Grassroots scaling up: Navigating algorithmic scales in a technopolitical landscape

Abstract

My theoretical aim in this article is to focus on an examination of processual enactments of scale in light of the technological affordances that are currently at the disposal of a significant majority of humans. I offer the terms algorithmic scales and algorithmic scalar affordances to describe one activist's engagement of practical theories of scale—her ‘algorithmic imagination’ (Bucher 2017)—which led her to design her audience in ways intended to algorithmically scale up, or amplify, her activities on Facebook—to enhance their spread numerically, rapidly, and translocally, making use of the algorithmically constructed communicative possibilities or affordances available to her on the site. (Social media, scale, algorithms, audience design, Facebook, activism, nation, politics)

Network affordances and the manipulation of scale in contemporary nationhood

My theoretical aim in this article is to focus on an examination of processual enactments of scale in light of technological affordances (and limitations) related to social media, internet, and mobile technologies. The specific focus is on an exemplar of one American activist, Dona's,¹ maneuvering upon algorithmic scales, as I term them, as she aims to effect political change regarding a voting-rights incident. To arrive at this notion of algorithmic scale, I review below a selection of prominent takes on scale in linguistic anthropological literature to date, and then provide a brief overview of how we have arrived at the set of scalar possibilities that technological innovations and affordances have recently made possible. Then, I engage with Dona's practical theories or ‘folk understanding’ (Bauman & Briggs 1990; Niedzielski & Preston 2000) of scale relying on these technologies—her ‘algorithmic imagination’ (Bucher 2017)—that led her to act in ways that she intended to algorithmically scale up, or amplify, her activist activities—to enhance their spread numerically, rapidly, and translocally. My focus on Dona's ideological conceptions of scale and scalar possibility in this article are based on actions that signal ‘audience design’ (Tagg & Seargeant 2014) through a series of posts based on activities and events that she chooses to entextualize in certain ways (Bauman & Briggs 1990; Leppänen, Møller, Nørreby, Stæhrc, & Kytölä 2015), rather than on any explicit statements from her regarding her thoughts and ideologies concerning how she navigates scalar affordances as an activist.

Scales, power, and politics

Blommaert (2007) notes that traditional (read mid-late twentieth century) terminological frameworks for sociopolitical organization were largely ‘horizontal’ and only spatial, and in being unidimensional they were sorely unrepresentative of the infrastructures of actual social organization. Invoking socially shared metaphors built on shared understandings of ‘TimeSpace’ (cf. Wallerstein 1998), Blommaert (2007) notes that social organization is built upon these ‘horizontal’ and ‘vertical’ shared (or not) systems of (e)valuation that are imbued with power and inequality. Moves in and across these scalar frameworks—sets of possible understandings of self and other positioning in space and time²—are moves reflective of claims to shared knowledge, power, and status or lack thereof: ‘moves across… scale-levels are moves within a power regime’ (Blommaert 2007:16). Such power regimes can be nested, or fractally organized (Gal & Irvine 2019)—one seen to be residing within the other. For example, and relevant to this article, sociopolitically established institutions such as federal (national), state, and local governments in the United States are generally seen to be nested one inside the other, reflecting an ideologized geopolitical makeup within the country of a political power hierarchy that is commonly understood to condition political participation and significance, with power and importance generally seen to be increasing as one moves from the local to the national.

In Gal & Irvine's (2019) exploration of scale, the authors add to Blommaert's move away from horizontal/spatial and geographical notions of scale into ‘interpretive schemata that establish a scale of spheres of relevance’, which are largely a matter of ‘perspective’ (2019:225). The authors note that scales (and ideologies about them) may be inherited from sedimented institutional processes and practices—as in the US government example above—but they also may be improvised, built on the connections and narratives that subjective actors build from their life experiences. Sociality amid scalar experiences, for Gal & Irvine, then, is a notion that invokes shared sets of understandings—interpretive frames that form the grounds for what can be said and understood in communication with others (see also Blommaert 2020). To ‘know one's audience’ is to understand and speak to commonly understood perspectival understandings and ideologized framings of social possibility as situated in time and space. They point out that Blommaert has referred to this idea as the ‘scope of communicability of… invocations’ (Blommaert 2015, cited in Gal & Irvine 2019:231). Based on this, we can treat scales of interaction as deictic in nature—actors index the materially constructed and/or socially agreed upon scales upon and within which they act, and they also act upon these notions of scale in order to shift, modify, change, or reinforce them. Carr & Lempert (2016) emphasize this notion of ‘acting upon’ scales by emphasizing the agentive possibilities available to us in acts of scaling. While not dismissing scales as preconceived—sedimented through process, practice, and power—the authors’ focus on scale's processual nature serves as a call ‘to leave behind a priori scalar distinctions and instead empirically track how social actors carve and cleave—or scale—their worlds’ (2016:4). Here they invoke Latour's (2007) Actor Network Theory, in which the focus lies in observing actors’ agentive practices in, among, and upon the material affordances and constraints available to them in their lived worlds. In this light, Carr & Lempert offer that the anthropologist's work is to pay ‘empirical attention to how bodies, technologies, commodities, communities, ecologies, and built environments afford scalar practices and impose limits on those who try to scale them’ (2016:10) and further, that we must be reflexive regarding the fact that, since scalar perspectives always select and deselect certain elements of understanding and experience, those of us vested with institutionalized authority must be wary of our own perspectives in selecting and deselecting the scalar perspectives that we invoke in our work, given its institutionalized potential for naturalizing the erasure of some experiences and prioritizing others. Likewise, Gal & Irvine (2019:241) assert that ‘The onus is always on the oppressed people to make themselves comparable—to make some aspect of their world of experience visible to dominant outsiders’. To move along or within, or to negotiate the scales upon and within which one is operating, then, is to engage in no uncertain terms in a negotiation of power and status; to situate oneself, another, and one's words and deeds in a specific evaluative system. A simple and static model of such evaluative systems is offered by Blommaert with his table describing language forms in Time and Space, where lower scales are seen to embody momentary and local actions and higher scales embody those that are timeless and translocal or widespread. This is perhaps most easily applied to ideologically conceived forms of language, where we can treat a standardized and written form of a language as more highly scaled in both space and time—more transferable across both—while local dialects and speech norms can be seen to be scaled lower: they change far more quickly than standardized forms, and are thus more momentary, and they don't travel well (with mutual comprehensibility) across space.

Table 1. Scales in space and time (Blommaert 2007:6).

Blommaert's model is, by his own admission, simplistic and insufficient for the description of situated interactions (2007, 2019, 2020). He claims to treat such a rubric instead as a mere starting point for sketching some of the ‘scalar imaginations’ (2019) that might be at play in a given situation. Indeed, these valuations of ‘low’ and ‘high’ in the above table are highly dependent upon subjective perceptions. For instance, with regard to the ‘nested hierarchy’ model of US political institutions given above, some, including the activist under consideration in this article, argue that there is actually more power to be gained at local levels of government than in dealing with federal legislature, because in short, one's voice counts for more in a group constituted by a city's population rather than a nation's (see Zentz 2021a).

All this is to say that actors operate among preconceived scales that measure value in several directions and on several planes, relying on sets of subjective experiences and differently shared notions of power and value in space and time. It is for this reason that amid scales that are sociopolitically ‘sedimented’ over the longue durée like American political institutions, and those that are materially constructed like the social spaces that are algorithmically codified within social media platforms (Blommaert 2020; Bucher 2021) as I discuss below, scales—as scalar practices and ideologies—must be assessed from the ‘ground up’, in situ, and of course, as these authors have already noted, linguistic anthropological approaches are particularly suited to examining these processes as they arise.

Communication technologies’ scalar influences on political possibilities

Advances in communicative technologies over the past couple of centuries have enabled changes in political and informational configurations in such ways as to have led people to reconfigure their political activities from regionally focused topics, events, and politicians, to issue-based politics in the late twentieth century, and more recently to what is frequently called ‘flash point’ political mobilization, wherein issue-based organizations use cheap or free communicative platforms in order to reach out to as many people as possible in order to activate them politically regarding singular events in time (Bimber 2003; Postill 2018). The major communicative technological changes that have accompanied these shifts—consisting of developments such as mass printing, developments in transportation, the invention of the telegraph, telephone, widespread (AM) and very localized (FM) radio broadcasts, television's ‘massification’ of media consumption and then segmentation of audiences as channel choices expanded, and most recently the internet and social media, which have enabled both massification and fragmentation—have enabled those with influence to gain more (van Dijk & Hacker 2018), but have also enabled grassroots actors with little (as in, not widespread) influence to occasionally ‘scale up’ messages to mass audiences in ways previously only guaranteed to institutional actors (Postill 2018).

Users of social media in general, and particularly among grassroots activists (who are my focus here), have made prolific use of social media networks toward their own communicative and political goals, acting locally, translocally, and nationally—as well as synchronously and asynchronously—in order to achieve specific outcomes. Activists and organizations operating at various temporal and spatial scales both materially and ideologically constructed have begun to coordinate differently across space and time. ‘Free’ social media communication platforms have afforded them opportunities to communicate and connect in ways that challenge mid-twentieth century ‘massified’ notions of space, time, and political representation and engagement (Dennis 2019; cf. Habermas 1991; Tilly 2004; Anderson 2006; Postill 2018; Blommaert 2020). Such platforms have also allowed them to challenge ideas that activism is constrained to limited times and spaces due to communicative technological infrastructures that previously limited both the temporal and spatial scales at which small organizations and individuals could communicate (cf. Bimber 2003; Leitner, Shepard, & Sziarto 2008).

In using social media platforms like Facebook, Twitter, TikTok, Snapchat, and so on, ‘end users’ can communicate and ‘gather’ across both time and space (Castells 2015; Blommaert & De Fina 2017; Tufekci 2017). They can speak to each other from opposite ends of a city, a nation, or the world; they can converse synchronously, with participants sending each other messages at the same time, or, due to the ‘archived’ nature of online interactions (boyd 2010), their communications can also take place ‘across time’, asynchronously. That is, a person can send a message or write a post, and a potential viewer—or a potential hundred, thousand, or million viewers—might see it in the next minute or after several days. Based on this temporal and spatial flexibility and unpredictability, we might say that social media users more generally—and, in the case shown here, activists specifically—depend on the breadth and depth of their online social networks in combination with the algorithmic infrastructure of the social media platforms they use in order to navigate and influence space, time, sociality, and politics. They can engage in ‘multi-scalar strategies’ (Harvey 2000; Leitner et al. 2008): they can be strategically present locally, translocally, nationally at the same time; they can (attempt to) translate local actions instantaneously into nationalized moments; they can take ideas circulating at a nationalized scale and ‘downscale’ them by generating local actions. This ‘algorithmic’ maneuverability that social media networks have afforded to actors highlights the importance of approaching research on scale in consideration of the material affordances that affect the very possibilities available to them as they experience and act upon various scales of interaction and evaluation.

Introducing algorithmic scales and algorithmic scalar affordances

Based on the above discussion, I now offer some definitions before moving to data presentation and analysis. Scale is the processual and value-laden construction, both upon and by actors, of relevant notions of spatiality and temporality in a given communicative instance. Scale is not an ‘ontological given’ (Carr & Lempert 2016:6; see also Gal & Irvine 2019; O'Connor 2020), and the relation between structure and agency is fluid and constantly negotiated as actors navigate ‘online life’ and ‘offline life’ (see also Glick Schiller & Çağlar 2011; Canagarajah & De Costa 2016; Carr & Lempert 2016).

I offer in this article a relatively new layer of scale that we must be aware of and engage with and that I have not seen discussed as such, and that is algorithmic scale. In order to define algorithmic scale, though, we must first understand (at least in a very basic fashion) what an algorithm is. At its most fundamental level, an algorithm is defined in Merriam-Webster (2022) as ‘a procedure for solving a mathematical problem (as of finding the greatest common divisor) in a finite number of steps that frequently involves repetition of an operation’ and ‘broadly: a step-by-step procedure for solving a problem or accomplishing some end’. Note that this definition is not limited to computers—algorithms at their most fundamental level are simply sequences of defined operations. We are, though, mostly familiar with this term of late as it relates to computers, the internet, and particularly social media—or ‘social networking platforms’. To clarify what I mean when I refer, then, to Facebook's algorithms, I refer the reader to Bucher's in-depth explanation of the company's shaping of our interactions on the platform (2021:117ff.). Importantly, Bucher points out that Facebook's ‘algorithmic system’ relies on over 100,000 data points in order to execute two acts simultaneously: (i) to produce sales for advertisers, and (ii) to keep users on the platform (see also Thorson, Cotter, Medeiros, & Pak 2019). Such proprietary computations are the lifeblood of the ‘big tech’ companies we are currently most familiar with—Amazon, Facebook, Twitter, Google, and so on. They provide entire architectures for human interaction on those platforms. These architectures are material and knowable like the architecture of a house is knowable; however, like architecture, they require much detailed and expert knowledge in order to be understood. Additionally, these ‘algorithmic systems’ are proprietary and largely held secret by their respective companies.

As these algorithms are proprietary, we generally cannot know what they really consist of; however, users of Facebook generate ‘folk understandings’ (Bauman & Briggs 1990; Niedzielski & Preston 2000; Silverstein 2010) of the ways in which these algorithms work and how they as end users are scaled amid them. As such, the notion of algorithmic scales that I engage with below in thinking about Dona's online interactions is in keeping with Blommaert's (2019) assertion that the very notion of scales in general is really one of ‘scalar imagination’—it is a broad conceptual tool the implementation of which relies entirely on our ethnographic examinations of social processes.

A final definition is due, and that is of affordances. Eisenlauer (2014:73) claims that this term was coined in academic literature by Gibson (1977) to refer to ‘“action possibilities” afforded by an object or an environment in relation to social actors and their individual capabilities’ (see also Blommaert 2019). My discussion in the following pages focuses on the algorithmic affordances available in Facebook as my participants made use of them; however, the definition of algorithmic scale I provide below, I believe, applies broadly enough to encompass scalar affordances—possibilities for scaling practices—as they have been newly influenced by these widespread proprietary computational infrastructures. Here, then, I offer my definition of algorithmic scale, based on the definitions of scale, algorithm, and affordance presented above. Algorithmic scale is

a set of scalar affordances made available within the constraints of human-built computational systems that delimit social interactions grounded in constructed communicative, geopolitical, social, spatial, and temporal orientations that span on- and offline contexts.

Algorithmic scales do not displace any other type of scale that we might already be familiar with, such as the aforementioned ideological orientations related to established political institutions. I see algorithmic scales, or algorithmic scalar affordances, as a set of communicative possibilities and constraints that we must grasp and incorporate into our analyses and understandings of scale in a world increasingly mediated by mobile technologies and internet infrastructures. That is, as Blommaert (2019) suggests, these algorithmic affordances consist of an expansion of the sets of communicative contexts and possibilities that have come to life with the advent of the internet and social media (see also Blommaert 2020), not a replacement of any of them. We indeed see in the set of actions described below that in addition to Facebook's algorithmic affordances, Dona's social and activist networks, members of the news media, and larger activist organizations are all essential to the amplification of Dona's messages.

I demonstrate through the data example in this article some of the ways in which political communication and organizing strategies are shaped by contemporary communication technologies, in particular with an eye to how Dona attempts to ‘algorithmically upscale’ or ‘amplify’ her activist ‘performances’ using both on- and offline social networks. When I use the terms scaling up or upscaling in this article, I intend to convey that Dona, the participant in question, has scaled up in two ways: (i) she has sought to relate a local occurrence to a national conversation, and (ii) she has sought to spread an intimate and personal message featuring herself using her own phone and camera to a ‘mass’ group of people likely consisting of some combination of ‘people who agree with Dona's viewpoint’ and ‘people who could stand to be convinced that voter suppression is happening on a regular basis’. This relates to the idea of ‘audience design’, which I define in the next section. For the purposes of the current conversation, I am treating scaling up and amplification as largely interchangeable. There is an inherent ‘verticality’ and ‘horizontality’ to a notion of upscaling that, in my view, amplification moves us away from in a way appropriate to the discussion of algorithmic socialization, wherein networked end users are able to post and access messages differentially based on their behaviors in relation to the ‘networked spreadability’ and ‘archivability’ that a given platform's algorithm affords (Bucher 2021). However, and in the case below specifically, such algorithmic reach is still deeply intertwined with the notions of upward and outward spread typical of horizontal/vertical approaches to scale and also typical of the institutionalized and ideologized notions of scaled government—spatial and fractally nested—upon which Dona's activism rests. Both types of action depend on algorithmic imaginations and possibilities for communicative distribution and spread to as many people across space and time as possible, as well as to as many ‘power brokers’—such as journalists, members of various levels of government, and large-scale activist organizations—as possible. I therefore leave the two terms interchangeable for now; however, in future theoretical work they may bear better teasing apart.

A brief primer on the ‘grammar’ of Facebook posts

For the purpose of the analysis below, I provide a simplified explanation of the structure of Facebook posts (I describe them at length in Zentz (2021a), and this is a structure that applied during the time I conducted research, from 2017 to 2019). Facebook is a ‘microblogging’ platform that relies on people connecting with each other online based on their offline social networks, or via interest groups formed on the site among relative strangers (in public, private, or secret group settings). Facebook acts as a ‘pre-author’ (Zentz 2021a) or ‘third author’ (Eisenlauer 2014) to the social interactions that take place within its bounds. That is, Facebook shapes the ways in which its users can communicate and connect with each other across time and space by combining various semiotic devices made available on the platform and various algorithmic models that drive platform engagement. When authoring a post, users can share information multimodally by authoring texts, posting videos or images that they themselves have recorded, or posting live-streamed videos. While Facebook's algorithms for the prioritization of which posts are more likely to show up in a user's News Feed³ are unknown, it tends to be common knowledge that posts containing images or videos are prioritized over simple written texts.

When authoring a text, the notions of audience design—indicating through textual constructions who are the poster's intended addressees, overhearers, and so on (cf. Goffman 1959/1991; Bell 1984; Goodwin 1984; Tagg & Seargeant 2014; Zentz 2021b)—and context collapse—the idea that no one posting on social media can really be exactly sure of who their audience members are (Marwick & boyd 2010, 2014; Androutsopoulos 2014; Tagg & Seargeant 2014)—are quite salient in a poster's construction of texts. Some of the sociolinguistic means through which a poster might design their audience might be informal and emotive expression for a more intimate audience or a topic one is emotionally invested in (which we see below), or formal speech or specific registers if one wishes to speak as a/to other categories of person like ‘mothers’, ‘fellow activists’, and so on. Additionally, as the architecture of Facebook would have it, posts can be set to public or private, a setting that can also indicate intended audience. If a post is set to private, only a user's Friends can see the post. If it is set to public, it can be viewed by non-Friends, and viewers of the post can share it to their own Wall. A shared public post will be shared in its entirety, including text and images from the original post as well as a link to the original poster's Facebook profile. For users who want more people to see their posts and are aware of how to navigate Facebook's post settings, they will set a post to public so that others can share it to their own Walls for more people to see the text in question. Posters can also ‘tag’ Friends when they author a post. When they do so, the names—and links to those Friends’ profiles—appear at the very top of the post following the word ‘with’ (e.g. ‘Dona—with Lucy and Joseph’). Interlocutors can also tag Friends in the comment threads below a post, and finally, Friends who want to ‘bump’ the visibility of a post might make comments or add humorous gifs that are relatively empty of meaning but that simply serve the purpose of ‘triggering’ Facebook's algorithm into making the post more visible to more users. All of these acts are acts of potential amplification or scaling up; they make it possible for more people, across space and over time, to see the user's original post.

One final note regarding what I intend to show with the data below. Bimber (2003:231) describes that ‘sociotechnological developments do not determine political outcomes, but instead simply alter the matrix of opportunities and costs associated with political intermediation, mobilization, and the organization of politics’. In keeping with the above, this work does not set out to say whether or not these activists’ political activities are effective in creating political change or changing onlookers’ minds (see also Dennis 2019). What I intend to do below is (i) to show a process in which Dona navigates scales using social media tools and her social networks in order to strategically effect representations of stories that she wants to tell to mass audiences; to do this, I assess Dona's actions in terms of audience design: her sociolinguistic and algorithmically informed orientations to various real and potential audiences; and (ii) to discuss how these acts exemplify ways in which social media have generated new possibilities for small groups of actors to scale their social and political activities.

Leveraging power: Geopolitical downscaling and networked upscaling

From August 2017 to August 2019, I served as the secretary for the board of the statewide activist group named Pantsuit Republic Texas (PSR), a largely women-led, ‘progressive’, and ‘intersectional feminist’ organization.⁴ PSR and its more local counterpart, Pantsuit Republic Houston, which I had been a member of and involved with leadership with since November 2016, consisted of two very large secret Facebook groups that formed after Donald Trump was elected President of the United States on November 8, 2016. Dona, a Houstonian, was the founder of PSR. When I became secretary of the PSR board in 2017, she and I, as well as another leader of both PSR and PSRH, Joseph—also a focal participant—agreed that while I was secretary of the board, I would also conduct an ethnographic research project among the leaders of PSR and PSRH. My data collection eventually became largely centered around all of their Facebook activities on both their personal Facebook Walls and within the PSR and PSRH Facebook groups—all spaces where I interacted regularly with the participants. The number of focal participants numbered eight in total, and these were members of leadership of PSR, PSRH, or both organizations. Dona was one of the most active organizers among my participants, and she was also one of the focal participants who engaged most frequently with me in both my researcher and secretary roles. For the former, she provided regular responses to research questionnaires and interviews in addition to allowing me to collect posts from her personal Wall and posts that she submitted to the PSR and PSRH groups. For the latter, we regularly interacted around scheduling and planning for the PSR board's activities and goals. For a lengthy discussion on my engagement as an ethnographer, conducting mostly online ethnography, among these focal participants and these two organizations, see Zentz 2021a:chapter 2.

I have shown elsewhere (Zentz 2021a) that most of PSR and PSRH's rhetoric and actions leaned toward ‘down-scaling’ political activities back into neighborhoods, towns, and cities in order to ‘upscale’ their power. That is, contrary to my initial discussion of scalar ideologies above, wherein as one moves ‘up’ the scale from local to federal one also moves up the scale of power, the group contested this conceptualization of hierarchically scaled political power by claiming that more power could be leveraged at more local scales (in local seats of government and in other awareness raising actions). Despite these contestations, though, their work could not help but be influenced by the strongly nationally mediated sociopolitical atmosphere within which they were situated (Hopkins 2018; Klein 2020). Questions of nationalized politics were unavoidable and also, frequently, locally relevant. Precisely because the scales of political discussion were so strongly dominated by nationally scaled media, then, outreach—activities aimed at changing political minds and opinions—at these scales were now essential even in local political activism. As such, despite these activists’ frequent downscaling rhetoric and activities (‘act local’, ‘boots on the ground’), they also knew that they needed to scale up their activities by making themselves and their ideas more visible to more people, far beyond the scope of their city. Local occurrences relating to national ‘hot button issues’ needed to be massively ‘amplified’, particularly via social media affordances, and given the scope of contemporary nationalized political discourse and massified and translocal, networked, and hyper-fast communication infrastructures.

In what follows, Dona engages in a very locally situated activity—becoming an authorized poll translator for Korean-Americans, advertising the event, and helping individuals to understand their ballots at a single voting site. But when things go wrong, she taps into a very nationally salient issue of voter suppression, and she quickly and astutely leverages her social and activist networks, largely via online algorithms and her understanding of them, in order to bring city/county-wide (her Friends, local activist networks, local news media sources, the county registrar), then statewide (the state elections official, which has little relevance in this sequence of events), and then nationwide (ACLU and other national activist organizations making a statement on the issue) attention to this incident that she has experienced. Throughout this sequence of actions, she successfully (thanks to her cultivated networks and their interest in and uptake of the issues that she invoked) ‘scales up’ or ‘amplifies’ a salient moment related to a nationally salient issue in order to draw widespread attention to the issue of voter suppression and how it occurs in singular and localized instances like the one she has experienced. The specific data now follow.

The escalation of voter suppression of the Korean community: Taking voter suppression viral

Share everywhere now

Late on Sunday morning, October 28, 2018, I opened my Facebook feed to a video of Dona, recording herself on a live-stream (I viewed it after its completion, thus in archived form), walking briskly along the side of a public building and describing an experience that she had just gone through. Before she had started her live-stream, she wrote a text header to her post: ‘Voter suppression at Trini Mendenhall in Harris County. Share everywhere now’. She was directing viewers of the post to ensure that the video would be seen by more and more people based on those users’ own Facebook Friend networks. To enable this, she was using a posting format that Facebook's algorithms were, at the time, privileging in users’ feeds, and she also set that post to ‘public’ so that it would be shareable by any Friend who wanted to click on the ‘share’ button on the bottom of her video post. This multimodal set of actions exemplifies a communicative ideology in which Dona knows that a singular live video post will reach the top of her Friends’ feeds faster than a textual post (she confirmed this to me when I asked her about it), and by expressing ‘share everywhere now’ in the header to that post, she knows that she can request to those Friends who are likely to see this post as she is streaming it in the live feed format to share it—to amplify its spread rapidly across time and space.

Figure 1. Share everywhere now (Dona, Personal, October 28, 2018⁵).

In the two-minute video she explained quite passionately that the event that she was there for—ballot translation help for Korean voters—had been advertised in local Korean language newspapers, and that she had already engaged in the exact same activity with senior citizens at the same polling place the previous Friday. She explained that there had been some confusion because there were two separate Korean-American groups there—some teenagers attempting to collect names in order to get a Korean language ballot (this should have been done through the census, she explained), and Dona's translator group. Dona explained that many of the poll workers knew her and her fellow volunteers’ faces as well as their families’. Despite this, the group was kicked out of the polling location and told to stay at least 100 feet away from the entrance door which, at this location, meant that they had to stand at the street. She stated that there were two translators who had been allowed to stay inside the building, and ended her video explaining that attorneys were coming from the Harris County⁶ voting registrar's office, that she and her fellow activists had a civil rights attorney coming to the premises, and that they would be contacting some voting rights organizations as well. In the thread below this post, Joseph tagged a PSRH leadership team member who was in charge of voting rights. At least one other leader from the organization tagged several local television news channels in the comments. By the time I watched the video it already had 18,000 views, and that number ultimately climbed to 22,000. This means that the video post was well amplified—it was shared very broadly by Dona's Facebook Friends and their own Facebook Friends—through acts of sharing and tagging—in order to reach such a large number of viewers.

Create multiple posts over time

In her second video, a one-minute live-streamed video posted not long after the first, Dona stated that all Korean language translators had been kicked out of the polling place and that not only were attorneys from the county registrar's office coming, but also now the constable had been called. This video received over 12,000 views.

In her final video that same day, she filmed from the parking lot of the polling location, standing at the back of the parking lot (by the street) where she had been told she must remain. In the video she attempted to show the two large SUV constable vehicles containing barking police dogs that had arrived at the premises. This video was about six minutes long as she tried to document what was occurring after the constables’ cars arrived, and to re-explain what had happened that led to her and her fellow translators’ ejection from the site. During this video she explained that there were now zero translation assistants in the building. She also reminded the audience that she and her fellow translators had taken an oath not to influence the voters—they were ‘authorized’, sworn-in translators. This last video received just 1,700 views. In the comments thread I noticed at least one local Houston television news journalist inquiring about people she could interview for a story on this occurrence. Therefore, despite the decreasing views across her video posts, Dona's persistence in continuing to tell the story via this medium allowed the uptake that her Friends had helped her out with—through their tagging/contacting of local reporters, contacting civil rights attorneys, and so on—to take effect. The reporters who contacted Dona were able to catch up with her side of the story via these posts and see if they would like to ask her more about what had happened.

Reach news brokers to expand audience

The event was in fact picked up by at least two local news organizations. In an October 29, 2018 article titled ‘Volunteer translators kicked out of voting location after confusion at Houston poll’ on the webpage of local television news station KTRK, an affiliate of the national ABC broadcasting network, the author (Dobbyn 2018) wrote:

At only nine days away from midterm elections, a voter and translator are speaking out after a misunderstanding at a Houston voting poll.

…

On Sunday, the translators say they were forced out of the polling site by a worker.

“Some Korean elders came out and the students greeted them and said some things in Korean and she (election judge) got very upset and said ‘you can't speak in another language here that we don't understand because that could be electioneering’,” said volunteer Dona Kim Murphey.

Then, she says, deputy constables showed up. “The poll workers are deciding to call law enforcement on translators. That's absurd. It's intimidating and absolutely voter suppression,” said Kim Murphey.

The county clerk's administrator of elections, Sonya Aston, says translators are provided in Spanish, Vietnamese and Chinese based on census numbers. But she adds, anyone may bring their own interpreter into the polls but that person assisting must take an oath before helping.

Aston also says volunteer translators may not solicit voters inside. They can only approach and ask if help is needed if they are outside and beyond the 100-foot perimeter from the front door. But volunteers fear not everyone will be reached.

“We are waving our hands, flailing to get people's attention as they approach the entrance hoping they see us so we can provide them with Korean translation services,” said Kim-Murphey.

A report containing similar content made it into the Houston Chronicle newspaper—Houston's largest circulation newspaper—as well (Foxhall 2018). These articles spread the reach (of course, much less in Dona's own words than in the editorialized format of a news article) of Dona's stories beyond her expanded network of Friends to potential city-wide audiences who likely, as news readers, were to some extent interested in political issues, and their publication was largely achieved thanks to her original posting on Facebook and the tagging of accounts of the several local ‘power brokers’ listed to this point: journalists, news organizations, activist leaders, lawyers.

Dona was not done here in her attempts to continue to amplify this story; perhaps just local news circulation and the original amplification of the event as it unfolded on Facebook live was not enough attention for this very nationally pertinent issue that had all of a sudden become very personal for her. On November 1, four days after the incident, Dona posted to her personal Wall a call for participation in an ‘emergency press conference’.

EMERGENCY PRESS CONFERENCE [note time change]. PLEASE PM me if you can join.

I need people to show up on at an emergency press conference at the TRINI MENDENHALL COMMUNITY CENTER tomorrow, the last day of early voting, FRIDAY, NOVEMBER 2, 4-4:30PM. Translators (of any language), people who have limited English proficiency (especially African languages, Arabic, Tagalog, Urdu, Hindi, Korean but also Spanish, Vietnamese, Chinese and others). And allies!!

…

Also looking for allies from differently abled community to join us. They won this fight not long ago and have someone at the entrance of Trini to ‘solicit’ people who appear to have special needs. (Dona, Personal, November 1, 2018)

The next day at the press conference, Dona stood in front of a small group of individuals holding posters in different languages. A friend of hers who was also the owner of a grassroots news organization (and therefore had a large number of followers on Facebook) live-streamed the video from her personal Facebook page.

Figure 2. We expect definitive action (Dona, Personal, November 2, 2018).

In Dona's fifteen-minute press conference, she reported that a translation activity she had been engaging in for several years was banned when she and other volunteer translators showed up to offer translation services, at a date and location that they had informed the Korean-American community of, and they were removed from the premises because, as the people who ran the polls stated, they could not understand the translators’ language. She explained that the people in charge at the polling site then got state endorsement of what they had done—that is, the Texas Secretary of State had endorsed their removal from the site (this is possibly also a result of Dona's amplification of the event). Dona invoked federal (national) voting rights laws that clarified that her activity was legal, and then she invited leaders from other local activist organizations to speak about the need for language access in voting. In the discourses on display in the video, Dona discursively scaled this local occurrence up to tie it directly to a nationwide issue, invoking federal and state laws that, again, were being hotly contested around the country at the time. Her involvement of other activist leaders also broadened the scale of conversation from just the local Korean-American community to the vast number of linguistic and differently abled communities who might need translation services when they went to vote. This video received over 72,000 views, demonstrating significant amplification again thanks to shares and likes given by viewers of that posted video, which were likely increased in number thanks to Dona's request for help from her journalist friend as well as her banding together with other activists representing different communities. This use of local activist and journalist networks again allowed the network of possible viewers and sharers to expand beyond just her momentary and specifically Korean-American poll translator experience. She had scaled up—expanded the scope of—the topic and the social networks involved from those specific to her and to Korean-American voters, to an issue more broadly relevant to all voters who needed any sort of person-to-person aid at the polls. This message was disseminated again to yet more viewers through the engagement of another local news power broker in addition to the diverse activist representatives who had joined her.

Reach national activist organizations

Finally, on official voting day, November 6, 2018—a day when Dona knew that people's attention around voting would be heightened—she authored another post on her personal Wall, stating,

If anyone will be attempting to offer language assistance today to someone you did not accompany to the polls and you get pushback, please private message me immediately. Twenty-four civil rights and voting rights organizations and advocacy groups nationally have signed onto a letter demanding that Harris County allow translators / interpreters to offer assistance at polling sites today after Korean American translators were banned from a site in Houston. The Secretary of State ultimately upheld the pronouncement that offering to help translate is equivalent to electioneering or loitering. ?!?? So election judges and poll workers across the state may attempt to deny voting rights to limited English proficiency voters. We may be suing and welcome plaintiffs from other language communities to fight this egregious injustice. In the meantime, any suspicion for voting rights violations? 866OURVOTE. (Dona, Personal, November 6, 2018)

In this final post, Dona made clear that this was an issue of national importance, so much so that based on this singular instance locally and all the amplification work that she had initiated and that her Friends and social networks had helped her to amplify, she had leveraged her activist networks to get organizations nationwide to sign on to a letter directed to local and state-level government authorities. In so many words, then, the nation's eyes were on ‘us’, in Houston, Texas. Her original post had been ‘maximally’ scaled up with a set of nationally scaled organizations writing publicly to address an issue in Dona's specific locale, and her knowledge that she likely had Friends throughout at least the state of Texas who might be involved in voting activities that day led her to author this post warning them of potential voter suppression based on limited English proficiency, and a national hotline to call in case they perceived any infractions by poll workers on that day.

In the above sequence of events and posts, Dona makes strategic use of her knowledge of Facebook's algorithms and her personal activist networked connections. She uses her ‘algorithmic imagination’ (Bucher 2017) to creatively assemble interlocutors and strategically author Facebook posts that capture the potential for amplifying her messages to relatively large audiences. Knowing that a live-streamed video will be prioritized on people's News Feeds, she broadcasts live from a voting site where she claims that she and her fellow translators have been, and are still being, subjected to voter suppression by local poll workers. Relying on this algorithmic prioritization (cf. Bucher 2017; Maly 2018) as well as her Friend networks specifically on Facebook, she is able to quickly get the short video's views to balloon to 22,000 viewers. Because she posts on a ‘public’ setting, the video is able to be shared by anyone using Facebook, whether or not they are her Friend, and that helps the post's ability to be spread far, wide, and quickly via her close and extended social networks (these viewers are located across a broad geographical space; however, I do not have information on exactly where the message might have spread to in terms of Facebook users’ locations).

Within the textual conversation threads below these videos, Dona and her activist collaborators quickly tag as many journalists and fellow activists as they can in order to bring the ‘newsworthy’ incident to their attention. These actions construct the journalists as (potential) ratified overhearers (Zentz 2021b; cf. Goffman 1979). Some of those journalists, having viewed these videos under which they were tagged, then contact people involved in the ‘voter suppression event’, and publish articles in visible local mainstream news sources the very next day, expanding possible knowledge of the event's reach to a broader, community-wide, and likely politically invested (as news readers) audience. Later in the week, in order to keep drawing attention to the issue and keep expanding her audience, Dona calls on allies in other activist communities—translators from other language communities and now also poll assistants for other vulnerable groups who have also had to fight to be able to help voters. She enlists her friend and fellow journalist/activist to use the latter's own platform and online visibility in order to draw more and newer viewers, and this live-streamed press conference on this public Facebook page attracts well over 70,000 views, potentially among viewers who are subscribed to her friend's news page and therefore likely interested in viewing local news with a progressive slant. Finally, on national Election Day, Dona shows that she has been doing more networking and collaborating on- and offline in order to act not only on social media but also institutionally: ‘Twenty-four civil rights and voting rights organizations and advocacy groups nationally’ are giving this issue attention and will potentially be suing the state. This post is written for her Facebook Friends, more specifically those throughout Texas, and perhaps beyond, who might be involved in voting activities on that day.

Scalar ideologies at work

Dona's act of posting her live video and calling upon her Friends to share, rapidly, multiply (across many networks/groups), and so on, is an act of orientation toward the scalar affordances provided on Facebook. It is also representative of Dona's ‘folk understanding’ or scalar ideologies regarding how Facebook's algorithms work and how end users navigate the platform. First, she believes that a live video is prioritized in feeds (again, she confirmed this knowledge to me when I asked her why she posted a live video in these three initial instances); second, she believes that the discursive deployment in her video of her anger at what has happened is something that will enhance ‘broadcast-worthiness’ among her peers, particularly regarding a local instance of a nationally scaled topic of voter suppression; third, she believes that more shares garner more visibility, which can lead to more informational power brokers picking up her story and writing about it in the news. Finally, a more virally shared incident as well as more ‘backstage’ networking in order to reach out to national civil rights organizations will continue to help bring attention to this event and therefore, hopefully, more action against voter suppression. With this sequence of acts, relying on several strategically networked F/friends, Dona is doing all of the things that Carr & Lempert (2016) reference regarding scalar agency: she is orienting to both an audience and a potential audience, where direct addressees are her Facebook Friends and their Friends. Potential addressees or ‘ratified overhearers’ are journalists, politicians, and more broadly scaled activist organizations. By sharing her rage at this particular occurrence and by tapping into these broader networks via her live posts, via her personal connections in ‘backstage’ interactions (Goffman 1959/1991; Treré 2019; Zentz 2021b), and via her repeated textual posts over a time period of nearly two weeks, Dona has leveraged the amplificatory power of social media and networked activism, spreading this story that must be made known to as many of Gal & Irvine's (2019) unconvinced outsiders as possible, across potential audience sets of direct addressees and various types of overhearers or onlookers.

Based on the above exploration, I have operationalized in Table 2 the expanded notions of scale in an algorithmic era that I have been exploring in this article, taking the liberty of re-authoring Blommaert's (2007) reduced and idealized table to account for an algorithmic era of expanded notions of scale.

Table 2. Language scales in time and space in an algorithmic era.

Through this one example of Dona's voter suppression incident, I have attempted to flesh these aspects out in a more detailed manner; however, as Blommaert suggested with his own table, the table itself is to serve as a simplified guide that should be applied to other situations in more detail as well so that we may continue to refine the model of, and our understandings of, algorithmic scales. In it, we can see that to ‘scale up’ based on an algorithmic platform is to be massively amplified or to ‘go viral’—to get as many shares and responses as possible. This scaling up then allows for the post to reach large networks of people across space and time by reaching the top of people's feeds for longer and being shared by more people for a longer period of time, continuing the post's reach to more people and continuing to enhance its virality. That original viral post, though, is likely not enough to spread Dona's message as far and wide as she wants to, and thus she keeps at it over time, both via Facebook posts and via personal networked communications, continuing to draw the attention of journalists and political actors. All of the scalar factors in the above table are clearly at play in various ways at various times throughout the journey of Dona's amplification attempts based on her singular instance, on October 28, 2018, of perceived voter suppression.

To keep this table broad and minimalist like Blommaert did is to allow for this complexity to be present, only using the table as a guide to help us consider how these different factors might influence expansions or contractions in the scale of textualized activities in a networked and algorithmically architected era. That is, the specifics of the US local/state/federal trichotomy, of Dona's use of a live video, and so on, are likely not appropriate to other contexts, and so such a basic table should hopefully be maximally adaptable to other research sites.

Grassroots (algorithmically) scaling up

What I have argued in this article is that the internet, mobile technologies, and social media platforms have added a new aspect to scalar possibilities of interaction and informational distribution. I have termed these algorithmic scalar affordances, and in order to examine these affordances, I have examined how one activist, Dona, made agentive choices in audience design in a sequence of efforts aimed at upscaling an event of voter suppression and the political outcomes that she wished would ensue from the event and her broadcast of it. What is not new in a situation like Dona's is the need, on the part of ‘grassroots’ actors, to labor to make their stories and their points of view with regard to wide scale ‘hot topics’ visible and understood (Gal & Irvine 2019). It is also not new that activists rely on social networks that are built via face to face and person to person activities. What is (relatively) new now, though, is the set of scalar affordances available within Facebook, social media, and other mobile and internet-based technologies that materially enable a rapid and widespread distribution of such activists’ singular and momentary messages to broader, networked audiences and potentially various power brokers. Thus, social media have enabled (albeit inconsistently) specific actors at specific moments to reach far broader audiences with their social media posts than has previously been possible.

The specific effect that algorithms potentially have on our ideological understandings of and actions upon scalar affordances is that dissemination of texts authored by non-institutional actors to enormous masses of people can happen in very short time frames across vast geographic distances. However, I have argued that our expectations of how to scale up a message from local to national, from intimately personal to a specific ‘public’ or mass of people, remain centered around traditional institutional centers that organize our lives. Every element of the process of the architecturalization of our social interactions on social media platforms is subject to social, political, and material influences, many of which well pre-date the internet. This layered combination of influences at every step shapes the ‘end user’ product that we encounter when we engage on social media platforms. With this article, I have offered to add to the conversation one single concept, through two related terms, algorithmic scales and algorithmic scalar affordances, that allows us to focus more on the ways in which mobile media and internet technologies have altered our to-date conceptualizations of, and enactments of and upon, the scales of power and possibility that structure our daily lives.