Semiotic Review 3: Open Issue - article published April 2013

Thumbstruck: The Semiotics of Liking via the “Phaticon”

Kane X. Faucher


Keywords: Facebook; thumbstruck; liking; semiotics; phaticon; sign; symbol

Image: you like this
I put up my thumb... and it blotted out the planet Earth
-Neil Armstrong

This article will be an early attempt to ground the ubiquitous icon of the “thumb” present on several SNSs and online comment fora in both semiotic and semantic registers. The digital convention of making use of the thumb must first be clarified in terms of its status as either icon, index, or symbol, and furthermore what role it plays in human-computer interaction (HCI), gamification of SNSs, digital gesturality, and the inherent mechanisms of arithmomania that guide approbation in the command and control environments of computer-mediated communication (CMC) that rely on prompting to guide online behaviour. In addition, we might ask if the thumb functions as part of the currency in online social capital accumulation and social transactionalism.

Phatic Phacebook

When Facebook rolled out its social interactive feature of allowing users to “like” their friends’ posts and photos, it would prove far too reductive to understand the sign of the upward-pointing thumb as solely an approbation cue given the multiple contexts of intentionality that prompt a user to decide on granting their digital endorsement. At issue here would be what the ubiquitous blue thumb means in terms of signs. It is not the case that the iconic thumb is meaningless, but that its meaning is rendered ambiguous after penetrating the surface level of mere approbation, and that its meanings cannot be inferred precisely from its social context. One of the functions of the thumb is a popularity metric that supplies three parties with data: 1. The recipient with validation or recognition for a post; 2. The other users who see the quantity of “likes” as indication of popular value which may lead to “herding,” and; 3. Facebook itself which relies on a sophisticated clustering algorithm to deliver “relevant” user activity to other associated nodes in the network as well as possibly providing data for targeted advertising.

The Facebook thumbs-up (including the variations on the same theme found on other SNSs) is a digitized gesture signaling approval, approbation, agreement, praise or even on occasion a reminder to the receiver of the sender’s existence. It is in some cases a form of reward feedback system. In other cases it signals the minimum amount of social effort required between two nodes or users on Facebook. It is also a Western-centric gesture given that some cultures, such as in the Middle East, the thumbs-up carries the connotation of insult (akin to North America’s middle finger or the British two-finger salute).

Not all SNSs or comment fora restrict the thumb gesture choice to the upright position. Initially, sites like “Hot or Not” allowed a choice between thumbs-up or down, and this is also in use on various other sites and fora such as YouTube and Yahoo Answers. Facebook, however, only allows one of two choices that the user must decide between: to click on the “like” or to refrain from doing so; there is no “unlike” button on Facebook, possibly for reasons of mitigating conflict scenarios online, abiding in part by Leech’s (1983) “maxims of politeness,”1 and ultimately to foster a high-trust culture by removing esteemdamaging options that might discourage or inhibit users from sharing more content that is vital to Facebook’s need to learn more about its users for the purposes of targeted advertising and “relevant” content display for its users. However, some users have adopted creative measures for expressing their “unlike” of posted content by qualifying their comment to that effect. Given the phatic function of “liking,” and its nebulous signification, posted content is effectively a “raising of the flag” in the attention economy to which other users are free to ignore, like, or qualify their dislike by means of the comment box area. In this way, the “like” may appear to be a sign, but more functions like a signal to indicate the presence of content that other users can interact with by giving their (dis)approval. Despite the popular meme of pressing Facebook to adopt an “unlike” button, we should clarify that this feature already exists on Facebook; technically, to “unlike” a post suggests that a post was initially “liked,” and that the user has opted to remove that “like.” The intention of an “unlike” button is more properly a call for a “dislike” button.

The issue of the thumb’s meaning is problematized given the number of intentions that motivate users to make the decision to click on the icon. If, for example, the user posts content along the lines of raising awareness of starving children in Sudan, does another user’s clicking on “like” signify the user’s approval of starving children or approval of the poster in raising awareness? When a user updates his or her relationship status as “single,” do the “likes” indicate approval of the event, support during an emotionally difficult time, or possibly signal romantic interest? Without any qualifying comments associated with the “like,” the intention remains obfuscated. It is also possible that the “like” decision represents the minimum of social communication required to still qualify as a communication event, such as in the case where a user feels obliged after a long period of non-communication to remind another user that s/he is still a contact but possibly not capable or willing to engage in a more substantial communication event such as sending an email, and so we might qualify the act of “liking” to be a purely phatic function. In such instances of self-referentiality as a motivator for “liking,” there is the creation of indexical feedback loop where the instantiation of the “like” is effectively a reference to the user who initiated the action. There may also be instances of indiscriminate “liking” when a user may simply click on all visible content in his or her network (what we can designate as “thumbing through” or “thumb-thru communication”). It may also be an event of reciprocity or trying to create the conditions of reciprocity, such as a user feeling obliged to click “like” on the basis of a history of having had content “liked,” or to initiate the possibility of exchange by being “prosocial” in “liking” the content of other users in the hopes that the gesture will be reciprocated.

Digitus Pollex: Background of an Indexical Sign

One of the foremost historians on the issue of the thumb in gestural and symbolic contexts in ancient Rome, Anthony Corbeill (2004), links the convention of making use of the thumb in the gladiatorial context as a gesture that indicated whether a fighter should be spared or killed with the etymology of the term pollex with pollet (power). However, as he points out, contrary to modern representations of ancient Rome, the gesture of a turned thumb signified that the fighter was to be killed, not spared. It should be noted that translations differ, and that it is unclear if the thumb for “to kill” was erect or simply exposed (pollice verso) as opposed to hidden or sheathed (pollice compresso) so that the gesture would signify the role of the sword to either be unsheathed to kill the fighter, or sheathed to spare him. The gesture in this context of the revealed thumb “to kill” is an active one, or requires an immediate action to be performed, whereas the hidden thumb (thumb pressed firmly into the fist) requires no action to be taken. The thumb in the Roman context is associated with being a power gesture given that it is the power of sparing or taking life. The reason for why the thumb initially came to be associated with power might lead us to speculation, but it may be tempting to consider from the biological perspective the advantages conferred by having an opposable thumb which allows more diverse functionality for the hand in terms of tool-use. Today, the thumb is an instrumental digit in the use of devices designed for convenient handheld use, beginning with the invention of remote control and with the advent of texting on cell phones. It is here that we might consider that technology in multiple contexts is oriented by the pollex as the digit of power, coded in part by its natural instrumentality. Space does not permit us to consider the gestural range of meanings of the thumb that can signify anything from approbation, insult, to hitchhiking.

In its digitized representation, the idea of the pollex as a power token may play a part in understand in the persuasive power of the icon in games of influence played on SNSs, comment fora, and purchase sites. The thumbs-up/down gesture in the digital context may be considered a recommender system that has some effect on the opinions of those who view them. For example, on sites that employ this function to aggregate the public’s opinion of a particular product, an overwhelming number of “thumbs-up” may influence a purchase decision on account of perceived relation between popularity and value, if not also a means of conformity to majority opinion. That this also leaves open the possibility of manipulation to ensure a kind of “momentum” where the artificial means of producing more “likes” from either a coordinated campaign among several users (or in the creation of multiple accounts directed to “like” a particular page or content) will precipitate even more “likes” from the community.

What is lacking in the decision-making process of choosing between thumbs-up/down is a more fine-grained system that is not a simple disjunctive. That is, the incorporation of Likert scales in the ranking of products according to stars or other objects as positioned on a scale of 1 to 5 allow for more specificity as to the opinion of the user that chooses to endorse or reject some page, content, sale item, movie, etc. Moreover, we cannot be certain as to why a user “likes” a particular item, the level of conviction behind the endorsement, or if it is according to central or peripheral processing. All that can be inferred from the communicative situation are binary sentiments of approval or disapproval. An abundance of literature in computational linguistics is dedicated to predicting the success of recommender systems according to binary classification of user sentiments by means of the thumbs up/ down decision (Turney 2002; Pang and Lee 2002).

Active versus Passive “Liking”

Arguably, the hyperboles of love and hate are among the most overused words in the English language, and given the nature of the online attention economy to rely on occasion on sensationalism to acquire more attention among a large number of competing users, it may appear odd that the thumb would represent “to like” rather than “to love.” It may be of some utility to consider a difference between what can be called active and passive acts of “liking.” In conventional contexts, to “like” something does indicate approval, but not in a manner that suggests adoration or full commitment. In “liking” content, we can signal to other on the network that we approve of some item, but in a way that keeps in reserve a full endorsement as such, or that the restricted options available on SNSs and comment fora will not permit differing degrees of said commitment to endorsement. If we were to assign a rough typology in terms of motivational factors in “liking” content, this might be done in terms of the locus or target of the “liking” intention:

Reasons to Like

  1. Content Only: User possesses conviction that the receiver’s content ought to be publicly endorsed or recognized by approbation.
  2. Receiver Only: User wishes to demonstrate support and endorsement of receiver’s content even if the content is only of peripheral interest to the user.
  3. Content/Receiver Blend: User demonstrates genuine support for both receiver and content.
  4. Self: User endorses receiver and/or content for the purposes of personal gain (reciprocity, status enhancement, contest entry, etc.)

The above factors concern primarily active “liking,” but there are forms of passive “liking” that take on the characteristics of indiscriminate endorsement. Various theories on persuasion and compliance (Lazarsfeld et al., 1948) can also involve persuasion without conviction. For example, if a majority of users on a particular user’s newsfeed indicate approbation for a particular content item, the user may comply with adding his or her own approbation to the content without considerable reflection. In such cases when the action of “liking” is meant to call attention to the sender, we can say in such instances that the thumb is pointing back at the sender.

Let us assume a male user (U1) with 99 (U2...U100) friends on Facebook. U1 posts a photo of his new puppy. Within the first five hours, 22 users in U1’s network have “liked” it, which must mean those 22 users saw the post by either visiting U1’s page or it was visible in the newsfeed, and furthermore made the decision to click on the “like” button. Five hours later, only 11 more users record their “likes” before no more “likes” are recorded for that event. U1’s photo has 33 “likes.” Two questions emerge: 1. Why did the other 67 users not “like” U1’s photo? and, 2. To what degree did the 33 users “like” U1’s photo? To answer the first question, we can appeal to some possible scenarios: 1. There are some users of the 67 that did not see the photo, and thus did not know of its existence in order to make a decision to like it; 2. There are some users of the 67 that did not like the photo, or did not feel the need to record their liking of the photo. Admittedly, this is an oversimplification of the motivations that might exist among the users who did not opt to “like” U1’s photo. We might also question why any of the users “liked” U1’s photo, returning to our list of motivators above. Was it to signify approval of U1 for U1’s benefit, for others, or approval of U1’s content for U1’s benefit or others? Was it a communication event to draw U1’s attention to the users who “liked” the content? A full disclosure on the etiology of the digitized thumb gesture in each instance would require access to the users themselves in terms of their intentionality of (in)action. The particular “why” the thumbs up appears, or its specific quiddity in the communicational context in which it appears, cannot be inferred from the receiver’s end.

Each instance of Facebook “liking” as a phatic function requires an agent, an action, and an object. What bridges the sender and receiver of the communication event of “liking” is the interactive icon or “clickee” that functions as a target for engaging response to the posted object (the trigger). This sets up a simple feedback loop:

Figure 1. Simplified Facebook Notification/Communication LoopFigure 1. Simplified Facebook Notification/Communication Loop

Although the nature of the feedback can vary (especially if the user decides to leave a comment in addition to “liking” the object), the main point of contact is the thumb icon. The object-icon of the thumb becomes the communicative conduit for phatic expression insofar as it contains no other information beyond an automatic social exchange to indicate a user having seen and acknowledged a posted object. This should not come as a surprise that much of our digital social interaction, modeled as it is on postcybernetic systems of feedback and control in new media, should increasingly take on more phatic forms of communicative exchange which streamlines communication so that it harmonizes with rapid feedback delivery of content in terms of display and engagement. A combination of uses and gratification theory writ for a persistent social networking environment that privileges perpetual presencing has several potential consequences for how network sociality is conducted (Miller 2008).2

In terms of Shannon’s mathematical theory of communication, the thumb icon is a prime example of information exchange in a channel. The sender intends to deliver a message of “like” and there is no information loss in transit, and thus the fidelity of the information is maintained. The element of surprise is preserved in the specific event itself in terms of who is sending it, and the change in the number of likes overall. From the purely technical standpoint, there is no difference that can be inferred from user A and user B clicking on the “like” button since the digital message is identical. As a signal, it is simply the addition of another of the same kind. One cannot deduce from one “like” and another since the signal will only carry the additional information of the identity of the user, not any qualifying details as to why or for what purpose the message was sent. In this particular channel, the receiver message can only be the addition of the “like” plus the identity of the sender; apart from this, there is no surprise: a “like” cannot suddenly become a “banana” or “stop.” There is also no way of being able to distinguish between two “likes,” if one user “liked” the post or content more than another user if there is an absence of qualifying remark, such as posting a comment. Although behavioural practices will differ in terms of motivations for “liking” content, if the communication is identical for two users in simply “liking” the content, there is no method by which it can be inferred if the event was an instance of active or passive liking.

Toward a Semiotics of the Thumb-Icon: Symbolic Sign

To apply Peirce’s triadic model to the thumb, we may arrive at the following:

Figure 2. The thumb in a Peircean triad

Figure 2. The thumb in a Peircean triad

The functionality of the icon is present in the way it can be directly engaged by the user in terms of a decision (the user can opt to simply acknowledge its presence and its meaning as proof of the content being “liked” and take no further action, or the user can interact with the icon to “like” the content). The thumb icon is technically symbolic given that it is a convention among digital users that the thumb in its broader context outside of Facebook as well as within it refers to the act of showing approval. The thumb icon is not indexical because there is no direct causal relationship between the object and the representamen since the thumb carries different meanings in different cultural contexts and situations (for example, a hitchhiker on the highway is not “approving” of cars passing by, just as a Facebook user is not signifying to another user the desire to hitch a ride).

The development of icons in human-computer interaction (HCI) shares its history with the rise of computing. The use of icons or pictographic symbols to convey a maximum of meaning in the minimum of (pixelated) space owes a great deal to the earliest forms of writing, such as hieroglyphs. In HCI, icons are used extensively to represent the data and processes in the computer (Gittins 1986: 523). Icons have the advantage of being more readily understood than text, but this is conditional upon the symbolic conventions and shared cultural meanings associated with various visual objects. Computer icons will either represent represent an action or an object (Barr et al., nd). Icons that indicate certain processes such as the icon for “home” or an icon of a magnifying lens representing search are generally understood, but may be of very low communicational value (Thiry-Cherques 2010: 709).

If we were to understand the thumb as a form of digitized gesticulation, a Peircean semiotics would apply a functional view where said gestures are referential to an extracommunicative object. In the theory of indexicals and demonstratives (Nunberg 1993, Perry 1997), for example, pointing plays a role in signifying for the observer that some salient object in a field is being indicated by this gesture. Karl Buehler ([1934]1990, and later revamped by Jakobson 1960) provides us with three registers for gesture that can be either emotive (expression of sender’s feelings, such as giving the middle finger to express anger), appellative (a gesture that is directed at a receiver, such as an accusatory finger), or phatic (signaling the sender’s interest in maintaining or regulating the communication event such as an open palm indicating to the receiver to continue speaking, or making a downward motion with the hand to indicate a command for the speaker to lower his or her voice). It may seem at first blush that the “thumb” can be considered under all three registers since in the emotive register the thumb can indicate the sender’s feeling or attitude toward the user and/or the user’s posted content; in the appellative register the thumb is intended to be directed at the receiver and/or the receiver’s content; and in the phatic register the thumb may be an invitation to continue communication or to post more content that will meet approval. However agile the thumb may be in its gesticulatory registers, we once again cannot fairly disambiguate its actual meaning by an appeal to Peirce, Buehler, or Jakobson. It may be of some benefit to consider Greimas’ (1970) fourfold typology of gestures to gain a better understanding. Greimas retains some of Buehler’s registers, such as the emotive that he assigns as attributive, and the phatic register which he attributes to modal gestures that include both phatic and metasemiotic functions. However, Greimas appends two more: the mimetic gesture (such as bodily icons of visual events) and playful gestures (which might include dance). We can, however, rule out these two additional categories when we consider the digital thumb, unless we are tempted to consider it under the mimetic, which it might only partially be, as a mimesis of actual thumbs-up/down mediated by digital representation. We may, however, be able to repurpose or broaden Kendon’s (1982, 2004) definition by claiming that the thumb icon refers to a distinct and visually-based, and effortful movement of part of a digital body, where the gesture operates as an abstract reference to a concrete manual gesture. In this way, the thumb icon might be part of multimodal gesture signaling as channeled through digital representation. However, in taking this view, we might align ourselves with the notion that digital representation of gesture is solely metaphoric.

Claude Cadoz (1994) gives us a typology of hand gestures as being either semiotic (the conveyance of information to an environment), ergotic (the alteration or modification of the environment) and the epistemic (gather information from the environment). In the case of the digital thumb, once again all three types or categories apply, for the thumb serves a semiotic function in conveying the information to a digital environment that some user and/or user’s content has met the approval of n others. It also may serve an ergotic function since a “like” may contribute to the higher probability of that content being visible in the newsfeed on Facebook, or may rise to the top of of a comment forum if the comments are sorted according to popularity. And, it may serve an epistemic function from the standpoint of the algorithm on a SNS that makes use of data-mining for the purposes of clustering, sorting, and displaying content.

The thumb icon is nonverbal, and somewhat mimetic of gestural language, although the same principles that apply to face-to-face (f2f) forms of communication that make use of gesture do not neatly apply in a digital context. Given that interface design and functionality has from its beginnings made use of icons to mimic the gestures that are representative of f2f gestural interaction, these representations are encoded, reductive, and ostensive (the same might be said about the constraints associated with the limited variety of emoticons to convey emotional tone in textual communication). For example, in f2f gestural communication we can convey with an index finger moving in a circular motion the idea of something that moves cyclically, or possesses the shape of a circle, whereas this is not necessarily employed in textual-based CMC (although there might be a means of replicating these gestured using animated GIFs, or in the use of avatar gestural non-verbal communication such as can be employed on Second Life; see especially the work of Antonijevic 2008). The thumb gesture in the digital context relies heavily on a flattened and conventionalized understanding of what the thumb signifies (generally a form of approval).

Perhaps none too ironically, the physical gesture that results in the displaying of the digital thumb is not always precipitated by an actual thumb. If, for example, the user clicks on the icon from a desktop computer that has a mouse or a trackpad, they will most likely use an index finger. However, it is quite possible that the physical thumb is employed if the device is a smartphone. Just as the writing of LOL rarely indicates that the sender is literally laughing out loud, the clicking on the thumb icon does not indicate that the sender is physically raising his or her thumb in a gesture of approval. Both these forms are figurative in CMC.

If we are to interpret the thumb as part of the symbolic mode, this leads us into murkier waters depending on the specific cultural competence of the interpreter to understand the thumb as an approbation marker or endorsement cue and not as simply a disembodied hand composed of pixels. What are we to make of the /thumb/ + n, where n represents the number of users who have given their endorsement? What we are given is an icon (/thumb/), sign, (the plus-sign representing conjunction), and figure (the number that refers to a quantity that belongs to the set).

Thumb: Sign or Stimulus?

Signs require interpretation whereas stimulus relies on “blind” reaction (Eco 1976). For example, the growl of an angry bear does not require interpretation as much as it does an immediate reflexive response to flee. The thumb icon as a prompt is perhaps a visual intonation as part of a digital symbolic frequency code. Yet what is interesting about the Facebook thumb is its reliance on a textual descriptor that is effectively redundant; that is, the icon of the thumb which ostensibly signifies “like” is appended with the actual word so that content level reads “[representamen] users 1, 2...n like this.” The interactive button itself contains the icon and the word “like.” The reason for this redundancy is unclear, but may suggest a prudent measure to ensure that users who may not possess the full conventional context of the thumb icon manifestation will be informed by the textual descriptor.

The “like” button does not qualify as mere stimulus, but neither is it wholly a sign or pure icon. The interactive potential of the “like” button is very low: one either clicks or does not click on it to register a communicative event. We might also suspend our earlier claim that “liking” is entirely phatic if we are to consider that at least some minimum of information is being exchanged. In the case of “liking,” the information carried by this communicative event will contain the identity of the user doing the “liking.” However, beyond this, “liking” remains ambiguous in terms of its motivation, and no idea or additional information can be inferred from the act itself. In many respects, the phatic function of “liking” is a form of small talk alongside prefatory words like “well.” Phatic functions align smoothly with general principles of politeness and prolonging communication. The thumb becomes a simplification of phatic expression in the form of interactive icon that automates the function. It in this way that we may describe the thumb icon as a “phaticon.”

We might also state that “liking” itself as a gesture of communication in general, plays a rhetorical role in conveying meaning and as part of persuasion in the social context. One of the earliest works to make a link between gesture and visual rhetoric would be Gilbert Austin’s Chironomia (1966 [1806]) where he devotes study to an almost mechanical process by which gesture can be used sparingly to supplement speech for rhetorical effect. Austin’s recommendation was that hand gesture should be a supplement to speech, not a substitute. In our digitized environment where the thumb icon is employed, we find that the icon commonly functions as a substitute for text-based forms of giving endorsement.

Structural Aesthetics in Design-as-Use

The thumb feature as a point of interactive contact for users is of a piece with the overall aesthetics of SNS platforms. To this end, leaving aside the possibility that this is part of an agenda to infantilize social spaces in digital environments, this process involves a reification loop. So, for example, the structure of the site as powered by programming code conveys functionality and features that produce the aesthetic of the site. This aesthetic is interpreted by the end user in terms of his or her engagement with the site–– generally the user interface design (UI). This engagement “realizes” the software structure. In this way, interface design attempts to draw together both form and function.

Just as shopping data is gathered for the purposes of predicting future purchases on the basis of complex probability matrices, the action of “liking” particular content is no different with respect to being exposed to certain content items. Most SNSs are, in fact, obfuscated prison-houses that guide and direct human behaviour in their environments making use of several prompts and cues that constrain choice under the illusion of freedom, be it the limited options in a drop-down menu, the limited options for accessing content by constraints in access channels, or the algorithm that determines what will be relevant to the user. In addition, the use of command verbs that have been naturalized in the digital environment such as “click,” “go,” “update”, “comment”, “rate/rank,” and “like” are emblematic of the process of insubordination to the digital environment.3 By promising inclusion, the precondition is this insubordination, and so we either “accept the paradigm or we are excluded.” (Thiry-Cherques 2010: 714-5). In terms of predicting social behaviour, probabilities are imperfect, but their predictions become more strongly confirmed the more social behaviour is shaped and structured by the environments in which they occur. The aestheticization of the digital environment functions in this respect to occlude from view the very constraints that allow the algorithms to shape social behaviour. “Liking” might prove a paradigm example of the fundamental irony of digital environments as a persistent environments of efficient feedback mechanisms and prompt-inputs for our opinions and sentiments: we can “communicate” in higher volume, but possibly at the expense of subjective variation in meaning content.

Digital environments such as Facebook are heavily hodological in nature, which is to say that they operate according to preset pathways that delimit options for access and interaction. Although SNSs such as Facebook do not determine the content variations that can occur, its enframing mechanisms ensure activity occurs within its organizational grid, and that communication flow is governed by the options available to its channels or conduits.

Thumb icons on Facebook are strategically placed so as to encourage users to provide their “feedback.” On specific Facebook pages, the opportunity to be reminded of the “like economy” occupies several points in the UI. There is usually a large icon as part of the specific page banner:

Figure 3. Enlarged “Reminder” of likes.

Figure 3. Enlarged “Reminder” of likes.

Nested in the comment and activity feed will be the names of users who liked specific posts (or the names of users plus a number representing other users if that number is larger than two), and an area at the bottom that displays activity on that page for the recent month followed by a large number indicating “likes.” On other sites that encourage visitors to “like” them on Facebook, the widget does not always come with the thumb icon, but instead relies on the recognizable Facebook “f” as brand reference:

Figure 4. Facebook-referential widget nested on non-Facebook site

Figure 4. Facebook-referential widget nested on non-Facebook site

Thumbification, or: A Game of Thumbs

Capitalism and gamification in non-game contexts share zones of overlap in terms of rules and outcomes. For example, popular video games rely on models that permit players to engage in strategies of accumulation for character-development, sometimes involving “level grinding.” Highly corporatized SNSs like Facebook cannot necessarily be considered out-and-out exploitationware, but many of the incentives by way of self-aggrandizing features are indexed on increasing the nebulous property of “social capital,” but in ways that encourage competition and self-disclosure which feeds directly into its algorithmic dataaggregation mechanisms.

When we consider what “gamifying” the SNS means, one heuristic model is provided by Roger Callois (2001) who distinguishes between paidia and ludus. Sites such as Facebook are not “freeplay” or nonlinear systems that encourage freeplay, but are essentially highly regimented spaces with rules and implicit objectives. In understanding liking as part of gamification (in terms of ludus, or rule-based games with goals and objectives that need not always be explicitly stated as opposed to paidia which encourages nonlinear free-play) we can apply the social software lens as a structural model for the semiotic transaction that is taking place. With respect to Facebook that abides by the popularity metric of “liking,” what we are presented with is a non-game with gamified elements or features. One of the shared zones of overlap between Facebook and traditional video games is the aspect of entertainment (Deterding et al. 2012: 12). By insisting on game “elements,” we can broaden the idea of gaming to contexts that are not specifically indexed on gaming as such to speak of gamified environments. On Facebook there are plenty of examples of game elements in play such as the ability to construct digital personae and explore identity experiments, the “turn-based” nature of comments, and the competition for perceived limited resources of attention in the discourse of “winners” and “losers” in the popularity contest of “liking.” Some corporations have already implemented game-like elements into their own Facebook pages by creating contest events where eligibility is to “pay” in the currency of providing one’s “like” to the corporation’s strategy of accumulating said likes for the purposes of proof by popularity (if not also as a means of targeting advertising at users within the SNS).

One of the common features of the web and information environment is the strong appeal to quantity considerations as a basis for making decisions and assigning value as part of a credibility assessment, and this has already been demonstrated to work in what is called “herding.”4 Social proof mechanisms may prove more significant in a semi-democratic setting of mass participation since no individual user may have the time or inclination to inspect the claims of each individual user.

When presented with user-generated and supplied information or commentary feedback, assessing the credibility of the online source may determine endorsement or rejection. Where communication is conducted in a format that allows for user-generated commentary, we might come to think that all users can be assigned the same (likely low) value of credibility, and this may continue until some users become defined as influence-leaders in the communication setting due to whatever metrics (such as number of thumbs-up) or non-metric (frequency of quotation or engagement) may assist in establishing credibility hierarchy. Generally, comment sections of a considerable size can measure credibility according to the long tail, a term coined by Chris Anderson (2005) that makes use of probability in statistics to determine popularity. The “head” represents in this case the influence-leaders of the particular web community who are assessed as having the most credibility, while the “tail” is the number of users with low to no credibility. However, we ought not to mistake or conflate credibility and popularity since although these may correlate, to possess popularity does not necessarily confer credibility.

In the online attention economy which involves games of social capital accumulation, we might suggest that most social transactions are indexed on producing a persuasive effect to solicit other users to validate both the existence of the user as well as the posted content by means of positive association and status enhancement. What is of interest in SNSs in general would be the admixture of a high trust environment where users may feel encouraged to engage in acts of self-disclosure, and the highly competitive nature of accumulating social capital. Given that resource scarcity can give rise to conflict, it remains to be seen if, in the context of an online attention economy, conflicts might arise. When taken in its competitive aspect, there are clear indications of how some users (be them individuals or corporations) attempt to “win” by playing the game in ways that might not have been intended, and which call into question the very metric of popularity as a credible model for assigning value. For example, users intent to aggregate an abundance of “likes” on their posts or pages can purchase the services of others through and other sites which promise thousands of likes added within a specified time period using a variety of means of achieving that end. In addition, the rise of digital astroturfing and persona management software makes use of automation to construct bot-profiles on SNSs that can be tasked with “liking” content.

A surfeit of studies on Facebook use among adolescents points to the highly competitive nature of acquiring “likes” and how these function as “badges of esteem.” An equally abundant supply of primers have been written that attempt to train marketers to leverage social media tools such as Facebook in ways that liking can function as a reward system by means of luring, generally by means of small social incentives (exclusive content access, contest entries, etc.) in exchange for the user “liking” the company Facebook page. New trends in social marketing now treat “likes” as insufficient given that Facebook users can “pile on” to “like” content, but never return to it. For that reason, the new turn has been to measuring beyond volume and valence of “likes” and to assess using analytic tools the depth of engagement with the content (for example, how many users that “like” a page return to it, share or post comments on their own pages?). Facebook itself has taken an interest in content engagement with its redesign of fan pages to reflect the metric of how many users are “talking about” the content which attempts to replicate through keyword data-mining more traditional forms of marketing such as word of mouth (WOM) which, in the social media context, exceeds paid marketing action up to 30 times in terms of performance effect (Stacey et al.: n.d.: 1). Such findings do not come as a surprise given the persuasive effects of influence leaders in peer groups (Lazarsfed, et al.: 1948).

In terms of basic return on investment for every post made by a user on Facebook, the number of likes will be contingent upon a variety of factors including, but not limited to: 1) size of the user’s network; 2) how the post will be purposed by Facebook’s EdgeRank algorithm which will determine the post’s visibility; 3) relevance of the post to the user’s network; 4) topicality of the post; 5) the presence of strong emotive cues (humour, outrage, etc.); 6) the communicational transaction rate of the user’s network; 7) the quantity of existing likes as a persuasive experience for other users to comply with adding their own “like”; 8) if the post is more photo-based than text-based (given that the algorithm favours the former over the latter).

With further research, we may arrive at the relation that possibly exists between automated phatic functions and how these facilitate the gamification of non-game environments. The plain fact that Facebook controls the phatic function in terms of providing only one means by which this can be expressed quickly is of a piece with its overall ideology of assuming control over certain communication functions in a way that is homogeneous and consistent which aids the algorithm, maintains the social “game” in terms of efficiency and speed, and ensures some measured degree of polite social interaction.

We might hypothesize that the larger the number of likes, the greater the share of the attention economy, and the higher the probability that a) the content will be distributed through the network, b) will attract more likes. So what is the ideal threshold required? The principle here of “like goes with like” does not refer to the atomist principle explicitly, but the notion of conformity and herding do suggest that the numerical aspect of “likes” will attract more of their kind, specifically on account of content redistribution throughout the network that will signal the attention more users who will be given the opportunity to view and interact with the content. On some sites, such as, a dropdown menu provides the option of organizing the comments according to highest and lowest rated. Given that a majority of visitors will not “backread” a comment thread if the comments number into the thousands, if this sorting option is selected, then visitors may be more likely to congregate to those comments that have received the most feedback by other visitors, and thus base their decisions to approve or disapprove according to the highest volume of interactions.

Facebook operates according to two functions: connection (the connecting of users) and interaction (conversational content between users). In terms of social network analysis, informational flow is interactional in terms of affinities, and connectivity is relational. When an algorithm effectively assigns topics of interest to users, this can be done through weighted keywords inputted by the user. So, for example, if a user posts on his or her page some comments about what s/he ate for dinner, those words can be flagged and the algorithm can place similar items in the user’s newsfeed that reference food. Should the user input a “like” on another user’s page, the algorithm can use this data to provide similar, and allegedly relevant, content higher up in the user’s newsfeed in the future. In addition, the algorithm can operate predictively by assessing a history of the user’s “likes” to determine what content should be displayed that the user may also “like.”

As Freberg et al. (2010) note, those who are more successful offline tend to enjoy equal success when their social activities are conducted online. Given that many Facebook connections are predicated on an initial offline connection, the game of “liking” might correspond to the same “social riches” of aggregation of attention as experienced offline. However, what of digital social environments where a majority of users interact that do not have a previously established offline connection? We might ask what role the “thumb” plays in screen nonymous environments in terms of approbation and rejection, and what motivating factors are at play in terms of providing a thumbs up or down to UGC. We might expect that such environments will possess weaker social ties, and thus might encourage more honest reaction to posts in terms of assigning (dis)approval, whereas on Facebook any implementation of a “dislike” button might precipitate conflict among users who are ostensibly connected according to the broad parameters of “friendship” as opposed to UGC comment fora where users congregate and supply content on the basis of a shared affinity underwritten by the purpose or topic of the site. However, these same sites would more likely have higher instances of flaming and trolling than Facebook, and thumbs are but one weapon in the arsenal of those whose purpose is to attack another user’s credibility.

Other sites that employ the thumbs-up/down such as news site fora operate in ways that differ significantly from Facebook. On, for example, visitors can give their (dis)approval without registering on the site (but comments are restricted to registered visitors only). The number of thumbs up and down are recorded beneath a number indicating an average that makes up a score, or “rating.”

Figure 5. forum posting rank

Figure 5. forum posting rank

By contrast, does not make use of thumb icons, and instead displays a score while keeping invisible the number of likes and dislikes.

Figure 6. forum posting rank

Figure 6. forum posting rank

YouTube, on the other hand, does not average out the thumbs up and down, and instead relies on matching number of views with the thumbs up and down:

Figure 7. YouTube aggregate rank

Figure 7. YouTube aggregate rank

Each of these contexts in which the thumb icon does or does not appear, allows for both thumbs up and down or not, and the preexisting connective relations between the users that congregate to these areas that permit UGC, will present slightly variable uses of the phaticon. That each expression of the phaticon does begin with an attitudinal orientation toward the object, the user, or both, remains unqualified without textual support in the form of comments.

Phaticons, as digitally embodied endorsement cues, provide for ease of “small talk” and rapid social communication with minimum effort by way of partial automation. Those who make the accumulation of supernumerary “likes” a point of self-validation (or, in the case of businesses, proof of popularity among a consumer base), we find that such gamified systems in entirely hodological spaces are indexed almost predominantly on numerical values. It is less a question of who “likes” our post, but how many do. From a semiotic standpoint, the thumb icon does not prove too problematic; from a social and semantic context we might question the value of accumulating what is the equivalent of empty “whazzup?s” or “how’s it going?s” as a measure of self-worth and self-value in the online environment.


Anderson, Chris 2004 The Long Tail. Wired 12.10.

Austin, Gilbert 1966 [1806] Chironomia, or a Treatise on Rhetorical Delivery, ed. Mary Margaret Robb and Lester Thonssen. Carbondale: Southern Illinois University press.

Antonijevic, Smiljana 2008 From Text to Gesture Online: A Microethnographic Analysis of Nonverbal Communication in Second Life Virtual Environment. Information, Communication & Society 11(2): 221-238.

Buehler, Karl [1934] 1990 Theory of Language, trans. Donald Fraser Goodwin. Amsterdam: John Benjamins Publishing.

Cadoz, Claude 1994 Les realites virtuelles. Dominos: Flammarion.

Callois, Roger 2001 Man, Play, and Games. Chicago: University of Illinois Press.

Corbeill, Anthony 2004 Nature Embodied: Gesture in Ancient Rome. Princeton: Princeton University Press.

Deterding, Sebastien, Dixon, Dan, Khaled, Rilla, and Lennart Nacke 2012 From Game Design Elements to Gamefulness: Defining “Gamification.” March 8th, 2013.

Eco, Umberto 1976 A Theory of Semiotics. Bloomington: Indiana University Press.

Freberg, Karen, Adams, Rebecca, McGaughey, Karen, and Laura Freberg 2010 The Rich Get Richer: Online and Offline Social Connectivity Predicts Subjective Loneliness. Media Psychology Review 3 (1).

Gittins, David 1986 Icon-based human-computer interaction. International Journal of Man-Machine Studies 24: 519-543.

Greimas, A.J. 1970 Du Sens. Paris: Seuil.

Jakobson, Roman 1960 Closing Statement: Linguistics and Poetics. In Thomas A. Sebeok, ed., Style in Language. Cambridge: MIT Press, 350-377.

Kendon, Adam 1982 The Study of Gesture: Some Observations on its History. Recherches Sémiotiques/ Semiotic Inquiry 2(1):45-62.

----- 2004 Gesture: Visible Action as Utterance. Cambridge: Cambridge University Press.

Lazarsfeld, Paul., Berelson, B., and Hazel Gaudet 1948 The People’s Choice. New York: Columbia University Press.

Leech, Geoffrey 1983 Principles of Politeness. New York: Longman.

Miller, Vincent 2008 New Media, Networking and Phatic Culture. Convergence 14(4): 387-400.

Nunberg, Geoffrey 1993 Indexicality and Deixis. Linguistics and Philosophy 16: 1-43.

Pang, Bo, Lee, Lillian, and Shivakumar Vaithyanathan 2002 Thumbs Up? Sentiment Classification Using Machine Learning Techniques. Proceedings of the 2002 Conference on Empirical Methods in Natural language Processing. 79-86.

Perry, John 1997 Indexicals and Demonstratives. In A Companion to the Philosophy of Language, eds. Bob Hale and Crispin Wright. Chicago: Blackwell.

Stacey, E. Craig, Pauwels, Koen H, and Andrew Lackman. n.d. Beyond Likes and Tweets: How Conversation Content Drives Store and Site Traffic. March 9th, 2013.

Thiry-Cherques, Hermano 2010 Intranets: A Semiological Analysis. Journal of Information Science 36 (6): 705-716.

Turney, Peter D. 2002 Thumbs Up or Down? Semantic Orientation Applied to Unsupervised Classification of Reviews. Proceedings of the 40th Annual Meeting of the ACL: 417-424.


1. One of Leech''s key maxims, building on Grice, would be to “minimize the expression of beliefs which express dispraise of the other; maximize the expression of beliefs which express approval of other.”

2. The LivesOn project more than demonstrates this point. LivesOn allows Twitter users to subscribe to its service which allows a program to process the user''s tweets and use an automated method for producing new tweets on the basis of said corpus, long after the user has died. This form of “ghosting” the web makes use of bot communication to maintain digital presence even if there is no more corporeal presence.

3. This should be distinguished from implicit computer prompts, such as a flashing cursor. Prompting in network cultures is fairly standard currency where, for example, users are prompted to fill in details such as their name, hobbies, interests, etc. In addition, we might speculate that in colloquial language the transition from proper nouns such as Google into verbs contains a similar function with respect to creating command-words. In offline contexts, communicative events such as interrogation or simply being asked, “give me your name” where the verb “to give” is a command, this sets up an instant power relation between the issuer of the statement and the respondent. For a more detailed assessment of the role of command words in linguistics, see Deleuze and Guattari, A Thousand Plateaus.

4. See Jen-Hung, Huang and Yi-Fen Chen 2006 Herding in Online Product Choice. Psychology & Marketing 23(5):413-428.