stage set and a mostly empty auditorium.

Figure 1: A screenshot of a YouTube video with the title, “The 2020 Presidential Debate But Only the Parts Where Someone Breathes Loudly or Sighs.”
Figure 1. A screenshot of a YouTube video with the title, “The 2020 Presidential Debate But Only the Parts Where Someone Breathes Loudly or Sighs,” available at https://www.youtube.com/watch?v=bBNV32aeVfE

Context 1: Deformative Memes

300K views is certainly a large number of views, but the view count alone may not warrant a designation of “viral.” Instead, Limor Shifman’s influential definition of a meme provides a useful framework that clarifies how my video participates in memetic expression. For Shifman, a meme is

(a) a group of digital items sharing common characteristics of content, form, and/or stance, which (b) were created with awareness of each other, and (c) were circulated, imitated, and/or transformed via the Internet by many users. (41)

Certainly, the video I have created exists within a network of other videos with similar content (political speech) and stance, but the most important characteristic that unites it to a thread of practice is its form. More succinctly, Shifman draws on Deborah Tannen’s discussion of linguistic prepatterning to observe that an essential reason reappropriation works in memetics is the interplay between fixity and novelty, weaving “established tenets, characters, and themes into new iterations” to give “audiences a recognizable premise by which to read a text” (Milner 87). Unlike image macro memes, however, the recognizable form which viewers may be responding to when appreciating “2020 Debate But” is not content (for example, an image of “sad Batman” or “bad luck Brian). Instead, the premise is form.

The evolving and multi-layered discourse of form in this kind of memetics makes nomenclature difficult, much less etiology, but I believe this video is an example of a coherent and recognizable memetic trope – even though it lacks a consensus name.

The earliest example of this trope documented on KnowYourMeme.com is a 2016 YouTube video (currently blocked on the grounds of its copyright violation), with the title “bee movie but it keeps getting faster” (Bee Movie but It Keeps Getting Faster). Other variations on this idea subject The Bee Movie to different arbitrary constraints: “The Bee Movie But Without Bees” (The Bee Movie But Without Bees), “‘The bee movie’ but only when they say ‘E’” (“The Bee Movie” but Only When They Say “E”) and “The Bee Movie at 3000% speed except when they say ‘bee’” (The Bee Movie at 3000% Speed except When They Say “Bee”). Content creators who work with this theme also tend to make similar permutations in combination with the movie Shrek, “All Star” by Smash Mouth, the “steamed hams” scene from The Simpsons season 7, episode 21, and undoubtedly many others. For example, when this meme structure was emerging in 2016, I created two videos based on Rebecca Black’s “Friday,” one where it speeds up every time someone says “Friday,” and another where it slows down. I consider the latter to be more successful (“Friday”).

Although all these videos work with different content and apply different processes to that content, they have several things in common: more of a “deformation” rather than “remix” because the latter term implies recontextualizing memetic expressions and their fungible content. A “deformation” instead recontextualizes the framework of the medial containers of that content.

Videos that demonstrate this trope of arbitrary deformation include:

  1. an arbitrary and (in most cases) programmatically implemented deformation of some media,
  2. source content with a banality and familiarity (perhaps through other memes) that makes it exploitable, and
  3. a title that flatly attests the deformation that the video delivers

The formal operations of these meme videos offer a predictable pattern, but that pattern tends to resist naming owing to its flexibility and its content-specific literalness. On the website KnowYourMeme.com, two candidate pages exists for naming this meme, user submissions that have not been fully endorsed by the site administrators. “Bee Movie But” points to the origins of this meme in those first deformations of Bee Movie, but the term may remain obscure to anyone not familiar to the meme’s origin with that film (Wesley et al.). “X But Without Y” is more generic, but it fails to account for all the deformations like those that arbitrarily insert other content, change the speed, add bass effects, and so on (“X but without Y”). A more inclusive but verbose term appears in a subreddit created around the same time, “r/XButEveryTimeTheyYItZ.” That subreddit is set to private now, and thus invisible, but its description is simply a link to “Bee movie trailer but every time they say bee it gets slower” (Avoid at All Costs). Failing to find a descriptive name, one recourse could be to designations of the rhetorical functions where these videos operate – procedural, operationalist, conceptual, carnivalesque – but each of these carries its own bag of unnecessary conflations.

In a 2017 article for The Verge, Lizzie Plaugic organizes these and similar videos into subtypes under the broader term, “technical meme.” For Plaugic, these are “jokes that require work” because part of the enjoyment of the joke is the recognition of the effort that went into producing it, and the viewer’s idea of that work may stand in place of any sense of enjoyment from the resulting artifact.

A technical meme gets its charm not from aesthetic pleasure, but a workmanlike commitment to an arbitrary premise. These songs or video clips don’t sound good, let alone make sense, but they flaunt a perception of technical investment.... A technical meme deconstructs its human-made source material with the cool distance of a molecular gastronomist. (Plaugic)

Plaugic finds earlier antecedents to the technical meme in artifacts like the ethereal “Justin Bieber’s “U Smile” Slowed Down 800%” (mesiuepiescha) from 2010, chipmunkson16speed, or the similarly-improved, slower version of the Olsen twins’ “Gimme Pizza” (Mary Kate and Ashley Olson - Gimme Pizza Song (Slowed Down)).

Earlier precedents may also be relevant. Looking beyond the YouTube ecosystem, antecedents of this trope in video art include several pieces by Nam June Paik (“Video Tape Study No. 3,” “Beatles Electroniques”), 24 Hour Psycho by Douglass Gordon (procedurally very similar to the slowed down Justin Bieber song), or Jack Goldstein’s Metro-Goldwyn-Mayer that recuts the MGM roaring lion into a 2-minute long loop.

Cory Arcangel’s Super Mario Cloudz also fits the “X but without Y” pattern (“Super Mario Bros. but without everything but the sky”), and Patrick Lemieux and Stephanie Boluk’s “99 Exercises in Play” is a richer demonstration of this pattern’s expressive possibilities because their games preserve through many permutations the playability of the original game rather than, as Arcangel does, transposing it into a different modality of viewing.

The arbitrary nature of these deformations alongside the implied effort required to produce the deformed artifact also calls to mind conceptual art and writing. Sol Lewitt’s recipe-like instructions are similar to the titles of the video memes. In the same way that Lewitt’s “Straight Lines, Approximately One Inch Long, Drawn at Random, Within a Square Using 4 Directions of Line each a Different Color” is exactly what it says it is, so “Shrek but only the adjectives” (Shrek but Only the Adjectives) is a supercut of Shrek with only the spoken adjectives. For both of these works, it is possible to engage with at least three parameters: 1) the field of play circumscribed by the terms of its title, 2) the process of executing that play, 3) the individual variations, nuances, and specificity of the artifact created in the process that still manage to delight and surprise viewers. Crucially, the second of those stages – process – connects an audience to the subjectivities of expression, and it does so in spite of conceptualism’s appeal to displaced subjectivity in expression. For technical memes, the “workmanlike commitment” Plaugic refers to envisions the person behind the process and holds them up against the video’s “aesthetic pleasure.” It is not so much that the process (implied or actual) is difficult so much as it is unnecessary. Lewitt’s compositions clearly work with a similar contingency and tension between idea and execution, though with an opposite effect, generally. Lewitt’s instructions are textually banal, but spectacular in execution. Deformative video titles are hilariously arbitrary, but the videos are only amusing for a few minutes at most. In the case of my “Debate But” video, the process I used to create the video is, similarly, only loosely connected to the premise in the title, because someone else might have accomplished the same thing through different means. What matters to an audience is that relationship between the title and the execution.

In all of these examples, the politics of appropriation and remix are important parameters that Limor Shifman’s idea of “stance” helps clarify. Shifman breaks down a meme’s stance into three “subdimensions”: participation structures (after Susan Phillips), keying (after Erving Goffman), and communicative functions (after Roman Jakobson) (40–41). With these in mind, it is possible to ask “what is the meaning of a technical meme, and who is that meaning for?”

Broadly, technical memes are closely related to “shitposting,” originally a term for low-quality, high-volume posts on internet messageboards, but more recently a term of art which Tom Whyman defines on behalf of the “Simpsons Shitposting” Facebook Group. Interpreting the stated rules of that group, Whyman explains that

Here, then, the shitpost is defined not simply by its low quality, nor by how annoying it is for others. It is defined rather by a sort of indifference, both to quality … and to reception—either positive or negative ... This invites users to produce a high volume of memes based on whatever happens to come into their head. (215)

The shitpost can also be an episteme with problematic applications in political discourse, as I will discuss below, but its rejection of evaluative criticism in favor of the fact of its existence again echoes conceptual art’s orientation toward ideas and processes instead of artifacts. As Lewitt confesses, “What the work of art looks like isn’t too important” (Lewitt). Whyman even includes technical memes – without that name – among the examples of Simpsons shitposts that make connections with other memes. That is, the shitpost example “Steamed hams but its All Star” makes a direct connection to the Bee Movie / Shrek vortex of meme signifiers, but as Whyman notes the prevalence of “Steamed Hams but its…” posts, he implicitly extends shitposting to include technical memes, without using the exact name. Whyman ultimately argues that shitposting is a response to digital overstimulation that forges new pathways and connections between cultural signifiers. “The shitpost, then, is the new art form we as late capitalist subjects need to help us unconceal our world: the art of overstimulation” (Whyman 230), and in this way, technical memes seem a particular efficient mechanism for accomplishing those connections thanks to their content agnosticism.

That said, the cynicism of shitposting in general, and “x but ..” technical memes in particular, does not absolve these memetic processes from ideological commitments. Rather the precession of procedure disentangles subjective reception from the work of production by using hypermediation as an echo of that process. Barnaby Goodman goes as far as to suggest that hypermediation (in the sense that Bolter and Grusin’s coined for that term) characterizes technical memes (and to this I would add other “post meme” genres like surreal memes, antimemes, literal memes, and deep-fried memes) as an opposition to viral media of the 00s that, by contrast, strove for immediacy (Goodman, para.11).

In the case of “2020 Debate But,” hypermediacy is evident in the way that the mediation of my deformed artifact depends parasitically on the original mediation of the debate as broadcast by PBS, and the organization, disposition, and framing of that original mediation remains evident in my finished product. This includes the PBS logo, the organization of specific shots, and sequence of those shots in both the original and the derived artifact. Unlike Simpsons shitposts and Bee Movie deformations, however, the content of my video clearly creates a premise that could be seen as political. The hegemonic, normative artifact of the 2020 presidential debate – a media object that was almost immediately subject to replay, remix, and decontextualization in news commentary – works on a different register than does the source material for other technical memes, but perhaps by inserting supposedly political discourse into a procedural meme form, my video places the original content of the debate into the same purpose as reruns of the Simpsons serve for the Simpsons’ shitposters: familiar tropes, formulaic gestures, and performative ideologies are as factitious as their remixes.

Whether or not my video is truly representative of this deformed, technical, memetic formula, thanks to the serendipity of the YouTube recommendation algorithm, it found an audience for a few months in 2021. Whatever the individual reasons for its success and evident popularity (24K likes vs. 240 dislikes), I believe some of its resonance has to do with the way it responds to the anxieties and absurdities of the year 2020. Writing about the Occupy Wall Street movement and the memes that propelled it, Milner is cautiously optimistic about how memetic media can support robust public conversation through its polyvocal potential, striving to “facilitate counterpublic solidarity without trampling opposition” (Milner 184). The putative counterpublic to the presidential debates may be unified in bipartisan cynicism about the mediated spectacle and grandstanding that passes for political discourse, but its visibility can hardly be considered a countercultural resistance when its being seen depends directly upon an inscrutable recommendation algorithm.

Context 2: YouTube

YouTube is a platform for sharing and viewing videos, and since its major source of income comes through advertising, the recommendations that drive users to watch videos are the heart and soul of what we mean when we talk about YouTube. Because content creators ultimately lead user’s engagement to advertising, YouTube provides detailed analytics about each video in order to help creators more successfully reach their audiences, with the goal that those audiences ultimately spend more time on the YouTube platform.

Google’s use of artificial intelligence to tailor recommendations, particularly recommendations that favor longer, more emotionally-triggering content, has been criticized for contributing to the prominence of right-wing viewpoints on its platform (Daniels; Roose; Ribeiro et al.), but along with more aggressive de-platforming policies in 2020 and early 2021, anecdotal evidence suggests that YouTube has tweaked its recommendation algorithm to recommend more of the same kind of videos instead of algorithmically-adjacent topics, novel topics, and more extreme versions of familiar content. Whatever its current priorities, the YouTube recommendation engine ultimately remains a proprietary black box, but because I created this video, I am privy to rather detailed information about the viewers whose engagement makes up the 300K views on the video. The previous section focused on finding formal, aesthetic, and memetic antecedents to “The 2020 presidential debate but only the parts when someone breathes loudly or sighs” in order to speculate about what the video may mean to those who enjoy it. In this section, I dig deeper into that audience by highlighting key data points provided in YouTube’s analytics.

What these data show is that the overwhelming majority (97.1%) of the video’s impressions arrived as a result of YouTube recommending it on their home screen or another video’s sidebar, and that 11.1% of those 2.1 million impressions resulted in someone clicking the thumbnail and watching at least part of the video. A high relatively high click-through rate confirms that the recommendation was a good one, apparently, since this click-through rate was significant enough that YouTube continued recommending it for about three months.

A screenshot of a graphic from YouTube’s analytics highlighting the statistic that the video received 2.1 million impressions, 97.1% of those coming from YouTube recommending my content.
Figure 2. “Impressions and how they led to watch time.” This is a screenshot of a view from YouTube Analytics. It is a graphic comparing impressions, views, and watch time in the form of a funnel.

In YouTube’s terminology, an “impression” occurs “if the thumbnail is shown for more than 1 second and at least 50% of the thumbnail is visible on the screen” (“Check Your Impressions and Click-through Rate”), so this figure includes appearances on the home page, the sidebar, and in the “Up Next” list that appears after other videos.

Another metric highlights one crucial mistake I may have made when editing the video. Figure 3 is a graph of viewer retention showing that, on average, viewers watched 44%, which is more than enough to understand the video’s premise. Overall, the declination of the graph shows that interest understandably wanes as the video progresses, but a significant dip followed by a spike around 10 seconds indicates that many viewers are using the playhead thumbnails to skip my on-screen title. Obviously, that title is redundant (as at least one commenter has complained), and it is inconsistent with the norm established in other technical meme videos that simply present the content of the video without commentary, explanation, or justification.

Figure 3: A screenshot from YouTube’s analytics view on my video with a line graph highlighting how viewer interest spiked in the first few seconds of the video and then tapered off steadily.
Figure 3. A screenshot from YouTube’s analytics view on my video with a line graph highlighting how viewer interest spiked in the first few seconds of the video and then tapered off steadily.

I can be confident that this dip is a significant figure from which to draw conclusions because it represents a statistically large set of viewers. And ultimately all those impressions and views are a result of some turning in mid-April 2021 – five months after I first posted the video – when the view count suddenly started spiking.

Figure 4: A line graph showing measuring views over the time period from 4/1/21 to 6/18/21. Views spiked to nearly 15K per day in mid-April with another, lesser peak near 10K in early-May. Views taper off steadily afterward.
Figure 4. A line graph showing measuring views over the time period from 4/1/21 to 6/18/21. Views spiked to nearly 15K per day in mid-April with another, lesser peak near 10K in early-May. Views taper off steadily afterward.

Prior to the period represented by the graph in Figure 4, the total views on the video were less than 1,000. So what changed in April? One clue may lie in the demographics of the viewers during that initial spike, which comes into focus when filtering the audience metrics to the first few days of the spike.

Figure 5: A dashboard view from YouTube analytics highlighting the demographic breakdown of viewers between April 6 and April 25, 2021. A series of bar graph clusters and a tabular summary reveal that the overwhelming majority are male.
Figure 5. A dashboard view from YouTube analytics highlighting the demographic breakdown of viewers between April 6 and April 25, 2021. A series of bar graph clusters and a tabular summary reveal that the overwhelming majority are male.

Figure 5 shows that for the crucial time frame of April 6, 2021 to April 25, 92% of approximately 100K viewers in this period were male and 8% female, and for both of the included genders, the largest age group by far is “18-24.” Most of these users viewed the video on a mobile device, and most of those devices were iPhones. A narrower majority (54.4%) of these users watched inside the United States.

These college-aged men using iPhones certainly make a desirable segment for advertisers, and the coherence of that demographic as it appears in my analytics may be enough to justify YouTube’s insistence on pushing my content in front of that audience: it seems to have worked. After that initial push, the viewer demographics have remained US- and iPhone-centric, but by June 2021, gender is more balanced with 77.6% of viewers recorded as male. The 18-24 age group still dominates with nearly half (49.8%) of all views, which is more than twice the second most active age group, 25-34.

Numbers like these take on more meaning in the context of other numbers, but this video is singular. Because its data is such an outlier among the videos I have uploaded, it is difficult to interpret this videos’ demographics by comparison with any others. In fact, the “2020 Debate But” video is actually a sequel to a very similar video I created for one of the 2016 debates, but that video only has fewer than 500 views so far. Earlier, in 2015, I created a similar audio artifact where I manually removed all of the words out of a lecture by William S. Burroughs. The video of that piece, published in a special issue of the journal, Enculturation, on cut ups, has even fewer views (“...”).

Did the young adult male audience drive the subsequent views of the video? And did YouTube’s recommendations turn to that demographic because it inferred that type of user’s affinity for technical, shitpost memetics? If so, is the recommender latching on to the meme format, or the premise that this video contains potentially political content? It is possible to gain some insight into that recommendation’s perception by its first audiences analyzing comments users have left on the video.

To undertake a more systematic approach to answering these and other questions, I scraped metadata for 2040 top-level comments and manually classified them in a tagging system. After classifying most of those comments, I weighted each comment by the number of likes it received and compared the two distributions. Table 1 lists the twelve most popular and noteworthy comment types, alongside their raw count (number of individual comments) and weighted score (number of comments plus the number of likes those comments received).

Label Raw Weighted
this is better than the original 344 28791
it sounds like something in minecraft 219 6714
you are now breathing manually 192 3478
sounds like when I (or my dog) exercise 155 10024
this must have been hard to make 148 8749
this could be music 130 1041
actual insights on debate broadcast 106 7123
why does this exist 100 5680
thank you recommendation algorithm 77 5424
puns 74 4913
asmr 72 2048
this sounds nsfw 47 58

Table 1: A summary of recurring comment types sorted by raw frequency.

Chart 1: A bar graph showing the like-weighted distribution of popular comment types.
Chart 1: A bar graph showing the like-weighted distribution of popular comment types.

As Table 1 and Chart 1 both illustrate, the most frequently-occurring joke involves some comment about how the edited video is an improvement on the original debate. This includes the top three like-getters.

They both made good points in this version.
The calmest version of this debate that I’ve ever seen.
This is the hard hitting debate that the government doesn't want you to see.

A smaller but still significant number make a specific comparison with Minecraft sound effects, "Lowkey sounds like eating in minecraft tho,” and comments comparing the emphasized breathing sounds to themselves or a pet after exercise, “Me, a physically unhealthy adult after climbing three storeys:,” or “The class after the GoNoodle video”.

Interestingly, a set of responses make some reference to the way that the video’s content makes them aware of their own breathing.

For some reason this video is stressing me out and I'm having a hard time breathing lol
Fun fact: you are now breathing manually
how is this calming yet giving me anxiety.

This effect, which I experience as well, can be attributed to the fact that most of the sounds spliced together are of sharp inhaling with no release until the final few seconds. The unresolved tension builds as the video progresses which may create an impression of hyperventilation. The phrase “you are now breathing manually” is also a verbal meme that KnowYourMeme.com traces to 2004 (tomberry). This and other comments demonstrate the participatory resonance of memetic expression; by affirmatively inserting references to other memes, participants in this comment thread signal recognize and amplify the intertextuality of the video.

Related to the “technical” trope identified earlier, at least 148 commenters make some reference to the effort that this video must have required to produce. I replied to many of the most-liked comments in this vein with a brief explanation of my method – which was not time-consuming at all – but the impression still persists. This interest in the editing process recalls Plaugic’s explanation of technical memes’ workmanlike commitment to an arbitrary premise: the idea that someone spent time on this is part of its curiosity.

*Someone sacrificed their time and energy to craft this masterpiece for us. Well done sir.*
Let’s take a moment to appreciate how much work this guy went through to make the presidential debate actually watchable
 Someone made this

Meanwhile, a less frequent but closely related set of comments seeks to undermine or discredit my effort.

this literally describes the entire debate. Each time someone spoke they did it, this is just the entire debate with no words.
This is fake 😐

And finally, another major theme that emerges in some comments is to recognize the algorithmic processes at play that made the video available to their attention. These commenters explicitly thank the recommendation algorithm or attest that the video will soon inevitably enjoy even more visibility.

The almighty algorithm has gathered us here today.
This is probably one of the best random videos that just popped into my recommended section
Ah, beautiful to see a viral video before it goes viral.
just reserving my “i was here before this got to 1,000,000 views” ticket

Like the comments imagining my efforts in editing, these algorithmically-oriented statements demonstrate an awareness of processes at work that prepattern and make possible the perception of the work. In this way, although the technical or procedural meme trope makes its deformational processes more visible by being arbitrary and literally attested in their titles, memes like this are always already the fungible units of internet attention machines.

The coding schema I employed is somewhat rough, and even though it accounts for 90% of the comments sampled (the remaining 10% being too idiosyncratic or ambiguous to warrant a specific label), the twelve categories listed here so far only represent a small portion of that 90%. A wider typology of comments (weighted by likes) and classified simply by whether they attempt to make some sort of joke provides a reliable sense that most viewers find the video humorous in some way. In other words, the purpose of leaving a comment is, in most cases, to show that they are “in” on the joke, whatever they perceive that joke to be.

Chart 2: A bar graph showing the total like-weighted comments that are jokes contrasted with all other comment categories.
Chart 2. A bar graph showing the total like-weighted comments that are jokes contrasted with all other comment categories.

When I created the “2020 Debate But” video, I was aware of the deformative technical memes circulating on YouTube. In general, these memes appear to select content arbitrarily or, at most, ironically, but deformation when directed toward a position of power can still have a critical point of view and purpose.

On reflection, if I had an intention when creating this piece, it may have been to hold up the absurdity of something like the debate and to critique the premise that the two candidates’ could impact their chances of being elected by their ethos and skills in persuasive rhetoric. The preponderance of jokes throughout the comments generally confirm my premise, but unfortunately, a small number of commenters have taken it in other directions. These comments are a reminder that formalist aesthetics, whether in conceptual art or technical memes, must also exist in a real world of heterogenous ideas, experiences, and positions. An apolitical pretense is an affordance of privileges that I benefit from, and the year 2020 witnessed several disproportionate tragedies that inevitably situate “2020 Debate But” into a social context.

Context 3: 2020

In the future, 2020 will likely be known as the year lost to COVID-19, but it is also a year during which the murders of George Floyd and Breonna Taylor by police sparked a summer of protests. I created the video with meta- or apolitical intentions, but in a nation and a time when ethnic and other identities are disproportionately oppressed, an apolitical stance is a privilege that does not extend to many beyond my white, able-bodied, cisgendered, tenured subject position. If it can be said to have a satirical purpose, I believe the video “punches up” (if it punches at all), but the fact that it does so with little threat of repercussion to me personally implies that I participate and benefit from the hegemonic status quo represented by the two figures in the video more than not.

In particular, taking Donald Trump’s speech as a field on which to play with my algorithmic deformation normalizes his language as political discourse. Trump’s policies, ideas, and statements are not funny, and making fun of him trivializes the real harm and immorality of those policies, ideas, and statements.

The former President’s role in failing to effectively manage a response to the COVID-19 pandemic is an important context worth considering that makes the video unintentionally relevant. I created this video as then-President Trump was undergoing treatment for COVID-19, and it is likely that he was experiencing symptoms during this debate performance which featured the heavy breathing that is the content of my deformation. My video became popular in April as the United States began to see the impact of widespread vaccinations making headway against the pandemic, so perhaps the distance from that event makes more palatable the schadenfreude of Trump personally experiencing the impact of a pandemic that his lack of leadership helped make worse. 17 comments on the video mention COVID explicitly, with several others making reference to oxygen or supplies (e.g. “So this is where all the oxygen went”) but it is not clear that these commenters or any others perceive Donald Trump’s audible breathing as a way to understand or empathize with his humanity.

The phrase “I can’t breathe” were among George Floyd’s last words as he was being murdered by former police officer Derek Chauvin (Video Of Fatal), in a haunting echo of Eric Garner’s and others’ last words before being killed by police. These events were not on my mind as I created the video, but it clearly occurred to a few of the users who left comments, and perhaps many others. Five comments (some now deleted) so far have attempted to interpret the video as a joke about George Floyd’s death. It is impossible to know how many others thought of the same joke, or how many cringed at the irony of a video about breathing and simply did not watch it. But it is also impossible for me to ignore the role I played in making this joke possible, even if it was just for those few.

Earlier, I associated my video with others that fall into a general meme meta-category called “shitposting.” Whyman’s analysis focuses on the formal qualities of shitposts, but in a social or political context, shitposting is one way of describing the 4chan-powered, metapolitical trolling that elevated Donald Trump’s presence on social media well before the 2016 general election and the alt-right before that (Switzer; Maly; Keenan). The white supremacist terrorist who murdered Muslim worshippers in Christchurch, New Zealand even referred to his act as “real life” shitposting (Munn). These disturbing contexts are as much part of this video’s background as the technical memes and “steamed hams” deformations.

What does a video with nothing but breathing mean? I remain ambivalent about whether or not “The 2020 Presidential Debate But Only the Parts Where Someone Breathes Loudly or Sighs” is truly viral or even an example of a technical meme, even though it is by far the most viewed, read, or watched artifact I have ever created. The technical apparatuses involved in its production, circulation, and reception all have antecedents and contexts, but the social reality amidst the collective traumas of 2020 and 2021 overshadow any pretext of ideological non-commitment. While I may some day decide to unpublish the video or at least provide more context in its description to make clear what my intentions were (and more importantly, what they were not), it exists for now as another artifact emerging from the milieu of 2020.

In the end, my video is just one drop in an ocean of buckets, and its daily view count eventually plummeted in June 2021 almost as rapidly as it increased in April. The algorithm has moved on.


This brief episode of popularity on YouTube was unexpected, but it can nevertheless be instructive. Although I originally produced the video without a clear agenda – artistic, ideological, or otherwise – the ways that I came to care about the significant strands of interpretation helped me understand a sense of purpose that I may pursue in future creative work. The deformative meme format may persist in various ways, but since its popularity coincided roughly with a period of heightened public awareness and scrutiny of recommendation algorithms, it will likely yield to other motifs that connect to emerging ideas and concerns in digital culture.

The video I created, “The 2020 Presidential Debate But Only the Parts Where Someone Breathes Loudly or Sighs,” may or may not be artistically significant in a grand scheme, but the number of commenters seeming to enjoy it for the same reasons that I did does provide a sense of validation for my creative work. Furthermore, the tendency of users to comment on my video with similar jokes demonstrates the participatory nature of memetic culture. Something about my video offered low-hanging fruit for users to signify that they were in on the joke, and YouTube’s comment threads gave those users a platform for seizing that opportunity.

Altogether, these algorithmic affordances fall short of providing an exact template for future work, but they do suggest the value of brevity and high-concept directness. For example, as Figure 3 shows, many in my audience skipped past the opening title text, suggesting that the video’s title was sufficient and that the on-screen title was redundant. This confirms the received wisdom from content creators: the recommendation algorithm will favor work designed to catch and hold attention in its first few seconds. This principle is is clearly responsible for the low-quality, engagement-driven click bait that floods these video platforms with content designed to manipulate viewers’ ever-shortening attention spans, but an optimistic interpretation of this phenomenon is that media artists and scholars may be able to use the same techniques to produce critical, conceptual content with similar attention to brevity, clarity, and the ever-evolving meme ecology.

Works Cited

Avoid at All Costs. “Bee Movie Trailer but Every Time They Say Bee It Gets Slower.” YouTube, web.archive.org/web/20170405105617/https://www.youtube.com/watch?v=ZW-lDX2UcL0. Accessed 4 June 2023.

Bee Movie but It Keeps Getting Faster. Directed by emilio, 2016. YouTube, youtube.com/watch?v=o0zqF3l0TsI.

“Check Your Impressions and Click-through Rate.” YouTube Help, support.google.com/youtube/answer/9314486?hl=en. Accessed 19 June 2021.

Daniels, Jessie. “The Algorithmic Rise of the ‘Alt-Right.’” Contexts, vol. 17, no. 1, Feb. 2018, pp. 60–65. SAGE Journals, doi:10.1177/1536504218766547.

“Friday” but It Slows down Every Time She Says “Friday.” Directed by Zach Whalen, 2016. YouTube, youtube.com/watch?v=y3k6hwRa0AI.

Goodman, Barnaby. “Amateur Content Creation as Compositional Practice: Viral Videos and Internet Memes in Online Participatory Culture.” Sonic Scope: New Approaches to Audiovisual Culture, Feb. 2021. www.sonicscope.org, doi:10.21428/66f840a4.b5f945b4.

Keenan, Dillon. “DON’T LET YOUR MEMES BE DREAMS!”: How the Alt-Right Uses Memes to Build and Further Their Dangerous Reach Online.

Lavigne, Sam. Antiboredom/Videogrep. 2014. 19 June 2021. GitHub, github.com/antiboredom/videogrep.

Lewitt, Sol. “Paragraphs.” Artforum, June 1967, radicalart.info/concept/LeWitt/paragraphs.html.

Maly, Ico. “White Terrorism, White Genocide and Metapolitics2. 0.” Diggit Magazine, 2019.

Mary Kate and Ashley Olson - Gimme Pizza Song (Slowed Down). Directed by Cretchy Chretien (cr3tchy), 2010. YouTube, youtube.com/watch?v=CJEoASUMZbI.

mesiuepiescha. Justin Bieber - U Smile (Slowed Down 800%). soundcloud.com, soundcloud.com/mesiuepiescha/justin-bieber-u-smile-slowed-down-800. Accessed 4 June 2023.

Milner, Ryan M. The World Made Meme: Public Conversations and Participatory Media. Reprint edition, The MIT Press, 2018.

Munn, Luke. “Alt-Right Pipeline: Individual Journeys to Extremism Online.” First Monday, 2019.

[NOISE]. Directed by Zach Whalen, 2016. YouTube, youtube.com/watch?v=-KTwEBDRlrM.

Plaugic, Lizzie. “Smash Mouth, Bee Movie, and Jokes That Require Work.” The Verge, 4 Jan. 2017, theverge.com/2017/1/4/14125366/sped-up-meme-slowed-down-smash-mouth-barenaked-ladies.

Ribeiro, Manoel Horta, et al. “Auditing Radicalization Pathways on YouTube.” Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, Association for Computing Machinery, 2020, pp. 131–41. ACM Digital Library, doi:10.1145/3351095.3372879.

Roose, Kevin. “The Making of a YouTube Radical.” The New York Times, 8 June 2019. NYTimes.com, nytimes.com/interactive/2019/06/08/technology/youtube-radical.html.

Shifman, Limor. Memes in Digital Culture. MIT Press, 2013.

Shrek but Only the Adjectives. Directed by Goo Lagoon, 2017. YouTube, youtube.com/watch?v=REN-a7ndbo0.

Switzer, Ryan. Unhooding the Alt-Right: Framing and Metapolitics in Far-Right Discourse. Central European University, 2019.

The Bee Movie at 3000% Speed except When They Say “Bee.” Directed by Esp, 2019. YouTube, youtube.com/watch?v=7apltfVJBwU.

“The Bee Movie” but Only When They Say “E.” Directed by Iss 2uugo, 2020. YouTube, youtube.com/watch?v=Z6a9_7ZU2Pg.

The Bee Movie But Without Bees. Directed by That One Ghost Named Asdfghjkl, 2016. YouTube, youtube.com/watch?v=eagjpMz2hHU.

tomberry. “You Are Now Breathing Manually.” Know Your Meme, 2014, knowyourmeme.com/memes/you-are-now-breathing-manually.

Video Of Fatal Arrest Shows Minneapolis Officer Kneeling On George Floyd’s Neck For Several Minutes. 26 May 2020, minnesota.cbslocal.com/2020/05/26/george-floyd-man-dies-after-being-arrested-by-minneapolis-police-fbi-called-to-investigate/.

Wesley, et al. “Bee Movie But.” Know Your Meme, knowyourmeme.com/memes/bee-movie-but . Accessed 4 June 2023.

Whyman, Tom. “Oh, They Have the Internet on Computers Now? The Online Art of The Simpsons.” Post Memes: Seizing the Memes of Production, PDF, punctum books, 2019, p. 50pp. doi:10.21983/P3.0255.1.00.

“X but without Y.” Know Your Meme, knowyourmeme.com/memes/x-but-without-y. Accessed 4 June 2023.