Self-Portrait, Selﬁe, Self: Notes on Identity and Documentation in the Digital Age

: Though the self-portrait has been hailed as the deﬁning artistic genre of modernity, there is not yet a good account of what the self-portrait actually is. This paper provides such an account through the lens of document theory and the philosophy of information. In this paper, the self-portrait is conceptualized as a kind of document, more speciﬁcally a kind of self-document, to gain insight into the phenomenon. A self-portrait is shown to be a construction, and not just a representation, of oneself. Creating a self-portrait then is a matter of bringing oneself forth over time—constructing oneself, rather than simply depicting oneself. This account provides grounds to consider whether or how the selﬁe truly is a form of self-portrait, as is often asserted. In the end, it seems that while both are technologies for self-construction, the self-portrait has the capacity for deep self-construction, whereas the selﬁe is limited to fewer aspects of the self. This prospect leads into an ethical discussion of the changing concept of identity in the digital age.


Introduction
The self-portrait has been hailed by art historians as the defining artistic genre of modernity, witness to the emergence of capitalism and liberalism. By now there is no shortage of commentary on self-portraits. However, we still do not seem to have a good account of what the self-portrait actually is. I contend that information science, as the study of documentary forms and processes, can provide such an account, which has to date been out of reach of art history and the philosophy of art. In this paper, I provide an account of the self-portrait using the analytical tools of the philosophy of information and document theory, i.e., I develop insights into the self-portrait genre by considering it as a class of document-specifically a kind of self-document, or one that provides evidence of some aspect of the self. Indeed, the self-portrait proves to be an illuminating example of self-document, as it is an early and culturally formative example of what is now a ubiquitous phenomenon [1]. Conceptualizing the self-portrait as a document, then, invites a reflection on how our understanding of the self may be changing in the digital age.
I begin by reviewing relevant aspects of document theory for its application to the question of self-portraits as documents. Next, I describe the conceptual methods used in this paper. Following that, I develop an account of the self-portrait as a document, beginning with a review of the literature on the self-portrait in art history and the philosophy of art, the two academic fields where most of the discussions on the topic have taken place. Following that, I analyze how the self-portrait expresses reference, furnishes evidence, and inheres meaning. Next, I discuss the selfie-for many, the most ready-to-hand example of self-portraiture-and consider whether selfies are necessarily self-portraits. Finally, the ramifications of these arguments are discussed, focusing on document-theoretic issues of representation and information-ethical questions of selfhood in the digital age.

Applying Document Theory to Art
On the grounds that self-portraiture is a genre of visual art, some points bear mentioning on the application of document theory to visual art. Overall, art is a little-examined domain within information science. In addition, though some research has been done on artists' information behavior [14,15], for the most part this has yet to benefit from the body of work in document theory [16]. There is an opportunity to specifically conceptualize artworks as documents, and art-making as documentation, and consider the attendant ramifications thereof. Kosciejew [16] argues that this is the case because art is both material and informational, and document theory allows for both these aspects to be investigated, whereas traditional conceptualizations in information science overlook the material.
To be sure, a few scholars have discussed how artwork can be documentary. This work comes from scholarly discussions on art documentation. In art documentation generally, Ørom [17] points out the scholarly paradigms and discourses embedded in the organization of artwork at the levels of classification systems, documentation and exhibition. Building on this, Régimbeau [18] interrogates the values laden in the classification systems we use, arguing that since the Middle Ages the classification of art objects has been subject to the head-hands bifurcation that is insufficient for dealing with contemporary art. Régimbeau [19] argues that documentalists should consider not only the art object's "content," but also critical analyses, popular discourse, historical trends and more. Such analyse documentaire is both interpretative and descriptive, and it engenders synthesis [19]. To this end, but independently of Régimbeau's work, Lopatovska [20] has begun to consider how art objects can be classified according to affect tags rather than merely their aboutness.
To speak of studies in documentation closer to human information behavior, Auslander [21], for one, problematizes the documentation of performance art by discussing what he calls "theatrical documentation": In the theatrical category, I would place a host of art works of the kind sometimes called "performed photography," ranging from Duchamp's photos of himself as Rrose Selavy to Cindy Sherman's photographs of herself in various guises . . . in which performances were staged solely to be photographed or filmed and had no meaningful prior existence as autonomous events presented to audiences. [21] (p. 2) Auslander [21] goes on to draw an account of how performance art and its documentation co-constitute each other. He seems chiefly interested in how performance art and its documentation co-constitute each other, but his work exposes something deeper: Whereas documents of performance art are traditionally or naively understood as proof of the performance that transpired, these theatrical documents firmly assert themselves as works of art in themselves. In a similar manner, Bénichou [22] discusses some examples of contemporary plastic art that blur the boundary between art and document. For instance, the "test pieces" by Eva Hesse were originally made as sculptural sketches or experiments; as such, they serve as documents for finished pieces (e.g., providing a model for preservation and restoration). However, some of these test pieces approach "finished" status, and moreover many of them are now exhibited on their own as artworks in themselves.
Auslander and Bénichou stop short of arguing that all art-making is essentially a kind of documentation. This is because they seem to conceptualize the document in a somewhat Otletian [7] way, seeing the document only as, fundamentally, directly representative of a work of art-even if, in the case of Auslander, that work may not actually exist, and in the case of Bénichou, they may be one and the same.
These views can be advanced by incorporating the neo-documentalist, semiotic perspective of the document. Latham [23] moves toward such an account in her framework for considering museum objects as documents: Museum objects are material, collected, deemed meaningful and wrapped up in cataloging and other processes [23]. Even accepting Latham's argument, it could be argued that would-be museum objects are no longer documents outside the museum setting, and so some further consideration may be necessary to secure the view that art, in general, is documentary. However, following more recent conceptualizations of the document [12,24], it is clear that the status of being a document is attributable even outside any formal institution; moreover, any document may provide evidence of any number of things. One scholar who has applied such a perspective to artwork is Walsh [25], who presents a study of religious icons as documents. Through a historical and theological discussion, Walsh explores how the icon functions as a document: The icon bundles text, image and material, and it references other icons, theological teachings and cultural events [25].
Building on Walsh's [25] work, I contend that all art can be considered a document, and all art-making documentation. Artistic works perform documentary reference in interesting ways. For example, if we consider the theatrical documents discussed by Auslander [21] to be artistic works in themselves, what they document is the artistic truth that the trust we place in documents may be misguided. If artworks are conceptualized as documents in the neo-documentalist sense, there are several implications that will have to be explored.
Of course, trivially, all artwork obviously provides evidence that it was made and therefore evinces the skill of the artist, but the document in the neo-documentalist sense is deeper: What chains of reference do artworks manifest? In what sense do they function as evidence? How do they cohere and propagate meanings? If we can answer these questions, we will have a deeper appreciation for the epistemic place of art in human life-a topic that has begun to be discussed in philosophy (e.g., [26]).

Conceptual Methods
In this paper, my overarching research question is: What sort of document is a self-portrait? My approach to answering this question is broadly phenomenological. I take phenomenology to be a method for allowing something that normally remains hidden to be revealed. This can be accomplished by beginning with everyday and uncritical understandings of the subject and then attempting to plumb deeper [27,28]. Thus, in conducting this study I began with the way self-portraits have been discussed in popular discourse, art history and the philosophy of art. Next, I used tools from phenomenological analysis, such as etymological study and the examination of phenomena similar to the target phenomenon [28,29]. For example, in their phenomenological study of screens, Introna and Ilharco [29] also contemplate mirrors; for this study of the self-portraits, I also examined the selfie. With the insights gained through this analysis, I continued to interrogate the self-portrait and selfie as documentary forms through reductive analysis [28]. I did this with reference to the document-theory concepts of reference, evidence and meaning. This analysis was based on a literature review of research in fields such as art history, philosophy and computer-mediated communication.

Self-Portrait as Document
At present, if pressed to provide a definition for self-portrait, one might come up with something like that articulated by the Oxford English Dictionary-"a self-made portrait of oneself." If honest, one realizes that this definition only turns the question into a Hydra: for what is the self, and what is a portrait? And how is of oneself to be understood?
To start with, as any number of commentators have remarked [30][31][32], the criterion of resemblance has no secure place in a definition of the self-portrait. For example, if the dozens of self-portraits of Rembrandt are considered to be faithful depictions, they would seem to depict different individuals. Moreover, artists have depicted themselves in situations that either never happened or that they could not have remembered: Michelangelo as Bartholomew's flayed skin; Caravaggio as Goliath's severed head; Frida Kahlo as an arrow-shot deer. We can better understand the self-portrait genre by considering it through the lens of document theory. As discussed above, documents (1) sit in referential systems and thereby furnish (2) evidence and (3) meaning. Thus, asking in what way the self-portrait is a document is asking about how self-portraits reference and how they furnish evidence and meaning. First, we will see how self-portraits have been conceptualized and discussed in art history and philosophy.

Self-Portraits in Art History
In art history, there have been innumerable commentaries on self-portraits and self-portraitists, but-remarkably-the self-portrait has hardly been considered to be a genre. Rather, art historians have worked with implicit assumptions about what is and is not a self-portrait; below we will see how this has been problematic. The first reflection on the self-portrait, Ludwig Goldscheider's Five Hundred Self-Portraits, first appeared in 1937 [31]. The book reproduces, as its title suggests, five hundred self-portraits, which are glossed by 50 pages of text that amount to little more than a laundry list. Still, Goldscheider does observe, crucially, that: the degree of "likeness" which this earliest of self-portraits achieves is not of much importance. It is a question of the degree of realism in the representation as a whole, depending not on ability, but on style, that is to say, on the aims of the period and the aims of the artist. [31] (p. 12) And so, from the start Goldscheider wrests the self-portrait from the question of naive resemblance, but he does not go further than that.
The subject lay fallow for several decades. In the 1980s, Sean Kelly, Edward Lucie-Smith and others hatched an idea that became the 1987 U.K. exhibition The Self-Portrait: A Modern View, in which dozens of contemporary British artists were invited to create a self-portrait to exhibit in a group show. The exhibit's catalogue demonstrates a breadth of styles and approaches to the genre [33]. Contra Goldscheider, each self-portrait in the book is accompanied by a photograph of the artist, inviting a back-and-forth comparative glance. Since then, three notable works have discussed self-portraiture as an art-historical genre: Joanna Woods-Marsden's study of the emergence and meaning of autonomous self-portraiture in the Renaissance [34]; Laura Cumming's art-criticism discussion [30]; and James Hall's cultural history [32].
All three shed insight on the historical circumstances that accompanied the creation of self-portraits. Notwithstanding an example from ancient Egypt, another from Greece and a handful from the late Middle Ages, the self-portrait is said to have emerged early in the Italian Renaissance. The first self-portraits, including those from Antiquity, depicted the artist within larger, typically religious works, functioning as a sort of signature or testament to the artist's skill. As the 15th century unfolded, the genre of autonomous (standalone) self-portraits emerged, soon burgeoning in northern Europe as well. Woods-Marsden writes of this emergence as artists' asserting their work as an intellectual activity-not just a manual craft-and, concomitantly, themselves as dignified members of society. This came along with the development of humanism, harnessing the classical dictum that "man is the measure of all things," and the liberal concept of the individual. Hall interestingly links this development to the availability of mirrors, and the reception and provenance of particular self-portraits.
Since then, many well-known artists carried on the genre, such as Albrecht Dürer (1471-1528) and Rembrandt van Rijn (1606-1669). It is known that Rembrandt's were used educationally, copied by his students. There is no evidence that self-portraits were ever commissioned; some, however, were given as gifts in exchange for favors. For example, Montaigne [35] (p. 101) writes that the king of Sicily presented the king of France with a self-portrait. Beyond this, the precise motivations for most self-portraits are not known. Art historians have suggested that first and foremost, self-portraiture is a matter of convenience-oneself is always the subject closest at hand-but it may also be a conscious effort to continually reestablish the myth of the artistic genius. In any case, self-portraiture flourished in the 19th and 20th centuries, with many artists becoming known specifically for their self-portraits, such as Vincent van Gogh, Egon Schiele, Edvard Munch, Frida Kahlo and Pablo Picasso.
Today, virtually all artists make self-portraits at one time or another, and the self-portrait is said to be the defining artistic genre of our age. In addition, indeed, the genre has continued to evolve; one of Hall's final examples is the 1995 Everyone I Have Ever Slept With by Tracey Emin, which takes the form of an appliquéd tent without any human depiction whatsoever.
As most of us are well aware, the widespread availability of camera-and internet-equipped smartphones has given rise to the selfie, which the Oxford English Dictionary defines as a kind of self-portrait. Indeed, for many, the selfie is the first thing that comes to mind today when they hear the word self-portrait. Some seem to equate the two. For example, Rembrandt's self-portraits have been described as "early" selfies [36]. More recently, the Philadelphia Office of Arts, Culture and the Creative Economy presented the exhibit Veterans Empowered Through Art: The Six Week Selfie Project, which was the culmination of museum tours and workshops and included sketches, complete self-portraits, poetry and personal photos [37]. Others, somewhat differently, see the selfie as the most recent manifestation of the self-portrait genre. In the view of Mirzoeff [38] (p. 31), the selfie "expresses, develops, expands and intensifies the long history of the self-portrait." Mirzoeff sees the selfie as a digital, networked outgrowth of this artistic genre. However, what is the self-portrait genre? For all the art-historical work recapitulated above, scholars have not yet answered that question. As a result, art history does not seem well-equipped to discuss challenges to the genre of the self-portrait, which have become more commonplace over the past century, such as Emin's tent self-portrait and the prolific selfie. Another example is the work of Cindy Sherman, an American artist whose oeuvre consists of photographs of herself in various guises. Art historians seem to regard these as self-portraits, as demonstrated in the collection of art-historical and -critical essays [39] and scholarly essays [40]. Still, Sherman emphatically claims that her works are not self-portraits. As she explained in a New York Times interview: "I feel I'm anonymous in my work.
When I look at the pictures, I never see myself; they aren't self-portraits. Sometimes I disappear" [41].
How can this be the case? Are Sherman's works self-portraits or not? And Tracey's tent? Are selfies self-portraits? These questions are growing in urgency as the idea of the self-portrait becomes bent and addled. To answer them, we must turn from the ontic to the ontological.

Self-Portraits in the Philosophy of Art
If art history has not questioned the self-portrait, what about philosophy? In the past decade, a few philosophers have examined the portrait, and within these strands are some insights on the self-portrait. Given that the self-portrait is a type of portrait, philosophical accounts of the portrait bear mentioning.
Freeland [42], in Portraits and Persons, develops three necessary and sufficient conditions for a portrait: first, it must show a recognizable physical body; second, it must show a sense of inner life; third, the subject must pose for the artist. Freeland encapsulates these conditions in defining the portrait as "an image that presents a recognizably distinct individual who has emotional or conscious states, and who is able to participate in the creative process by posing" (p. 284). Thus, for Freeland, sketches made in a life drawing session, people in commercials and CCTV footage do not qualify as portraits. Paolo Spinicci [43], building on a different literature and line of argumentation-both phenomenological-largely agrees, except that he says a portrait need not be posed but can merely look as if it is posed. Thus, whereas for Freeland there can be no such thing as an animal portrait (assuming that animals cannot pose), for Spinicci this is no problem. Moreover, Spinicci suggests that we should identify portraits based on how they are used (e.g., as a stimulus for remembering a loved one), rather than merely their formal aspects. Still, for Spinicci the possibility for use does seem to rely on some resemblance; just as the shape of a screwdriver makes certain uses (i.e., in certain types of screws) possible or impossible, Spinicci says a portrait must be suitable perceptually and/or imaginatively.
Maes [44] responds to these accounts. First, he argues that Freeland's account is overly restrictive. Agreeing with Spinicci, he says that not all portraits are posed, such as the famous image of Che Guevara. Maes also questions her "inner life" condition, arguing that deathbed portraits (which Spinicci uses as an example) cannot show inner life because there is none, and nor can those portraits not showing the face. We find an example of the latter in Goldscheider's book: Wilhelm Busch's depiction of his own hand. Finally, Maes points out examples of very abstract portraits that do not seem to depict any part of a recognizable human body and yet still qualify as portraits. Furthermore, Maes suggests that the pose may not be enough to qualify a portrait. He prods at Freeland's example of a tourist's snapshot in front of the Eiffel Tower, taking the position that such an image is not a portrait, though an individual is recognizable, as is their inner life, and the photo is certainly posed. Responding to Spinicci's point about usage, Maes gives the example of a bust thought to be a portrait of an ancient philosopher and which was used as such for centuries, but was recently discovered to be a general depiction of old age. In Maes' view, it should stop being regarded as a portrait on Spinicci's account, and yet that has not happened.
Based on these criticisms, Maes proposes his own account of the portrait. He says that an artwork may be considered a portrait if, first, the artist intended to create a portrait and realized that intention successfully. This requires that the artist have a concept of the nature of portraits which matches that held by a group of prior portrait painters and that the artist seek to realize that concept in their work. Maes argues that necessary and sufficient conditions need not be specified, and that this account can incorporate other philosophers' accounts felicitously. He offers several "portrait-relevant features" that may be part of the concept of the portrait, including the revelation of the subjects looks, inner life and/or social standing. Applying these insights to the self-portrait, we can see why a self-portrait must do more than simply convey the external appearance of the artist; it must show the subject's inner life, which may include how the artist felt, what they wanted to show, how they wanted to see and be seen, their struggles regarding fame, their impressions of their work and their thoughts about mortality. As Cumming writes, these concerns may indeed trump that of external appearance: they may be less true to appearance than portraits. However, they are not just portraits, for all that art history often treats them as a subset; and they often specialize in other kinds of truth. Artists have portrayed themselves, improbably, as wounded, starving or unconscious beneath a tree, as a baby being born or a severed head dripping blood, as younger or older or even of the opposite sex. . . . However, no matter how fanciful, flattering or deceitful the image, it will always reveal something deep and incontrovertible (and distinct from a portrait) . . . the truth of how the artist hoped to be seen and known, how he wished to represent (and see) himself. [30] (pp. [4][5] Following Maes, we may posit that self-portraitists seek to realize a concept of the self-portrait that they hold in accord with a tradition of prior self-portraitists. The self-portrait presents the appearance and/or inner life of the artist through a depicted pose. This explains why Cindy Sherman's works are not self-portraits: they do not present her inner life, but rather those of her invented characters. And insomuch as her characters are not recognizable individuals-being, in some sense, her anonymous doppelgängers, who only resemble her by coincidence-it is doubtful whether they should even be read as portraits. Of these philosophers of the portrait, Freeland does devote some consideration specifically to the self-portrait, though chiefly as a means to critique the narrative theory of selfhood. On her analysis, narratives entail a distinction between narrator and product, clear causal unity among the events and narrative closure. Freeland surveys a handful of serial self-portraitists, arguing that these self-portraits do not form narratives in that sense. While there are examples of narrative art (e.g., biblical and mythological), Freeland finds portraits, including self-portraits, to be decisively non-narrative. Rather, they seem to be a way for artists to make themselves directly present to far-flung observers. Freeland writes that self-portraits are: visual artefacts that are made in order to draw our attention to the depicted person as a subject with his or her own intentionality; the artefact itself thus manifests two distinct sorts of purposes (both intentional), that of the creator and that of the subject. We can appreciate both at once, in a complex grasp of the meaningfulness of other people's actions and awareness. (p. 174) In critiquing the narrative theory of the self, Freeland seems to sense that there are problems with separating the self-portraitist from the self-portrait (as this is unlike separating the author from the narrative), and she notes as well that lived humanity is not always a matter of clear causal chains and it is certainly not closed down. However, Freeland writes that the self-portrait is a matter of an artist's conveying "what they see and what they want to project about an inner self" (p. 143). However, is there such an "inner self" to be conveyed? Is that the correct, or even a useful, way to think about the self-portrait?

Self-Portraits through the Lens of Document Theory
Further insight into the self-portrait can be gained by considering it as a document. Doing so will show how it sits within systems of reference and how it furnishes evidence and meaning.

Self-Portraits in Systems of Reference
As theorized since Otlet [7], a document is part of a system of reference. Considering a self-portrait as a document means, first, regarding it as something situated within a system of reference. Information science has generally been concerned with information rooted in particular technologies, such as in books or on computers, but systems of reference can also be conceptualizations of broader practices [45], such as religious iconography [25]. Indeed, a document can be said to be situated within any number of systems of reference, depending on which aspects of it are deemed relevant for the purpose at hand.
In memory institutions, self-portraits are described according to controlled vocabularies along with other cultural objects along dimensions such as geography, physical characteristics and subject matter [46]. Memory institutions' collections are often heterogeneous, comprising far more than merely self-portraits (or even just portraits), and so these vocabularies (e.g., the Categories for the Description of Works of Art developed by the J. Paul Getty Trust) are not specific to self-portraiture [47]. Thus, they situate self-portraits within a broad system of cultural organization (including other artworks as well as historical periods, geographic locations, etc.).
Self-portraits have been classified specifically in art history. One mode of such classification is according to the format of the portrait. There are several conventional portrait formats: profile (the face depicted from the side), three-quarter view (the face depicted on an angle), half-length (or bust, where the head and shoulders are depicted, often straight-on), and full-length (the whole body depicted) [48]. Next, self-portraits have been classified according to the number of people in the portrait. Self-portraits depicting only one person (the artist) are referred to as autonomous [34] or independent [32] self-portraits. Those depicting more than one person comprise two categories: First, face-in-the-crowd [34] or bystander [32] self-portraits depict the artist in a minor position amid a larger scene, often of a religious nature, while multiple self-portraits depict the artist alongside other people, such as a spouse or other family members [32]. Lastly, self-portraits may be categorized according to the activities depicted. Artists may depict themselves at work as artists (often showing their implements, canvas, hands or mirror), or simply posing (possibly in costume) [32,34]. It may be noted that contemporary self-portraits sometimes defy these traditional classification schemes, such as those that depict the artist in abstract shapes and fields of color [32].

Evidence in the Self-Portrait
Next, considering a self-portrait as a document means considering how it provides evidence, and of what. Self-portraits could be said to provide evidence of many things, including the artist's technical abilities and the materials that were available in a particular setting. Here, however, I will focus on how self-portraits provide evidence of the self, seeing as the capacity to provide this kind of evidence is what distinguishes the self-portrait from other artistic genres.
Prima facie, one might assume that a self-portrait provides evidence of the appearance of an artist. However, as noted above, this is dubious. As Freeland [42] writes, any portrait must do more than simply convey the external appearance of a person; it must show the subject's inner life. In the case of the self-portrait, this "inner life" may include how the person felt, what they wanted to show, how they wanted to see and be seen, their struggles regarding fame, their impressions of their work and their thoughts about mortality [42].
This sort of evidence can be understood through the framework developed by Goodman [49] for the philosophical understanding of art. Goodman begins with the premise that we use symbols in perceiving, understanding and constructing the worlds of our experience. He views artworks as entities composed of symbols, which sit in relation to other symbols in the world; thus, they classify aspects of reality for us, as do such things as scientific theories and what makes up common, ordinary knowledge. In this way, artwork requires interpretation, which amounts to understanding how artworks perform reference. Art does not provide representation in the sense of resemblance; rather, it references through denotation and/or exemplification. Through denotation, art constitutes a pictorial label for its subject (e.g., The Mona Lisa denotes the sitter); through exemplification, art is a sample or has particular aspects of some entity or concept (e.g., The Mona Lisa exemplifies intrigue). It is the notion of exemplification that explains how abstract and non-objective art, such as instrumental music, can be meaningful even though they do not represent or denote anything that can be described through propositional statements [49]. As a symbol system-Goodman hesitates to call it a "language" (cf. p. xii)-a work of art is so semantically and syntactically dense and symbolically replete as to defy clean analysis; rather, in art, "familiarity is never complete and final" [49] (p. 260), which is what makes the work of art a site for questioning and the building of understanding.

Meaning of the Self-Portrait
Last, documents have meaning [2]. What does a self-portrait mean? One route to considering the meaning of human practices is through myth [28,50]. In popular culture, the self-portrait is sometimes linked to the myth of Narcissus. For example, a 2010-2011 exhibition of self-portraits at the Pennsylvania Academy of Fine Arts was titled Narcissus in the Studio. In Greek mythology (one rendition is given by Graves [51] (p. 276)), Narcissus was a youthful hunter known for his beauty. One day a vengeful goddess lured him to a pond where he saw his own reflection and fell in love with it. Narcissus was unable to look away, and he eventually died there staring at the image of himself, and in his place sprung up the Narcissus flower. Connecting the work of artists (not limited to self-portraitists) to the Narcissus myth goes back to the 15th century, when the artist and writer Leon Battista Alberti presented this myth as the origin of painting in On Painting [52]. Alberti makes this connection because Narcissus was turned into a flower and "painting is the flower of all the arts" [52] (p. 25). To be sure, linking Narcissus to self-portraiture specifically is more immediately defensible. However, does the meaning of the self-portrait stop at narcissism? Already we can answer in the negative: Recall the discussion above on self-portraiture as expressing the changing role of the artist in society, manifesting the new philosophy of humanism, and positioning artwork as the union of intellect and labor [32,34].
Further insight still into the meaning of the self-portrait can be gleaned through etymological analysis, which is a mainstay of phenomenological inquiry [28]. According to the Oxford English Dictionary, self-portrait originated in English in 1831 as a transliteration of the German Selbstbildnis or Selbstporträt. Selbst is reflexive (meaning it refers to the agent of the phrase) which comes from the Old High German word for master [53]. The word Bildnis refers to a form, figure, effigy, painted likeness or example [54], while the word Porträt comes from the French portrait.
Though self-portrait came to English through German, the English word portrait came directly from French [55]. The French portrait comes from the Latin verb protrahere [55]. The prefix pro-(a variant of por-) means forth, forward or outward [56], while trahere (from which come also the English words train and tract) means to drag something slowly or to draw something out in the passing of time, resulting in some consequence. Thus, protrahere has the sense of drawing forth, pulling out, prolonging or bringing to light [57]. (This last sense is not unlike illustrate, which etymologically means to bring to light.) The phrase draw something out has at least two meanings. On one hand, in its primary meaning, it signifies an effortful pulling, as in drawing a bucket up from a well or drawing a shy person out of reclusion. On the other hand, it could mean using a pen to make marks on paper, as in drawing out a schematic. The former seems to be the original meaning of the word, having entered Old English as dragan (think of the modern drag). In the 14th century, the word came to be used in the sense of creating something with a pencil or pen, constructing a figure or document (OED sense 59). Drawing in the sense of creating a resemblance of something else did not emerge until the 16th century.
So, we are invited to think of today's drawing a picture in light of this history: it is not a question of mechanical duplication, but of realizing something that was latent in the world. Likewise, we should recognize that the portrait, in its primordial sense, is not a matter of depiction or representation. Rather, it is a matter of revealing, of bringing something into being. Key elements here are time and effort.
If that is the nature of the portrait, then the meaning of the self -portrait has to do with bringing oneself forth over time. It is not that there is a preexisting self that is being duplicated or even represented on paper, but that the self is being constructed, coaxed into existence, drawn out, as the self-portrait unfolds.
This would seem to suggest that whether a work of art is a self-portrait or not comes down to what the artist experienced in making it. I agree with Maes, then, that the artist's intention does matter-but it is not just that the artist intends to create an object that looks like a frozen mirror; rather, that the artist intends to create their self. So, the success of a self-portrait is not a question of verisimilitude, but of the quality of the artist's personal experience.

Are Selfies Self-Portraits?
As mentioned above, many commentators link the selfie to the self-portrait. For the most part, this may be a matter of naive conceptual confusion or commercial appeals to buzzwords, as in the Philadelphia Six Week Selfie Project and articles hailing Rembrandt and Parmigianino as early-modern selfie tastemakers. However, in a few cases, the link is earnest and examined. Mirzoeff and Rettberg, for instance, do maintain that the selfie is a manifestation of the self-portrait genre [38,58]. Others moderate this, saying that even if this is the case, the two cannot simply be equated. For example, Saltz [59] argues that selfies and self-portraits are sufficiently different because of the necessary skill and training involved. As it turns out, even what does and does not count as a selfie is unclear and is subject to debate in popular forums [60].
Given this confusion and disagreement, it is worth spending some time conceptualizing the selfie, which I will do here. At the very least, studying phenomena that are similar to the target phenomenon can shed light on the nature of the target phenomenon; sometimes this is simply because the boundaries between the phenomena are clarified, and other times this is because studying a similar phenomenon brings to light some aspect of the target phenomenon that had gone unnoticed. Thus, a discussion of the selfie can, at the very least, serve to ground a deeper understanding of the self-portrait. As we will see, this discussion also provokes a deeper reflection on the question of identity in the digital age.
The selfie is defined in the Oxford English Dictionary as a photograph that a person takes of themselves, generally with a smartphone, which is then shared with others online [61]. As this suggests, selfies are ways of documenting social facts, such as showing someone else one's having been in a particular place with particular people. To this definition accounts have added that selfies capture spontaneous and casual moments which are shared immediately, and as such they promote a focus on the present [62,63]. The selfie developed alongside the smartphone, which integrates a camera with web sharing capabilities [62]. Visually, selfies are distinct for some of their formal aspects: They are taken with a wide-angle lens, often with the photographer-subject's arm showing and at angles that belie an amateurish composition [62][63][64]. Indeed, part of the selfie's proliferation may be that it does not require technical skills [64].
The selfie generally depicts the photographer-subject's body and surroundings. Wendt [65] discusses how the selfie limits itself to the photographer-subject's external appearance, minimizing the non-visual aspects of the person; this contrasts with the accounts of the portrait given above, in which the expression of inner life is a central aspect. This is perhaps most clearly the case with so-called medical selfies, which have been used to track progress on burn healing and other health issues [66]. Still, Qiu and colleagues [67] found that certain personality traits, such as agreeableness and neuroticism, can be ascertained through viewing a selfie. Zhao and colleagues [68] offer a differing account, arguing that the most salient aspect of the selfie is not the representation of the self, but rather the image's capacity for evoking different perspectives (i.e., emphasizing intersubjectivity), inviting onlookers to look not at the photographer-subject necessarily, but from the photographer-subject's point of view; this seems consonant with Freeland's [42] (p. 174) assertion that the self-portrait draws attention to the depicted person as an intentional subject, discussed above. Similarly, Donnachie [69] discusses that selfies serve as a way for people to greet each other and thereby serve as a path for encountering the other.
Moving the focus away from the selfie's content, Frosh [70] has argued that one cannot recognize a selfie just by looking at what it represents. Rather, Frosh says that recognizing an image as a selfie requires people to "make inferences about the nondepictive technocultural conditions in which the image was made" [70] (p. 1608) and to have been socialized into reading such images. Thus, the selfie's nature as a social and digital artefact must also be considered. Indeed, both Levin [71] and Mirzoeff [38] suggest that the defining aspect of the selfie is its networked nature. Mirzoeff calls the selfie a form of "predominantly visual conversation" [38] (p. 63), emphasizing its communicative capacity. More precisely, Frosh [70] describes the selfie as a form of phatic communication-that is, communication whose primary purpose is the production, expression, and maintenance of sociability and wherein the denotative meanings of the words (or, in this case, images) are not important. Rubinstein [72] develops this point further, arguing that we should not think of selfies as referencing their subjects but as expressing the networks within which they sit. Rendering this in more detail, Wendt [73] mentions that selfies do not stand alone, but inherently include metadata, including hashtags and timestamps. So, selfies are seen not merely as images, but as embodiments of relationships. On this view, the uniqueness of the selfie is that it is self-same across many iterations on various screens and devices, wherein the self is articulated anew in each image, and thus each selfie is also self-similar to other selfies [72].
Along this vein is what Rettberg [58] calls the selfie's serial nature. Producing a selfie generally involves a person taking multiple photos until they are satisfied [58,64]. Because of this, Lüders sees selfie-making as a "reflexive processes of visual self-authoring" [64] (p. 947), while Rettberg [58] suggests that this corresponds with one's never fully being able to capture or convey what you want to about yourself. Similarly, several scholars of the selfie describe the selfie as a site for self-authoring, not only for challenging existing structures but for realizing one's "true self" through "authentic" expression [74][75][76][77][78]. However, somewhat less Panglossian, Wendt finds the selfie to be symptomatic of an endless quest for the ideal self: "As if we are unable to understand our being-in-the-world, we become accustomed to our being-in-the-image" [65] (p. 45).
So, where does that leave us in our comparison of the selfie to the self-portrait? Though the two may have seemed quite similar or even coextensive at first blush, it now appears that they are really quite different. First, self-portraits can be in any medium, but selfies can only be photographic-and generally only taken on smartphones. Both forms of representation depict their creators, but the self-portrait seems to emphasize the creator's inner life, whereas the selfie emphasizes the outer life: self-portraits manifest meditations on, for instance, possibility and death, while selfies limit themselves to the immediate environment. Self-portraits are generally kept to oneself, while selfies are virtually always shared. Self-portraits are singular, while selfies are multiple. Selfies are a form of phatic communication; to the extent that self-portraits are communicative, they constitute substantive communication. Self-portraits are made to last, but selfies are for now.
More deeply, the self-portrait and the selfie manifest differing epistemologies. The selfie seems rooted in a correspondence theory of truth-a positivism in the manner of the early Wittgenstein-while the self-portrait manifests something more along the lines of Heidegger's aletheia, truth as unconcealment. We can understand this in terms of Heidegger's argument in "The Question Concerning Technology" [79]; whereas the selfie manifests standing-reserve (a control-based, representational, calculating way of thinking), the self-portrait exemplifies poiesis (an open, poetic revealing and belonging).
In light of the etymological exploration recounted above, we can also appreciate the gravity of the fact that self-portraits are made over a long period of time and with skillful effort, whereas selfies are nigh-instantaneous and celebrate maladroitness. On one hand, this could be interpreted as part of the ultimate democratization of self-portraiture. Whereas the fine-arts practice of self-portraiture is the purview of an elite few, requiring technical and aesthetic skills as well as time for leisure and a certain level of wealth, taking selfies is a practice available to a vast number of people (digital divide notwithstanding). However, on the other hand, it may be that the very aspects of self-portraiture that were worth venerating have disappeared from the democratic selfie.
Perhaps this can be better understood through an etymological metaphor. Given that the portrait part of self-portrait is what carries the meaning of time and effort, it is striking this is precisely what has been excised in the transition from self-portrait to selfie. As this analogy implies, effort, drawing out and time are present in self-portraiture but not selfie-making. The selfie leaves us with only an instantaneous and present-limited-perhaps rightly rendered in the diminutive-obsession with the self.
That being the case, it is instructive to consider how the self-portrait and the selfie differ in terms of the construction of the self. Based on the above reviews, it seems that discourse about both seems to posit an already-present self being depicted. More accurately, both forms of self-documentation can be seen as forms of construction-not just depiction-of the self. However, the self-construction of the self-portrait is much deeper than that of the selfie. Celebrations of the power of the selfie for self-authoring and authenticity notwithstanding, it does not seem that selfies offer the same opportunity for deep reflection of one's inner life, integration of the past and future, etc., that are so inextricable from the self-portrait. Moreover, whereas the self-portrait as a finished product draws in onlookers as an intentional subject, the selfie is emphasized as a node in a network rather and not a being in itself. In addition, as Rubinstein [72] wrote, each selfie is self-similar to every other selfie. So, does this not, in the end, efface the self? If selfies really are self ies, as their name suggests, then they seem to propound a very particular understanding of the self: as something networked and indistinct. This issue will be taken up below in Section 4.2.

Implications for Theories of Representation and Description
This study has explored how self-portraits are documents. This has sometimes been implied [80][81][82], but it has not yet been explored deeply. In this article, I have discussed how reference, meaning and evidence manifest in the self-portrait; comparing these features of the selfie served to sharpen these findings. This work contributes to document theory particularly in considering the questions of representation and temporality.
A key issue of interest in self-portraiture is that of representation. As we saw, the Oxford English Dictionary defines the word self-portrait as a "portrait of oneself" [83]. In an Otletian [7] theory of representation, this "of oneself" is a matter of visual similarity between the person and the object. This would seem to be the predominant interpretation: In general, people expect artworks to look like the things depicted, and a self-portrait thus is expected to look like the artist. In other words, this is artwork as mimesis, a Platonic or Aristotelian theory of art [84].
However, this is not the only way in which "of oneself" can be understood. For example, the self can be represented in terms of style rather than, or in addition to, content. Thus, "of oneself" can be understood not only as depiction but also in the sense that a self-portrait comes from oneself. Indeed, the etymology of self-portrait testifies to this fact. In this sense, representation is understood in Briet's [8] terms, as indexical reference (pointing by association), rather than Otlet's [7] sense of representation as depiction.
Whereas both these senses of representation are at play in the genre of the self-portrait, discourse on the selfie focuses on representation in Otlet's [7] sense. In the selfie, the self only seems to go as far as what can be seen. To be sure, some researchers have discussed how selfies perform indexical reference, but the referents are not the self nor aspects thereof, but rather other people, objects, events and ideas. It may be the case that self-portraits reference the self in both ways and selfies only in one, or it may be that the question has not yet been asked of selfies. It seems possible that particular individuals could become known (even in small social circles) for their style of selfie, though the literature on selfies does not show it. Along these lines, Lim [85] suggests that further research on the selfie should examine the extent to which aesthetic appreciation matters in the selfie phenomenon.
Beyond a deeper understanding of these genres, what does this mean for practitioners? In its origin, documentation entailed the description of documents to facilitate their retrieval within formal memory institutions [4,7]. In recent decades the scope of documentation has broadened to recognize far more documentary forms and sites, but description is still a central part of documentation and information science. For the description of art, guidelines exist such as those published by the American Library Association [47] that offer a framework for describing both the form (e.g., measurements and materials/techniques) and content of a work of art. In the self-portrait, the two epistemic senses of "of oneself" blur the distinction between form and content, as the form is also content. Recent work has been exploring the prospect of developing affect tags for art-ways to describe art in terms of emotional expression rather than just representative depiction-but these focus only on the content [20]. My study suggests that the form may also be important for such description, at least in some artistic genres (the self-portrait being one of them).

Implications for Identity
This work also has implications for our understanding of personal identity or self (here I use these words interchangeably). How the self is understood changes over time and place [86,87]. Considering our own context, Capurro, Eldred and Nagel [88] write: In the age of the Internet the question concerning whoness is posed anew because the ways of being in time and in space that characterize human being, along with togetherness in the digital medium of the cyberworld, are going through hitherto scarcely imaginable reshaping and recasting.
How can we understand what it means to be a self today? As Capurro and colleagues point out, this is tantamount to asking: What options are there for existing today? How are contemporary technologies changing the ways we can reveal and conceal our selves? While a complete analysis of these questions is beyond the scope of this paper, a few points of discussion bear mentioning.
As discussed above, art historians have shown how the self-portrait has, since its inception, been a harbinger of new conceptualizations of the self. Then, insomuch as the selfie is advocated uncritically as the current, de facto form of the self-portrait genre, it would seem to signal a new understanding of the self. This is demonstrated in the documentary characteristics of the selfie, which express a networked, data-driven view of the self. Others have written as well on how personal identity may be changing in the digital age, for example being more wrapped up in products and brands via social media [13,[88][89][90][91][92]. A recent feature in Wired points out that this may particularly be the case with the present generation of teens, who "view their identity as a curated composition" [93] (p. 260).
However, there seems to be some tension in this emerging sense of self. In brief, emphasizing the digitally networked nature of the self seems to be at odds with the previous idea of the self as a discrete, coherent entity [90]. Furthermore, there is some evidence that the newer conceptualization of the self undermines psychological well-being [94,95]. Naturally, people experiencing tension in their self-concept seek to resolve that tension. Anderson [89] (p. 12) sees this as "precisely what lies behind the current trend of self-documentation online . . . a reassertion of the visible self as a gesture of defiance at having one's identity reduced to abstract metadata." However, perhaps as long as these reactions happen in digital systems, they cannot escape the problem [96], as computing platforms are already arranged according to the logic of networking.
Self-portraiture holds potential as an effective reaction against the pressures of the emerging, digital-technological sense of the self. As an art form, it allows great freedom of expression; in the parlance of Goodman [49], self-portraits are syntactically dense and semantically replete (i.e., they offer a bigger vocabularies and syntaxes). Given this capacity, there are endless ways to interpret a work of art. In addition, every time one returns to a piece of art, they come with a new situation-with different things on their mind, etc. As Elgin [97] (p. 287) writes, "The picture is inexhaustible. There is always more to be found." The question is whether self-portraits, in the sense of this potential, can exist online. More to the point: whether selfies can be more like self-portraits-whether they can be artful. It would seem that the answer is yes, at least in some cases [98]. Indeed, there are surely examples in which selfies are personally meaningful, and perhaps there are cases where they involve expending effort, drawing out and taking time. However, what distinguishes these cases from those where selfies are, to quote one of my teenage sisters, "just something everyone does"?
As we move further into the digital age, we must work to better understand the space of possibility for free self-expression, including the current realities of personal and social practice, both with digital technology and without. "Only when we are able to see the issues clearly are we in a position to assess also which practices and customs are to be valued and which are to be kept out of a cultural way of life" [88] (p. 29).

Conclusions
In this paper, I have conceptualized the self-portrait as a kind of document. In self-portraiture, a person draws out their self; making a self-portrait is an act of self-creation just as much as self-representation. Of course, today we have myriad ways to self-create through documentation, and thus self-portraiture becomes an ethical question: Which forms of self-documentation contribute to the flourishing of selves?
As Hepworth and colleagues [99] have argued, phenomenological findings are useful for identifying tacit and implicit dimensions that should feed into the design of future information systems. Recall that social computing has been implicated in the creation of a new, troublesome concept of self-identity. As social computing as proliferated and we rely on it to function in contemporary society, any way forward must take place within the rubric of social computing. To speak of social networking, for instance, we might imagine a platform where users do not identify themselves in terms of types and groups, but rather as unfolding, ever-in-progress, multivalent individuals. However, is this possible, either logically or practically? Answering this question is a vital task for the future of humanity.