Literature: Picture, Language Game, Picture Again
Literature: Picture, Language Game, Picture Again


The paper discusses possible analyses of literature in the framework of Wittgenstein’s views on language. The analysis of fiction as picture in the Tractarian sense, though interesting and possibly fruitful, is not without problems. The analysis of it as language game presents it as a defective, private-language-like type of game. However, picturing can be understood as certain function of language games, expressed especially strongly particularly in “literary” utterances.

Table of contents

    This paper discusses possible analyses of literature in the framework of Wittgenstein’s views on language. The analysis of fiction as picture in the Tractarian sense, though interesting and possibly fruitful, is not without problems. The analysis of it as language game presents it as a defective, private-language-like type of game. However, picturing can be understood as certain function of language games, expressed especially strongly just in “literary” utterances.


    The culmination point of Ludwig Wittgenstein’s early philosophical phase, his Tractatus logico-philosophicus (Wittgenstein 1921), provides a consequent theory of language. The language Tractatus talks about is language consisting of “propositions” (Sätze) depicting the world. A proposition is something that is – if not nonsensical – either true or false. True propositions are those that correspond to (mirror, reflect) an actual fact (Tatsache, matter of fact) in the world. False ones are those that correspond to a fact, which is not actual (actually present in the world).

    However, several serious objections have been raised to this conception. I will confine myself only to some more specific points. The sketched notion of propositions and truth is rather narrow. Wittgenstein never provides an example of what exactly he considers to be a proposition. Most probably, he talks about certain ideal of language – a language consisting only of what can be “said” (which means, what can be said clearly – which means, what can be said in the manner science talks about things). The language is declarative, a-temporal, im-personal, only factual, etc. Most of the language we know as language isnot considered as language here. However, if as far as language is to be understood as an isomorphic system reflecting the world, most of what we know as language could qualify – though perhaps only more or less – as such. Even in very “everyday” talk we make utterances/utter sentences (“Bob, you’re damn’ stupid!”) that can be meaningfully thought of as either agree with reality (the world) or not (true or false in virtue of that).

    But the situation is not that simple: how is – for instance – a negative true proposition true? Does it correspond to a “negative fact”? Or to the sum of all facts (which don’t include the required “positive” fact)? The case of negative true proposition is probably the most painful for the Tractarian system, because it surely should somehow deal with them. But there are also others. Propositions about the past or the future must be either analyzed as a present-tense-type proposition with time coordinate added; or stated to be no genuine propositions. A similar problem concerns indexical utterances (“I am a human being.”).

    A particular difficulty is represented by literature. There are no doubt both true and false propositions concerning the worlds of fiction: such as “Sherlock Holmes was a detective” versus “Sherlock Holmes was a criminal villain”. Can the former be denied to be true and/or the latter to be false? That is, can they be considered as equal as to their truth-value (truthfulness)? (Actually, to be no meaningful – neither true, nor false – propositions at all?) That doesn’t seem correct. Hence, in virtue of what corresponding fact is the proposition “Sherlock Holmes was a detective” true?

    It cannot be a fact in the straightforward “scientific” sense: an actual constellation of objects within the only, real and objective world. Moreover, whatever fact it is, it must be the one which the proposition “Sherlock Holmes was a detective” is a picture of. (Let’s confine here to those sentences in fiction that can be thought of as true or false (having truth-value). Of course, not all the sentences in literary works are of this nature – are about something, make us “see”: Sherlock Holmes for instance.) As far as “Sherlock Holmes was a detective” is meaningful and true, it has a logical form which it shares with the “fact” (?) it pictures. (Certainly, the notion of shared “logical form” doesn’t mean much more than the close relation binding the picturing and the pictured.) Perhaps, there is a candidate to being the “world” or the “fact” with respect to which the proposition – as its logical picture – can be true: it could be the Opera by sir Arthur Conan Doyle. Whoever reads through the four novels and fifty-six short stories, cannot doubt that the proposition is true, because Sherlock Holmes really was a detective (and not a criminal villain). Hence, the proposition can be understood as sharing the logical form with the Holmesian canon (as a true logical picture of one fact constitutive of the respective fictional world, contained in Doyle’s books).

    This world is of course much less solid than the world in more common sense. And the solidity can even decrease, depending on the degree of canonicity of the reference “world”. This is an interesting problem is the case of Sherlock Holmes. In his short story “A Study in Emerald”, Neil Gaiman presents an alternative Lovecraftian world, where Sherlock Holmes (supported by Doctor Watson, of course) can be understood – at least according to the ethical standards of this alternative reality – as a criminal villain. (And – both logically and paradoxically – Professor Moriarty and Colonel Moran are detectives going on Holmes’ track.) Both Doyle’s “Study” (in Scarlet) and Gaiman’s “Study” are intended to be read as portraying (picturing) the same personality – can they be both true logical pictures of Holmes’ character? We can perhaps refer to Doyle’s writings as the deciding authority (as they are more canonical); but the same problems concerns also Ancient Greek gods and heroes – is the deciding authority Homer, who portrays Ulysses as a hero, or Vergil who portrays him as a villain? There is no canon here; yet both speaks about the same character, and with respect to these reference “worlds”, both – incompatible – descriptions of Ulysses are in a sense true. This can be evidence in favour of non-propositional nature of literature: however true can sometimes “propositions” of fiction seem, they cannot count as real propositions that have one uncontested truth-value in virtue of correspondence relation to the real world.

    This is hardly a satisfying analysis. Descriptions of features of fiction seem to be meaningful and such that can be true (or false), that is, they have to be a picture (a logical picture?). As a picture of something – though not in the Tractarian sense – they can be also exceptionally vivid and graphic. But their being a picture doesn’t probably consist in the correspondence with reality. Or more precisely: what they are picture of, is not a reality in a standard sense, as it can sometimes disturb laws of logic (such as the law of contradiction) – which reality usually doesn’t do. How are we to understand this strange kind of “picture” (picturing)?


    Wittgenstein is known to have recognized himself the weak points of his older conception of language – he replaces language as a picture by language as a variety of tool sets, “language games” (Wittgenstein 1953) The meaning of so understood a language is established in public (as patterns of use); Wittgenstein emphasizes the importance of rules governing the usage in games; the term “rules” covers “grammatic” rules of correct formation of words and sentences, as well as complex and contextualized patterns of (linguistic) behaviour. (Whereas the former are much the same or similar in most games, the latter differ (often dramatically) from game to game.) The obliging power of rules lays in their embodiment within the linguistic practice of the speakers’ community that warrants them.

    Of course, when speaking of practice and embodiment, the rules in question are – expectably – implicit rules. However, according to Wittgenstein, philosophy should describe the use of language; the slogan of inferentialism (a philosophical position elaborating many of Wittgenstein’s insights) is that the rules can be “made explicit” (see Brandom 1994). It doesn’t have to be always equally easy to identify the rules and make them explicit, but it is assumed to be possible. This is not only a “scientific program”: the ability to distinguish explicitly between following the rules and not following (i.e. playing the game correctly/well and not playing) is often very useful and sometimes even necessary.

    Now let’s consider the situation of literature in this new context. Apparently, fiction shouldn’t be understood as picturing anything. A new, more appropriate analysis sees it as a particular language game (or type of language games). As a language game worthy to be called as such, it must have rules identifiable to the extent that literature could be distinguished from non-literature. (Insofar as in the practice of our world, there really is both literature and non-literature, which is not one and the same, however vague the borderline between them is.) Can the supposed rules of the literary game (games) be made explicit, and if so, how?

    There are various definitions of what literature is, what it does, what function it fulfills or should fulfill (for them see e.g. Rexroth 1987). These apparent descriptions include often also certain prescriptive moment. Be they made from an ex post position, or establish a future program: in both cases they provide a guideline of what one should do, if what she/he does is to count as a piece of literature. This points to another problem: most games are not interested in their own definition and rules; literature does. As a counterweight to the demands on what literature should do, players of literary games often try to do something that can count as literature, yet (intentionally) breaks the explicitly stated rules. Hence, once some defining rules of literary game (games) are explicitly stated, they cease to be exhausting (at best).

    Moreover, the situation of literature is complicated also by the fact, that it is not quite clear, whether it is a single game, few games (representing several defined types of artistic expression), or indefinitely many games. (Wittgenstein himself is not explicit in this point – it seems that the complexity of a game can be low as well as high, and more complex games can include simpler ones.) If the last option is true, than each writer – in constituting a “literary space” on her/his own – creates her/his own game. But this would make literature close to the “private language”. In fact, there are analogies between literature (conceived as such special type of language game) and private language. First of all, it’s difficult to identify the rules of literature (as actually valid). And quite like in private language, the player can never be sure, whether she/he really plays the game (well, correctly), or just thinks to do so. The game displays no stability through time: does the novel I am writing now represents the same game as the novel I wrote two years ago? (Or, though I am still working on the same novel, is what I am doing now the same game as what I was doing yesterday?) Can it be uncontroversially decided?


    We have seen that though seeing the whole of language as a set of contextualized tools (language games) is probably more appropriate than seeing it as a picture of the world, this shift doesn’t have to work equally well in the case of literature. And on the other hand, the view on literature as certain kind of imagery (that is, picturing) proves itself to be more intuitive (though perhaps misleading). Let’s try this direction: can literature be conceived uncontroversially as a picture?

    In Philosophical Investigations, Wittgenstein introduces a very famous and influential notion of picture: “Ein Bild hielt uns gefangen.” – a picture held us captive. A picture is present within our language, and this graphic (pictorial) language, through which only – as a medium – we can see things, represents metaphorical “glasses”. Of course, the capacity to hold its users captive doesn’t have to be equally characteristic for any picture whatsoever. However, it is clear that there are pictures having this capacity, and essentially within the realm of literature.

    Literature can be conceived as a picture in this sense. Like the logical picture criticized by Wittgenstein, literature, too, purports us certain seeing of things. But the notion of picture is “functional” here: picture as something that is able to hold captive, to make someone do, and perhaps see (or even aware of seeing) – something. On the face of it, this may be a characteristics of – so to speak – pictorial (depicting) language games. However, this functional characteristic refers to all types of language games. Language game is an activity adapted to a situation, in which it should serve the wished purpose. And typically, playing a language game involves an interaction (communication) with another person or persons – so to perform a successful communication means mostly to make the other do what I want her/him to do (this is quite broad – including also the simple “make the other understand what I wanted to tell with my words”). Does therefore all language (all language games) have the nature of picture? Maybe, but to say this would be not more than to generalize a metaphor, though perhaps a true one (“Life is struggle”).

    What makes literature a picture in a bit more sophisticated sense is the fact that this is its aware purpose: literature focuses on providing pictures (in the sketched sense). The producers of literature often want to show something – they may reflect on their activity as picturing in both senses, the trivial (correspondence theory-like) as well as the sophisticated (functional, pragmatic). (Though the players of other games also provides these “pictures”, it is mostly implicit and unaware, and – consequently – not equally effective.) The quality of such picture, able not only to strike the recipient, but also to keep her/him (hold her/him captive) and make her/him do something, can be described in Heideggerian way as “Ursprünglichkeit”, originality (Heidegger 1959). The literary pictures are strong, urging, attractive, vivid; what they once “show”, cannot be then easily altered, neglected or forgotten – they place the depicted world very decidedly in front of our eyes. Note that this doesn’t have to be a trademark of “literature” as an established public institution; it concerns certain quality of utterance – various utterances can have this “literary” strength, but not the ambition to be a literature (preaching, a love-letter, skillful lying). And though literature attempts such making the recipient do/see something (being the functional holding captive), its self-interpretation can oppose the pragmatic explanation; so it is rather difficult to identify the contextual function of playing literature (the purpose of it as a language game).

    To sum, “picture” of “something” may be more appropriate analytical tool (at least an analytical metaphor, that is, analytical picture) in the case of literature, than the notorious machinery of language game is.


    Work on this paper was supported by grant No. P401-10-P266 of the Czech Scientific Foundation.


    1. Brandom, Robert 1994, Making It Explicit, Cambridge (Mass.): Harvard University Press.
    2. Heidegger, Martin 1959 Unterwegs zur Sprache, Pfullingen: Neske.
    3. Rexroth, Kenneth 1987 “The Art of Literature”, in: Kenneth Rexroth, World Outside the Window, New York: New Directions, 275-302.
    4. Wittgenstein, Ludwig 1921 “Logisch-Philosophische Abhandlung”, Annalen der Naturphilosophie 14.
    5. Wittgenstein, Ludwig 1953 Philosophische Untersuchungen, Oxford: Blackwell.
    Ondrej Beran. Date: XML TEI markup by WAB (Rune J. Falch, Heinz W. Krüger, Alois Pichler, Deirdre C.P. Smith) 2011-13. Last change 18.12.2013.
    This page is made available under the Creative Commons General Public License "Attribution, Non-Commercial, Share-Alike", version 3.0 (CCPL BY-NC-SA)


    • There are currently no refbacks.