“I Like What I See” is a Chrome extension created by the artist Steve Klise. When activated this experimenting piece of software will click the Like bottom on every post you encounter, while scrolling down your newsfeed. Consequently, this means that you do not have to neither Like nor like anything, the “I Like What I See” application will, literally, Like what you see for you. Klise’s experiment is obviously a response to the paradoxical numbness in liking, which arguably is installed in and intensified by the Facebook like-economy; within this social site it is extremely easy to express approving emotions detached from any truly embodied affect. In fact, it only takes a click. However, Klise’s project is also shedding light on other relevant troubling aspect encompassed in the realm of interface culture, which became evident when Mat Honan through his Facebook profile liked everything he saw manually; this liking-everything-activity not only blurs the profiled user’s true interests, but, as Honan reported, it also explodes the newsfeed with strange, irrelevant and – for the greater part – commercial content (Honan).
How does this pollution of irrelevance happen? Because inherent in Facebook we find the newsfeed algorithm called Edgerank, which monitors and analyses a user’s Like, shares and comments. This in order to show the user, what he or she is most likely to be interested in – maybe especially provide the user with ads for products, he or she is most likely to buy. Thus, if everything is Liked, the algorithm will push everything in the user’s direction.
The fact that our activity online is monitored by algorithms and therefore available for analysing and tracking, is the condition we accept as users of social media sites. The premise is, as is the point of Tiziana Terranova as well Bernard Stiegler, that the asymmetrical, passive/active relationship in communicative situations between human beings (like/being liked for instance) has become objects of grammatizations; discrete, written data-entities within online forums. This means that buttons constituting our social and interactive behaviour online are “linked to underlying data structures (…) and subjected to the power of ranking algorithms”. (Terranova 394). Consequently, due to an underlying code, Like, Share, Go, Tweet, Accept, buy etc. are indexically referring back to the clicking user.
With the point of departure that monitoring and analyzing algorithms are part of our everyday communication practice, I am in the course of this paper going to investigate subjectivity in language when the subject is constituted in language plus an underlying computer code. As is the case of Facebook, when we click Like and socially express approval and at the same time are indexed and analyzed accordingly. In other words, how can we approach subjectivity in language, when we speak in language plus an underlying code, which operates simultaneously with the enunciating practice on the interface of the user?
As “I Like What I See” illustrates, even innocent Likes can cause an excessive online existence, since our information is perpetually available for algorithms owned by advertising agencies, when the click links to invisible data structures.
I am the one who says I
In order to discuss subjectivity in language, I draw on the linguist Émile Benveniste and his theory of enunciation. In his line of thinking, the unique person, the subject, is constituted within language in use – or language in discourse – and only here. In other words, it is in exercising language, we can articulate ourselves as ourselves; it is in exercising language we are able to say “I”. Thus subjectivity happens in language, when the enunciating speaker is setting himself up as such – pronouncing himself the speaker in terms of appropriating I.
To emphasize, linked to subjectivity in language, the personal pronouns become significant, especially I. In the article “The Nature of the Pronouns” Benveniste shows that by contrast to other nouns referring to a fixed object in the world, personal pronouns are instruments without a definable, identical object. And as a result, because of the particular referential organization of I “[e]ach I has its own reference and corresponds each time to a unique being who is set up as such” (“The Nature” 218). Not to be taken as a figure, but a linguistic form indicating ‘person’, I is an empty sign, always available and becomes “full” when converted from the language system to language use, from langue into discourse by a unique person in a unique time and place. Accordingly, I signifies “the person who is uttering in the present instance of discourse containing I” (“The Nature” 218). Poetically, Benveniste calls the consequence:
Since they lack material reference, they cannot be misused; since they do not assert anything, they are not subject to the condition of truth and escape all denial (“The Nature” 220)
I, by its virtue of being empty, exists not before, not after, but only within an actualization in an instance of discourse. Consequently, it is only “by identifying himself as a unique person pronouncing I that each speaker sets himself up in turn as the “subject”” (“The Nature” 220). It is when I designates a speaker the speaker becomes the subject, and as language is the instrument of communication, the speaker, implicitly or explicitly defines a you, when defining herself as I. Subjectivity, then, is not the feeling of being oneself. Instead, in this line of thinking, it emerges as a property of language; “ego” is the person who says “ego” (“Subjectivity in Language” 224). Benveniste concludes: “And so it is literally true that the basis of subjectivity is in the exercise of language” (“Subjectivity in Language” 226). Thereby he says that subjectivity is literally, when I is transformed from an element in a system to a unique designation. Consequently, Benveniste shows, that temporality and subjectivity is inseparable, the time at which one is, is the time at which one is speaking:
[…] there is no other criterion and no other expression by which to indicate “the time at which one is” except to take it as “the time at which one is speaking.” This is the eternally “present” moment, although it never relates to the same events of an “objective” chronology because it is determined for each speaker by each of the instances of discourse related to it (“Subjectivity in Language” 227)
Every time the I is employed by a speaker, the present time of the utterance and, inevitably, the presence of the speaker is introduced as well (”Le Langage et L’Expérience Humaine” 67-68). And this is how the “present” moment becomes eternal. The moment of the utterance cannot escape its own presence. The present time is always inherent in the discourse and, thus, always inherent in the constitution of subjectivity. An utterance is never repeated, it is enunciated again, every time employing this present moment, every time producing a new subject, every time a new I, here and now.
Click – enunciation online
To sum up, according to Benveniste, language as system holds voids in the shape of pronouns, and subjectivity is constituted when a speaking subject enters language, fills up the void and designates itself as the speaker in a present time discourse. I is the person uttering within the present discourse containing I, or, in other words, the subject is the I speaking here and now.
So how is this relatable to the constitution of subjectivity online? Where is the empty signifier I, which can be filled by a human subject in language plus code? Why is it relevant in relation to compulsive Liking, which subjects us to the power of algorithms? Christian Ulrik Andersen argues that the linguistics voids in a computer interface equates the buttons connecting the user with a given function (Andersen 208). Following this point, the buttons we manoeuvre with the mouse or the keyboard of the computer can be said to equate the empty signifier I as they offer themselves to be appropriated by a speaking subject. However, in the matter of subjectivity, it would be problematic and imprecise to make a complete parallelization between buttons and the linguistic I, since the computer interface already functions within the symbolic realm of language. The empty signifier I in computing seems to be several different buttons, all of them, very literally, offering themselves to be clicked by referring to their individual function in a particular situation either symbolically (i.e. search) or iconic (i.e. thumbs-up). Opposing the empty character of the personal pronoun by already entailing a referent, which is fixed in a given function, it is evident that buttons in computing are already filled and thus hold value before the instance in which they are used.
Nevertheless, for this study, like several scholars before me (Friedrich Kittler, Alexander Galloway, N. Katherine Hayles, Florian Cramer, Wendy Chun etc.) I perceive code as a sign system, a structure similar, or even parallel, to language; a sign system in which we speak and communicate (emails, status updates, comments, search queries etc.) and therefore a system, which can be actualized or realized on a semantic level. I would then argue that computational media holds the quality of being conditioned by a discrete data structure, which enables activation of language in two semantic layers, with two different semantic values of meaning; a “conscious” layer in natural language (the interface of the user) and what I will call a sub-semantic layer in the underlying code.
Following this, these parallel semantic layers must also allow for two different referents, each employing and producing meaning in their respective system. This then qualifies the postulate that the interface button – being natural language plus an underlying code – can be filled in natural language and at the same time be an empty, available signifier employed in code as a language system. Accordingly, the interface button, as for instance Like, has two referential organizations in its click-ability: the filled, fixed Like of natural language, and the empty, available I, unique in its use each time, depending on the individual click. In other words in the instance of clicking Search, Share, Like, Post, Tweet as well as Buy and Accept you are also saying I. Hence, the click is the act of enunciation, the act of transforming an empty available signifier in code as a language system into a unique designation of a speaker – just as saying I transforms the system of language into discourse.
To crystalize the point here, the click automatically prompts the coupling of natural language and computer language, since it is in the click and only here this extra language is “spoken”, converted into discourse, into use, by a speaking subject. Basically, this means that even when typing 140 characters into my Twitter-profile without pressing tweet, or creating a status update on Facebook without clicking post, while this designates me as subject of the given enunciation in natural language, it will not be an enunciation within language plus code. The code as language has not yet been actualized in the click, and so the enunciation remains on the level of natural language. With click the subject couples the two systems and is, thus, produced within both.
Furthermore, in Benveniste’s theory of enunciation it is important to clarify that subjectivity is a property of language – that is to say, subjectivity takes place only insofar as a person speaks. The subject exists solely within its own utterance, in the immediate act of the enunciation. Inevitably, as a consequence, the subject does not exist before or after the instance of discourse in which it is speaking. Following this, given that the subject is constituted in the appropriation of the I, then the click in this analysis must include a double subjectivation of the speaking subject online. Accordingly, the subject is constituted within and produced through not only language, but is in fact a property of language plus code, when the click as an empty signifier is appropriated by an (already) speaking subject.
What “I Like what I see” demonstrates is the fact that I with my Like also Like what I do not see. And what I do not see is underlying data structures such as news feed algorithms. Consequently, I become a subject of the algorithms flooding my Facebook with customized offers; algorithms producing me in not only language as a system, but also in systems mirroring my predicted consuming, political etc. intentions. Click produces an excessive me.
Andersen, Christian Ulrik: Det Æstetiske Interface, Digital Aesthetics Research Center, Aarhus: 2009 (print).
Benveniste, Émile: ”Le Langage et L’Expérience Humaine”  in Problèmes de Lingustique Générale II, Paris: Édition Gallimard, 1974 (print).
Benveniste, Émile: “The Nature of the Pronouns” [La Nature des Pronoms, 1956] translated by Mary Elisabeth Meek in Problems in General Linguistics, Miami: University of Miami Press, 1971 (print).
Benveniste, Émile: ”Subjectivity in Language” [De La Subjectivité Dans La Langage, 1958] translated by Mary Elisabeth Meek in Problems in General Linguistics, Miami: University of Miami Press, 1971 (print).
Honan, Mat: I LIKED EVERYTHING I SAW ON FACEBOOK FOR TWO DAYS. HERE’S WHAT IT DID TO ME, Wired, 08.11.14.
Stiegler, Bernard: “The Most Precious Good in the Era of Social Technologies” in
Unlike Us Reader, Institute of Network Cultures, Amsterdam: 2013 (print).
Terranova, Tiziana: “Red Stack Attack!” in Avanessian Armen & Mackay, Robin #ACCELERATE, pp. 379-399, Falmouth: Urbanomic, 2014 (print).