That poetry is a marked instance of language, and that poems make expressive use of marked stylistic devices in both vocabulary and syntax, have been known facts about poetry for decades (Mukařovský 1989, Jakobson 1960, 1968, 1973). Somewhat accordingly, many scholars have brought concepts from linguistics to bear upon literary texts, specifically by analyzing the stylistic devices that seem to make the literary text effective (Austin 1981, Epstein 1981, Fairley 1981, Freeman 1981b, Keyser 1981, Kiparsky 1981, Tarlinskaya 1987). Nevertheless, many such stylisticians commenting on poetry have tended to focus on these stylistic expressive devices while often seeming to miss the point of the texts they are reading. These often inadequate, superficial readings result not from mere accident, but from unexamined and incorrect assumptions about the nature of poetry and the literary text, assumptions that neglect the text's literariness in favor of what is believed to be its communicative function as a message.
Normal communication requires not only utterances, but exchanges of utterances — a communicative context — to establish reference. People must be able to ask each other questions and have answers; the communicative exchange is a constant dialectic aiming toward mutual understanding. But the literary text, as Socrates (Plato 1982, 564-567 [275d-e]) points out about the `written' text in general, cannot answer any questions — it always says the same thing, no matter how things change around it. By contrast, a poem, and, indeed, any literary text, is a special case of the use of language, in which the referential — and even the apparently pragmatic — sense of the utterance is permanently removed from its communicative context and therefore cannot truly be said to communicate anything at all. The literariness of the literary text is, first of all, not a matter of its inherent makeup, but of how it is received. Literariness is, in the first instance, a cognitive, not a physical, category. Indeed, the person who, in a lover's quarrel, responds to a partner yelling in anger at him or her by observing, `You're so cute when you're mad', is, in a sense, transforming a communicative message into a literary text. This person is inappropriately regarding the utterance that expressed the anger as an object of contemplation and aesthetic pleasure, attending to the attractiveness of the form of expression rather than the anger it is intended to convey. A response that shows appropriate communicative engagement in this case might be fear, righteous indignation, or defensiveness, rather than admiration at the display of anger as if it were intended as a work of art. But the literary text is intended as a work of art, even if it represents an utterance expressing anger.
As in the foregoing example, literariness is not so much a quality of an utterance itself, but rather a quality of how an utterance is perceived, a frame of mind on the part of the hearer or reader that treats the utterance as something worthy of contemplation, discovery, and admiration, rather than the conveyance of an intended message. Nevertheless, literary texts usually have formal features that prompt readers to receive them as literary rather than communicative. When the reader perceives the text as literary, the apparent communicative content of the utterance is a sort of virtual rather than actual communication, one that would only take place within an imaginary world that would provide it with a communicative context. Since that imaginary world is created only out of inferences derived from the text's signs, it has limits that distinguish it from the infinitude of the real world. Insofar as the literary text communicates an intention, it does so at a level higher than that represented directly by its own language — outside, as it were, of the text's fictive frame. The fact that the literary text is bracketed off from the exchanges of normal communication also causes the text to be apprehended as a totality, rather than strictly as a sequence. Whereas normal communication is a potentially endless sequence of signs and responses through time, the literary text is a completed whole unto itself, acquiring a timeless spatiality in the reader's mind (upon completion of reading) at the same time that it also produces the appearance of a normal sequence.
The markedness of a literary text such as a poem will then serve simultaneously to express intention and to provide the aesthetic pleasure of surprise — the two, intention and surprise, become inseparable — but only insofar as such markedness is placed within its properly literary context and the reader assumes the correct default expectations for what would be unmarked. In short, the language of a poem does not mean, but seems to mean. The poem means only insofar as it uses this seeming within a larger context, which is not communicative in the normal sense, but literary and aesthetic.
By aesthetic, I mean appealing to a sense of pleasure in the contemplation of the object, in its totality and as a pattern. This is the purpose, or primary function, of an aesthetic object. By contrast, the primary function of communication is to convey meaning. If one communicates effectively, the aesthetic pleasure given by the form of communication should be negligible or trivial compared to the content being communicated. Effective communication usually means clear, relatively transparent, easily understood communication, with little distraction of the listener's attention to the formal features of expression. But if an utterance is considered as an aesthetic object, it does not matter so much what meaning it conveys as how the parts of the utterance work together to make a satisfactory or successful whole. Two opposite mindsets are required for appropriate responses, respectively, to what is understood as a communicative utterance and to what is understood as an aesethetic object — a literary text. The proper response to a communicative utterance will be directly relevant to the content of the utterance — a reply, an action, or an appropriate affect (such as fear or defensiveness in the case of angry speech). By contrast, the proper response to an aesthetic object, such as a literary text, is contemplation, interpretation, the pleasure of discovering how the parts come together to make a whole, and a sense of wonder at its wholeness, completeness, perfection, ingenuity, and forms of surprise.
Communicative messages are not subject to the same interpretive attitude as aesthetic objects, because communicative messages are embedded within ongoing communicative contexts that can restrict and control their meanings. By contrast, literary texts, as aesthetic objects, are appropriate objects pondering their meanings, since they are bracketed off from any communicative contexts other than the limited, imaginary ones they imply. Much of the aesthetic pleasure of the literary text is precisely in the pondering over possible meanings and the construction of plausible patterns of meaning. By contrast, communicative messages should aim at clarity, and ambiguity is not generally a source of pleasure but of temporary frustration, remedied by further exchanges of communication intended to clarify and disambiguate.
A poem by Emily Dickinson, `I Cannot Live With You' (1960:317-18, no. 640), illustrates some of these points. It begins with the statement (or, to be more accurate in speaking about poetry, the pseudostatement), `I cannot live with you — / It would be Life — .' We can infer that the speaker must just have been asked by someone to live with him — that someone has just proposed marriage to the speaker (and that the speaker is a woman, the person charged with either rejecting or accepting a marriage proposal). Later lines in the text clearly corroborate this inference. Thus we imagine a communicative context, in which an interlocutor, the suitor, has asked a question, `Will you marry me?' and what we `hear' is the response. But we cannot infer how the interlocutor would respond to this response, nor is it a relevant consideration for the experience of the poem. The focus of the poem is, rather, on the abstruse reasons the speaker gives for declining the proposal, all of which involve, ironically, a way of living up only too well to the (Congregationalist Protestant Christian) ideal of marriage, given the inherent sex inequalities within that ideology. The poem thus allows an ideological system to play itself out against itself. The neatness with which the speaker, in all innocence, can give voice to a standard ideology in such a way as to make the fulfillment of that ideology — marriage — impossible is, itself, a source of great aesthetic pleasure: we wonder at the brilliant wit of this argument that turns so seamlessly and yet ironically against its own assumptions. At the same time, the neatness of the poem's metrical and syntactic arrangements coincides with the feeling of inevitability about its argumentative structure, again to contribute to our delight in its ironic perfection. The meter and other poetic elements do not, then, contribute to the `message' of rejecting a marriage proposal; they contribute to the surprising aesthetic effect of the text for us, its readers. The rejection is not aimed at us. We get to delight in the ingenuity, brilliance, and perfection of the utterance. Within the imaginary world implied by the text, the imaginary recipient of the `message' would not be able to enjoy these aesthetic qualities. For him, it would be a message, not a literary text. By contrast, in our real world, outside the imaginary world of the text, we, the readers, can and should take pleasure in the wit and ingenuity with which the utterance is crafted — that is, its aesthetic qualities. The reader experiences much of this aesthetic pleasure in the process of solving the riddle of the poem — figuring out that it could be imagined as the rejection of a marriage proposal, and then marvelling at the surprising form of logic such a rejection takes.
This essay will attempt to clarify the problems of stylistics when applied to poetry, and, in so doing, will sketch a concept of the literary text — the poem — in a way that is more rationally defensible than the concept apparently held by many stylisticians. I hope that this concept of the literary text would also be more useful for teaching people how to read and appreciate poetry and other literature. I shall discuss the concept of literariness first, defining poetry accordingly. Then, I shall examine some errors that I think typify the potential pitfalls of stylistics-oriented reading. With these criticisms, I shall urge readings of the same texts more faithful to the concept of literariness, and show how markedness functions in these readings in the realm of the literary rather than in the realm of normal communicative utterances. Finally, drawing upon these examples, I shall review the phenomenon of markedness in relation to communication in the special case of the literary utterance, suggesting how what is `communicated' is an aesthetic experience, rather than a message.
A literary text is not communication in a direct and simple sense. If the utterance takes a propositional form ('Roses are red'), it is subject to the truth test in normal conversation ('Are they really?'), but not in a literary text (i.e., it matters not whether they are or not; the poem says they are, for its purposes). If the utterance takes the form of a question, it would be liable to an answer or, if a rhetorical question, a response in real life; in literature, it is only open to literary interpretation. The same with imperatives: in real life, one responds to `Take out the garbage!' by doing so or by refusing to do so, whereas, in reading poetry, one responds by figuring out the relevance of the seeming imperative to the rest of the seeming utterances that make up the poem as a whole, in order to discover the whole poem's point, its overall aesthetic effect. In poetry, even illocutionary acts have no force in the real world, but only contribute elements to an aesthetic whole. If a character says `I cannot live with you', this does not mean that a real marriage proposal has been declined, but that the formula of rejection is presented to us in a surprising way that gives us aesthetic pleasure.
The communicative context ennabling a communicative utterance has two elements: (a) a social context of exchange, and (b) a referential context, giving the utterance meaning and relevance. The social context includes within it the possibility that the recipient of the utterance will be able to ask questions or answer the utterance in some way, either directly or indirectly (as, for instance, by writing a letter after hearing a lecture in a large lecture hall). Even the refusal to respond, the failure to respond, or the accidental impossibility of responding contribute to the social context. The referential context must be clear enough to all parties to the communication in order for communication to occur at all. Reference is not exclusive to propositional utterances. If I say, `How are you today', it must be clear who the you is and which day today is in order for the utterance to qualify as normal communication. By contrast to all of this, the literary text requires (a) a reader, and (b) inferrable fragments of a communicative context (with both social and referential elements) sufficient to make the utterance make imaginary sense in an imaginary world. The elements of this `world' would be insufficient to provide the infinite possible preconditions, environments, and consequences of a real utterance. The fragments of a communicative context that can be inferred from the clues in the text are just enough to make the literary text `work', that is, deliver its aesthetic effect of surprise and satisfaction. The limitation of the communicative context to these few inferrable fragments of an imaginary world gives every literary text an air of ambiguity and invites the reader to ponder its interpretive possibilities. But that same limitation, the result of the text's being bracketed from the constant give and take of normal communicative exchanges, is what gives the text its finitude and completeness as an aesthetic whole, so that the entirety of its pattern can be appreciated and enjoyed.
Because the communicative message and the literary text are received cognitively in completely different ways, the appropriate responses to each are different. When the speaker of Wallace Stevens's `Blanche McCarthy' exhorts, `Look in the terrible mirror of the sky', our first task is not to look up (and thus cease reading the poem), but to consider what the `terrible mirror of the sky' might, figuratively, mean, and thus how it will fit into that text's system of figurative signs as an aesthetic experience. When the speaker of John Keats's `Ode on Melancholy' admonishes, `No, no, go not to Lethe', we are to imagine that the utterance would be properly directed to an imaginary person — a virtual literary character — who is about to commit suicide. It is unlikely that we would happen to receive the message appropriately when we, ourselves, are on the verge of committing suicide, and would pause long enough to read the poem through to take the rest of its admonishments. Here, as elsewhere in poetry, even the seeming message — the content of the imaginary communicative utterance in an imaginary context — is far too complex and ambiguous to serve as a practical message in a real-life situation. This richness is not a necessary feature of a literary text, but it follows naturally, on both the poet's and the reader's parts, from the assumption of the appropriate frame of mind that the reader would hold in receiving a literary text rather than a communicative message.
That the appropriate response to the imperative beginning `Ode on Melancholy' is the imaginary reconstruction of a possible scene in which the utterance would be a communicative message introduces a crucial paradox about the literary text. This paradox is that the literary text's existence depends upon the hypothetical possibility that the very same utterance, the same signs, could be employed in an imaginary communicative exchange within an imaginary referential context.
Literary meaning, then, is a special case of meaning. What a literary text `communicates' to its reader is its aesthetic effect, which may be understood as some form of surprise. Literary texts use the material ordinarily used for communication in order to deliver surprise, which pleases the reader, edifies him or her, sharpens readerly skills, and performs other functions associated with aesthetic experience rather than with communication per se. Literary texts achieve surprise by means of various kinds of markedness. The very removal of the utterance from any conceivable actual communicative exchange and referential context already makes the utterance as a whole, by definition, marked as literary, and it is in part the task of the literary writer to produce sufficient linguistic evidence to make sure that the reader does not mistake the text for a communicative message. Verse in poetry has, among its functions, that of setting the language of poetry off from that of normal communication, since, in verse, attention is drawn toward the mere surface of language — its pure sounds or visual appearance in writing — rather than to its meaning. This shifting of attention would be inefficient and counterproductive if the utterance were intended for communicative purposes, since the aim of a communicative utterance is, first and foremost, transparency. When a parent tells a child to wash the dishes, the parent does not want the child to dwell upon the sound of the parent's consonants or the repetition of sh in wash and dishes. This abnormal focus on the surface of language rather than on its meaning is specific to poetry. Nonpoetic literary texts use a variety of other devices in order to ensure the reader's response to the text, as a whole and in the last instance, as aesthetic rather than communicative.
But the mere existence of verse in poetry is not generally sufficient to deliver surprise, except in a very vague way. Poetry produces surprise by means of marked choices against default assumptions on many different levels, all depending either on literary contexts alone or on the peculiar and paradoxical interaction between literary contexts and the nonliterary, real contexts that they imitate or to which they allude, in actual communication. For example, Emily Dickinson's `Because I Could Not Stop for Death' (1960:712, no. 350; see Aviram 1994:263-279) surprises the reader in many ways, including the following.
The speaker apparently has already died and is thus speaking from beyond the grave. This doesn't generally happen in real life. In literature, such speeches are usually attended by signs of the supernatural, and the speeches are often cryptic and oracular. Here, the speech is a matter-of-fact past-tense narrative of the events leading up to the speaker's present (dead) state.
The speaker has died, but she does not sound as if she is in heaven or hell, but rather in a state of complete emptiness, where nothing happens. (This also makes the idea of her speaking from such a state even more unfathomable.)
The poem is in tetrameter quatrains (four beats per line, four lines), with the second and fourth lines truncated so that the fourth beat is observed as silence rather than being realized with a syllable of a word. This is the conventional hymn form, which would be conventionally associated, in the Christian context to which other signs in the poem allude, with praises for the happy afterlife promised to the saved Christian soul. Yet this poem contains no such praise, and thus the poem forms a sort of anti-hymn when read against the cultural context of the Christian beliefs to which it alludes or that can assumed to be known (and possibly held) by its immediate readership in 19th-century America.
The allegorial figures of the poem describe death as if it were marriage — to the figure Death. This itself is a small surprise — the bridegroom being death. But death has long been figured in Christian literature as like a marriage — to Christ, not to death. So having Death appear in the place of Christ is another surprise.
The surprises all work together to create what appears to be the speech of a naive speaker who now knows that death leads, not to heaven or hell, but simply to an immense vacancy. In other words, the poem dramatizes a reversal of fundamental Christian beliefs about the afterlife. The effect of the poem does not depend on this apparent revelation being either true or false; it depends on its being imaginatively successful and giving us the pleasure of surprisingly reversing ordinary, normative expectations. All this is accomplished by an intricate network of markedness on thematic, formal, and figurative levels.
The capacity of a text to deliver pleasing surprise depends, therefore, upon the specific knowledge of its readership, knowledge drawn both from life and — especially — from literature. What is delightfully surprising to one reader may be simply chaotic to another and too predictable to a third. This difference is what accounts largely for differences in literary tastes, critical receptions of various poets, and the constant changes in academic and culture-wide canons of good poetry. This fact, too, creates a paradox, this time about the teaching of literature. Students appreciate a literary text insofar as they have sufficient prior knowledge to view the marked features of the new text with the delight of surprise. Yet it is that very pleasure that would motivate a student to wish to acquire sufficient knowledge — sufficient literacy — to have it in the first place. In teaching literature, we constantly find ourselves in the position of trying to get students to read things in the hope that, eventually, they will realize — always retrospectively, and often too late for us — that the texts we assigned did, in fact, give them pleasure. But this is precisely why academic canons cannot be based upon prevalent popular tastes and are, in some ways, necessarily conservative. The task of teaching literacy includes that of continually reintroducing sufficient literary context from the past to make the pleasurable surprises of texts available to succeeding generations of students.
Learning to read poetry is not the same thing as learning to read messages, although the former recruits the skills of the latter in order to gain access to surprise and literary pleasure. Yet the approach that some linguists have taken, in commenting on the style of particular poems, presumes an identity between these two skills. As a result, stylisticians of high professional repute and considerable expertise in their field have, in their commentaries on poems, completely missed the surprise constituting the point of those poems, and belabored issues that are somewhat more obvious. Teachers and students, in turn, read these commentaries, and are likely to repeat the stylisticians' mistakes: to confuse the literary text with a communicative message, to relegate literary devices to the role of `boosters' for the message's meaning, to read parts in sequence at the expense of the whole, and to oversimplify the sense of the literary text.
Because stylisticians assume that what makes poetry poetry is the use of various marked linguistic devices in order to convey meaning — i.e., to communicate — stylisticians tend to assume a given meaning for a poetic text and then elaborate all the various devices that `boost' the expression of this meaning. In many cases, the meaning they infer as their starting point is oversimplified, and sometimes illogical when applied to the totality of the poem, the whole text rather than parts of it in sequential order. This superficiality, flatness, or even wrong-headedness is not simply an accident. It proceeds from the project of `revealing' the expressive devices of a text rather than, first, understanding the way literary texts such as poetry work as aesthetic structures whose function is to deliver surprise when given the right context and when read as totalities rather than sequences of elements. Stylisticians tend to assume that the literary text is a message like any other, only more punchy or expressive, using more elaborate devices to get the message across. Accordingly, the message is a sequence of signs, bearing the normal relation of beginning to end, rather than a sequential elaboration of a complex structure whose parts must be remembered and held together in the reader's mind to appreciate the totality. These assumptions are incorrect and lead, ultimately, to disappointing misreadings of rich and powerful poems.
My first example of this problem is E. L. Epstein's (1981) otherwise interesting essay, `The Self-Reflexive Artefact: The Function of Mimesis in an Approach to a Theory of Value for Literature.' Epstein discusses many great poems in a variety of languages, including, briefly, William Blake's `The Tyger.' His general thesis is that literary value, especially in poetry, derives from the extent to which linguistic devices are used in a mimetic fashion — to imitate the content of the poem. Thus, Epstein's sense of the term self-reflexivity differs sharply from its meaning in literary theory. For Epstein, self-reflexivity means the `reflection' or imitation (or merely expression, representation) of the meaning of the message in its form — as if the form were a dressing or clothing for the meaning. By contrast, in modern literary theory, the term refers to the way that the literary text, bracketed off from normal communication, represents in its thematic content its own nature as a literary text. That is, the poem is `about itself.' (See Aviram 1994:17-28, 43-58, 223-246.)
In the specific case of `The Tyger', Epstein spells out quite neatly the content he finds imitated (or expressed) in the form:
What Blake achieves in Tyger [sic] is a subjective mimesis, based on syntax, for a very high degree of value. It seems obvious from the poem that what Blake is coneying is his own awe at perceiving the energy that drives the universe and the poet, a force here symbolized as a tiger, a power beyond good and evil. (Epstein 1981:187)
Obvious, indeed. Epstein (1981:186) also notes, beforehand, that, after the reader does `his "first complete reading", as it could be called, all further readings contribute nothing new and become increasingly automatized scannings of a familiar linear pattern'.
Is rereading `The Tyger' that useless? I doubt it, especially if the results of the `first complete reading' are as Epstein summarizes them, quoted above. Let us consider the poem as would a good literary reader, rather than a stylistician. The good reader's goal will be, first, to read the whole of the poem, and to understand how the parts relate to the whole that comprises them.
It is striking that `The Tyger' is made up exclusively of questions.[1] From an interpretive point of view, these questions beg a question: are the questions to be taken as straightforward, wherein the speaker seeks answers, or are they rhetorical questions, not intended to be answered or having self-evident answers? Rhetorical questions are marked uses of the normal question structure, which is normally assumed to seek an answer. One of the ways that rhetorical questions are marked as such is by containing within them the information that would otherwise be left for the answerer — that is, the rhetorical question contains its own answer, and thus prompts a response that accords with its implied answer.
The third question would likely be understood, in conversation, as a rhetorical question, wherein the statement, `You are not going swimming', is contained, as it were, within the question. Whereas (1) leaves open any possible answer suggesting a plan or activity, and (2) leaves open either a yes or no answer, question (3) requires either assent — 'Yes, you're right, it's too cloudy' — or remonstrance — 'Actually, I was planning to swim; why not?' Question (3), in other words, is a version of the statement, `You shouldn't go swimming today', or even the command, `Don't go swimming today.'
In `The Tyger', the questions are not obviously rhetorical, but not clearly straightforward, either.
Tyger tyger, burning bright
In the forest of the night;
What immortal hand or eye,
Could frame thy fearful symmetry?
As a rhetorical question, this sentence would be equivalent to the statement, `Surely, no immortal hand or eye could frame thy fearful symmetry.' Yet, such a statement makes no sense, semantically, within the context of social and cultural beliefs that would seem applicable to the poem. If, that is, the tiger were `created' at all, then it would have to have been created by some sort of immortal being — a being with powers sufficiently divine to create mortal beings, and thus, itself, immortal — whether God or a classical figure such as Prometheus or some other imaginary immortal being. In this case, then, the question is not rhetorical but straightforward, asking for the identity of the divine being who created the tiger. And yet, like a rhetorical question, this question contains within it sufficient information to make the answer necessarily one from a limited number of choices — some being that has an immortal hand or eye.
In the second quatrain, the speaker further informs our concept of the being about which the questions are posed:
In what distant deeps or skies,
Burnt the fire of thine eyes?
On what wings dare he aspire?
What the hand, dare seize the fire?
The maker of the tiger must have had wings, and must have dared to aspire. It would be inappropriate to speak of God as daring to do anything. The image in the questioner's mind seems to derive from that of Lucifer, leader of the rebel angels, in Milton's Paradise Lost. (The image of Daedalus aspiring to fly may be relevant, but probably not as clearly so as the image of Satan.) Following this thought, the last line presumes that the tiger's maker had seized the fire used to make the tiger's eyes, in a way reminiscent of Prometheus's theft of fire to give to man. The transition from Lucifer to Prometheus is possible because (a) both of them are types of rebel divinity, and (b) in some sense, neither of them could really have made the tiger — neither the story of the Fall of the Angels nor that of Prometheus's creation of man (after which he gives man fire, against the will of the other gods) quite fits in an imaginary story of making the tiger — and, in any case, there is no actual story of the tiger's creation in canonical literature and myth. Milton's Satan specifically is incapable of creating life — a notion following that of Spenser's arch-demon Archimago, who can only make illusions, not real living beings. And Prometheus only creates man and steals fire to give it to man; Prometheus would not make an animal that man finds frightening. So, in both cases, the questions contain within them hypotheses that make the questions unanswerable. These are special, marked questions, because they contain enough information to act as rhetorical questions, but the answers that they would prompt if they were rhetorical questions would be patently incorrect.
One of the most problematic aspects of Epstein's in reading of `The Tyger' is his assumption that the speaker of `The Tyger' is the same as the actual author of the poem, William Blake. As we have seen thus far, the speaker of the imaginary speech that the poem represents is somewhat befuddled by his own questions. He asks questions that are impossible to answer, because they at once contain too much information to be open to an answer as if they were straightforward questions, and, as rhetorical questions, presume incoherent or impossible answers. The befuddlement of the speaker becomes even more dramatically obvious in the next quatrain:
And what shoulder, & what art,
Could twist the sinews of they heart?
And when thy heart began to beat,
What dread hand? & what dread feet?
The first question here continues in the line of the previous ones, suggesting that, whoever made the tiger, he must have (or have had) some great physical powers to make something of such power to terrify. But the second question seems to confuse the object of creation with the creator. The `dread feet' couldn't really be attributes of the creator. Whereas the hand would be used to make the tiger and would follow in the series of metonymies, shoulder, art, hand — as well as recalling `What immortal hand' from the opening and `What the hand dare seize the fire' in the second quatrain, feet are not used to create. This is not to mention that it would be hard to imagine in what way feet can be dread, except insofar as a tiger has claws on both fore and hind paws and springs to the attack on its hind legs, and thus has dread feet. The very confusion or misattribution of qualities from the tiger onto the creator, or perhaps vice versa, illustrates something about neither the tiger nor the creator, but rather the speaker himself. The speaker seems to be caught up in a sort of hysteria, compatible with his inability to accept the real answer to his original question.
The mood of hysteria rises even higher in the next quatrain:
What the hammer? what the chain,
In what furnace was thy brain?
What the anvil? what dread grasp,
Dare its deadly terrors clasp?
One can imagine a hammer and an anvil involved if the making of the tiger is likened to a product of the Greek god Hephaestos (Roman Vulcan), yet another potential-but-impossible creator, since Hephaestos also never actually makes living beings, but rather artefacts such as Achilles' shield. It is hard even to imagine how a chain would come into play here. And, while it makes some sense to think of the tiger's brain as being full of `deadly terrors', it is hard either to imagine the brain, even a tiger's brain, as the product of a hammer, an anvil, and a chain, or to imagine why a `dread grasp' would be necessary to take hold of it. Most of all, the questions at this point seem to fall thick and fast, contributing to the mounting emotional pitch of the speaker.
The next quatrain both alludes again to Paradise Lost and makes clear to us, if it hasn't been before, why the speaker is having so much trouble:
When the stars threw down their spears
And water'd heaven with their tears:
Did he smile his work to see?
Did he who made the Lamb make thee?
The first three lines, again, seem to presume that it was Lucifer — or Satan, as he is known after the fall from heaven — who created the tiger, again, a doctrinal impossibility. Here, the presumption leads to a fearful imagining of Satan taking pleasure in making something terrifying to humanity. But the closing line asks the very opposite question. If God made the tiger, it asks, how could the very same God who make a soft, friendly, and woolly thing like the lamb also make the ferocious tiger? How could the God who made Good also make Evil? This last gloss of the speaker's question is further assisted by the proverbial identification of the Lamb of God with the person of Jesus, and also by the poem, `The Lamb', in Songs of Innocence, composed a few years before Songs of Experience in which `The Tyger' appears. The title, `The Tyger', may in some sense allude to this earlier poem, as both titles are generic names of animals. In `The Lamb', the speaker, who identifies himself gleefully as a child, says that the lamb was created by God, and both the lamb and the speaker (a child) typify the innocence that God has created. The contexts thus reinforce the idea that the problem the speaker has in `The Tyger', the problem pushing him into a state of hysteria, is the traditional Problem of Evil: How could God, who is good, have made evil things in this world?
The final quatrain, repeating the first but for one word, once again seems to settle — and yet not to settle — upon the alternative answer, that it was not God, but Satan, who created the tiger:
Tyger Tyger burning bright,
In the forests of the night:
What immortal hand or eye,
Dare frame thy fearful symmetry?
The word dare recalls `On what wings dare he aspire', alluding to Lucifer's rebellion against God, as if to suggest that the creation of the tiger was, itself, a gesture of rebellion and usurpation of God's creative powers.
To review, two important points come out of our close reading of Blake's `The Tyger.' First, the speaker's questions have the form of rhetorical questions insofar as they contain enough information to imply their own answers, but the answers that such rhetorical questions would prompt would be both inconsistent with each other and generally impossible according to the doctrinal and mythological contexts to which they allude. Second, the speaker's insistent questions seem to boil down to a refusal to believe the most obvious answer to the question of who made the tiger, God, because it seems impossible to the speaker that a good God could make evil things.
But the very same logic that has enabled us to question the assumptions that the speaker's apparently rhetorical questions assume should also enable us to question the assumptions underlying the central question, the Problem of Evil. For this problem is only relevant insofar as it is assumed, to begin with, that the tiger is evil, an assumption manifest with all the terms suggesting terror, fear, dread, etc., throughout the poem. Just as the speaker has created for himself his own incapacity to answer his own questions by insistently posing questions that are too closed to allow the correct answer (God), so, too, he has created his own problem to begin with by assuming that the tiger is evil. In a sense quite consistent with Blake's other work, the speaker himself has `created' the tiger, insofar as he has created in his own imagination the object of which he is terrified and projected that object onto the actual being, the tiger.
This idea that evil is created in the eye of the beholder is consistent with other poems in Songs of Experience, in particular, `A Poison Tree':
I was angry with my friend:
I told my wrath, my wrath did end.
I was angry with my foe:
I told it not, my wrath did grow.
And I water'd it in tears,
Night and morning with my fears,
And I sunned it with smiles,
And with soft deceitful wiles;
And it grew both day and night,
Till it bore an apple bright,
And my foe beheld it shine,
And he knew that it was mine,
And into the garden stole
When the night had veil'd the pole;
In the morning glad I see
My foe outstretched beneath the tree.
Blake's retelling of the Fall in the Garden of Eden — a story originally about the acquisition of knowledge of the difference between good and evil — draws our attention to the minimal difference between friend and foe — supported by the minimal phonological difference in the first and third lines: I was angry with my f—. We have no reason to know, or to think, that the friend and the foe are inherently different, until the speaker chooses to treat them differently. The speaker feels wrath equally towards his friend and his foe. What defines the foe as a foe is not that the foe incurs the speaker's wrath, but that the speaker does not tell his wrath to his foe, leading eventually to disastrous consequences, in which the foe lives up to his role as foe and the enmity between speaker and foe reaches the extremes of deception and death.
This idea also corresponds to the meaning of the phrase in `London', also in Songs of Experience, `the mind-forg'd manacles':
I wander thro' each charter'd street
Near where the charter'd Thames does flow
And mark in every face I meet
Marks of weakness, marks of woe.
In every cry of every man,
In every infant's cry of fear,
In every voice, in every ban,
The mind-forg'd manacles I hear.
Everywhere the speaker goes, he witnesses a suffering connected to restriction and confinement — even the Thames river is `chartered' and thus not free to all. But the manacles of confinement are forged in the mind. They are, implicitly, the products of bad ideas and lack of imagination.
Finally, `The Tyger' has its own most immediate visual context to give us a strong clue to corroborate the above reading. Blake originally published his Songs of Innocence and of Experience himself, etching each poem by hand into a copper plate, as if to imitate medieval illuminated manuscripts, along with illustrations to the poems, some of them exquisite, which he colored in by hand-painting them on the newly printed books. `The Tyger' has one such illustration: a picture of a tiger, below the words of the poem and in the bottom quarter of the page. (A tree along the right side of the page may suggest an Edenic setting and pehaps reinforce the connection with `A Poison Tree.') The tiger hardly looks frightening or terrible. If anything, with eyes wide open and shoulders low, he looks a bit frightened himself, which would make perfect sense if he is the target of the speaker's frightened — and thus frightening — projections. Even if the other literary contexts mentioned here as corroborating evidence should be considered inaccessible to the beginning reader — and they should not — still, the picture itself is evidently as much a part of the `text' of `The Tyger' as the words are.
We have come a long way from Epstein's reading of `The Tyger' as an expression of Blake's `own awe at perceiving the energy that drives the universe and the poet.' The shallowness of Epstein's reading comes despite his meticulous commentary on specific stylistic choices within the poem, some of which may be correct, but all of which are trivial if, to begin with, they are assumed to support an unsatisfactory reading. But the reductive, superficial reading is not an accident. Epstein's failure to recognize the various kinds of irony at play in `The Tyger' arises from the theoretical assumption that the poem is a communicative message — in this case, expressing Blake's awe — and that the task of a good reader is to notice how expressive this message is. `The Tyger' is not a communicative utterance. Rather, it is a dramatic exercise, involving an imaginary speech. Our task as good readers is to notice where the surprises lie in the content and form of this speech, and what sort of meaningful pattern the surprises form. This pattern gives us, simultaneously, the poem's meaning and its aesthetic effect.
The poem's surprises depend on markedness on a variety of levels, many of which require knowledge of appropriate contexts. For instance, the clash of implied answers between Satan, Prometheus, and Hephaestos depends upon our recognition of literary and mythological allusions. The summary of the questions as the Problem of Evil depends somewhat on familiarity with this problem in Western thought, especially in religion. In the former case, the unmarked condition would be a single, coherent field of allusive reference, making the questions neatly rhetorical and their answers easy to produce. In the latter case, the unmarked condition would be the simple answer, yes, God made all animals, both the lamb and the tiger. Without these contexts, one cannot fully appreciate the way that the speaker seems to be answering his own questions with answers that he cannot accept, and yet cannot accept the answer (God) that would seem more obviously appropriate, given his concerns, because of the Problem of Evil. And without these contexts, the reader cannot arrive at the point of noticing that the speaker is in some sense inventing his own difficulties by categorizing the tiger as `evil' from the start. The unmarked condition here would be, either a celebration of the tiger as a natural creature — an attitude somewhat typical of the Age of Sensibility — or a complaint about the appearance of evil in an otherwise good world — a sentiment we might find in Renaissance literature, in which things of nature might be employed allegorically to represent moral categories such as good and evil.
Blake creates a marked utterance on this level by bringing together these two, incommensurate viewpoints — an allegorical and a sentimental view of Nature — in order to reveal the factitiousness of both views as products of the perceiver's projections onto the world. This last point, however, is not Blake's message, as such, but rather the effect of his drama. Poetry does not tell, it makes things happen with words.
In `The Tyger', Blake causes us to recognize the difficulties someone can get into by making moral assumptions about things in nature that have no a priori moral essence. In this sense, we could say that the poem `communicates' to us, but what it communicates is not a message, but an experience. This experience — what happens if you have this attitude — is indistinguishable from the poem's aesthetic effect, which is the effect of surprise. Here, the surprise is, ultimately, a series of surprises: (a) the speaker is saying nothing but questions; (b) the questions can't be answered, even though they sound like rhetorical questions; (c) the questions reveal great consternation; (d) the speaker has created the cause of his own consternation; (f) the answer to the question turns out to be `God' only in the first instance, but `the speaker' or `the perceiver' in the last. It is this succession of intellectual surprises that makes the work simultaneously deep and delightful.
My second (and last) example of an unsatisfactory reading produced by a stylistician is Samuel J. Keyser's (1981) reading of Wallace Stevens's (1954:9) `The Snow Man.' Let us first recall the poem.
The Snow Man
One must have a mind of winter
To regard the frost and the boughs
Of the pine-trees crusted with snow;
And have been cold a long time
To behold the junipers shagged with ice,
The spruces rough in the distant glitter
Of the January sun; and not to think
Of any misery in the sound of the wind,
In the sound of a few leaves;
Which is the sound of the land
Full of the same wind
That is blowing in the same bare place
For the listener, who listens in the snow,
And, nothing himself, beholds
Nothing that is not there and the nothing that is.
Keyser (1981:112-120) comes closer to a convincing reading of this poem than Epstein does with `The Tyger', but not close enough to keep him from missing the point entirely. In order to decide the semantic meaning of the poem, Keyser first quotes Frank Kermode and then Stevens himself, obviously according more authority to the latter than the former. Here is Kermode first:
Out of `The Snow Man' grows the recurring metaphor of winter as a pure absracted reality, a bare icy outline purged clean of all the accretions brought by the human mind to make it possible for us to conceive of reality and live our lives. So purged, reality has no human meaning, nor has a man; he is
...the listener, who listens in the snow,
And, nothing himself, beholds
Nothing that is not there and the nothing that is.
(1960:34; qtd. in Keyser 1981:119)
Keyser (1981:119) continues:
In contrast to Kermode's interpretation of The Snow Man we have Wallace Stevens's own explanation which he gives in a letter to Hi Simons, dated 18 April 1944: `I shall explain the Snow Man as an example of the necessity of identifying oneself with reality in order to understand it and enjoy it.'
The New Critics of the early-to-middle twentieth century generally urged that readers view a poet's own comments about his poems, especially brief and cryptic comments in personal correspondence, with the utmost skepticism: we cannot really know how much thought or knowledge was assumed as the context for the letter, and we cannot know whether Stevens willfully, or perhaps unknowingly, reduced the complexity of his own poem for his correspondent's sake. Nevertheless, not heeding this principle, Keyser takes the beginning of the poem, the words `One must', to indicate that the entire poem is an admonishment to the reader, a behest telling the reader that he or she must be or do as follows. Granted, the bulk of Keyser's comments focus — correctly — on the remarkable structure of the poem's single sentence, which is built so that the reader will complete a given portion and think that that is the whole of it, only to keep reading and find out that what he or she had thought to be a whole turns out to be only a part of a much more complex sentence.
Despite this fine insight, ironically, Keyser still allows the apparent imperative form of the opening, `One must', to dominate his sense of the whole. He treats the complex structure of the sentence as an expressive ornament to dramatize the sense of the original imperative, in effect, 'One must be like a snow man.' Keyser therefore misses the point that the sentence turns out to have an altogether different, surprising meaning when read as a whole. The sentence begins as an apparent imperative, but ends up actually being more like a complex conditional-resultative sentence.
It helps if we reduce the poem's sentence to something of a skeletal paraphrase. We can take advantage of the poem's many parallelisms — such as the parallel of regard in the first tercet and behold in the second, and the frost and the boughs and pine trees crusted with snow within the first tercet, as well as both of these with the junipers shagged with ice and the spruces rough in the distant glitter / Of the January sun.
One must be a creature of winter to perceive wintry things and not perceive emotion (misery) in any of them; only a snow man, who is nothing himself (not a perceiving subject), can (as it were) perceive nothing that is not there and the nothing that is.
The sentence thus has an idiomatic structure something like that of the sentence, `You have to be heartless to know her and not love her.' This latter does not urge someone to be heartless, it explains what it would have to take to avoid loving a lovable person.
Stevens's sentence is not a recommendation to become like a snow man, which would be impossible. If it were possible for the reader to become like a snow man, that would mean that the reader would bring nothing to the reality he or she perceives — would impose no conceptions such as `misery in the sound of the wind.' But that would also have to mean that he or she would not bring any knowledge to bear upon perception, even the perception of the marks on the page that are the letters and words of the poem. In short, a snow man cannot read poetry. So, if the reader were (hypothetically) successful at becoming a snow man, he or she would do so only at the expense of his or her ability to read the poem or even to know that he or she has succeeded. The snow man is nothing himself. Even before this, the poem gives us a strong clue in the phrase equivalent to `have a mind of winter', that is, `have been cold a long time.' The latter phrase sounds like a reference to the dead. This does in fact make sense later on, once the reader figures out that the snow man is not a human being and that one would have to be dead — be nothing — to `see' the world the way `he' does — that is, not see at all. The poem then works, as it happens, as a nice allegory for one of the main points in this paper: the importance of the extrinsic literary and cultural contexts brought to bear upon the literary text.
The surprising shifting of the sentence in Stevens's `The Snow Man' is not a demonstration of the openness that being a snow man would entail — we cannot possibly think like a snow man, since a snow man does not, by definition — even the poem's definition — think. Rather, it is precisely the turn, from apparent injunction to actual observation about the nature of humanity, that surprises and shocks the reader out of the complacency of his post-Romantic expectations. A commonplace of Romantic thought was and continues to be the idea that humanity is alienated from Nature, in permanent exile, because of its self-consciousness. The very things that make us human — language, reason, morality, civilization — also make us somehow separated from Nature, and fill us with anxiety and a sense that we do not belong where we are. With this idea comes a yearning for the blissful state of oneness with Nature. To put the thought simply (at the risk of caricature), squirrels are lucky because they never have to worry about whether they are doing the right thing and never have to wonder about the meaning of life. But whereas, in the Romantics, the object of such yearning might be a bird (such as Keats's Nightingale), which may or may not in fact have some awareness of its own mortality (for all we know), Stevens's snow man is most certainly without any awareness of anything, since it is not a living being at all. Indeed, even the name for it, snow man, is the product of human projections and human creation: it is only figuratively a `man' made of snow. The unmarked condition in this case is, precisely, the expectation that one must have a mind of winter and be like the snow man. The surprise comes when we realize that, far from having to be a snow man, one has no possible choice of being a snow man, since being one would be neither being nor choice at all.
There is, indeed, a deeper paradox in the poem, however. The common human attributes that include the hearing of misery in the sound of the winter wind also include our ability to contemplate what it would be like if we were not ourselves. Imagining other subjectivities is part of the human condition because it is one of the capabilities of imaginative projection, and this includes at least trying to imagine impossible subjectivities, beings that cannot possibly be subjects at all, such as the snow man, `nothing himself.' The poem does not so much tell us that this is an aspect of the human condition as allow us to experience this paradoxical fact as another result of the surprise that the poem delivers. Again, the poem's `meaning' — that is, its point — is inseparable from its aesthetic effect, its surprise and its paradox.
Like Epstein, Keyser makes his mistake precisely because he expects the poem to express the author's ideas or feelings as a communicative utterance. This renders him insensible to the possibilities of surprise on the level of ideas, even though Keyser is quite sensitive to surprises in syntax. He assimilates the surprises of syntax to an incorrect reading of the content because he assumes a simplistic reading of the content to begin with and then assumes that the syntactic patterning will enhance or boost the message. This logic is very similar to that of Renaissance theorists who viewed such things as verse or poetic imagery as a kind of rhetorical `garb' to `clothe' the truth carried within the text as its message. Then as now, this confusion results in a tendency to read parts at the expense of the whole.
Perhaps most fascinating here is how the poem itself seems to prepare the way for the mistaken reading — not only by setting up the surprise, but also by then addressing the very questions about reading that the mistake — and the surprising turn to correct it — would naturally pose. The poem presents to us the paradox of how, like reading, human perception always includes projection, but our ability to project can also include our imagining impossible nonsubjects as test cases, nonsubjects who, if they could exist, would not project when they perceived. The reader who most `truly' understands the text is the one who projects nothing at all, but this would mean that he or she actually understands nothing at all and is not in fact reading. Reading always involves distortion, even though its aim is the absence of distortion. It is this very paradox that keeps the poem from being either a simplistic recommendation to a Zen-like nonprojection or an equally simplistic celebration of human projections. Rather, it dwells upon the paradox that makes (perfect) reading simultaneously possible and impossible. It takes a good enough reader to recognize, ultimately, that the poem is also saying that the perfect reading is impossible — and yet that is the perfect reading, since no other reading will do. This self-reflexivity — the poem allegorically representing how poems work — rather than the mimesis of imitative form that Epstein calls `self-reflexivity', accords best with literariness. For when the utterance is removed from normal communicative and referential contexts, it can only ultimately refer to itself and illustrate the paradox of language that only appears to communicate — having no communicative exchange to make it do so; and yet, in a sense, that very paradox is what the text `communicates.' Every literary text approaches this paradox of self-reflexivity in a different way, but the ultimate pleasure of surprise and fascination with the paradox of a language that has the substance but not the social circumstances of communication remains always at the end of the road of interpretation.
As we have seen, literary texts, including poems, have the function of delivering surprise, provided that the reader can recognize the appropriate contexts as the unmarked conditions against which the texts' features are marked. This means that literary pleasure presupposes by its very nature literacy — or what Jonathan Culler (1981) aptly calls literary competence. Literary competence is not a stable state, however, but an infinitely progressive accumulation. Reading is a lifelong apprenticeship in the art of literary competence, as an increasing number of texts becomes increasingly accessible to the studious reader. But what about readers who aren't there yet? How does a reader get started? Are we not faced with a closed loop, where the pleasure of reading requires prior reading, which requires the capacity for the pleasure of reading?
The pleasure of literary texts, fortunately, is available to various degrees on various levels simultaneously. At the most rudimentary, the necessary context of the appropriate `unmarked' default expectations comes from very broadly shared cultural traditions, ideas, beliefs, concepts, or images. For instance, almost everyone who has learned to read in a Western country will recognize the collocation of a tree with a dangerous fruit and a garden into which the foe steals as some sort of allusion to the Fall in the Garden of Eden, because this web of myths is so widely known in some form or other throughout Western culture, even among people who have never read the Bible or Milton's Paradise Lost. In the case of `The Tyger', the incoherency among the specific divine figures to which the questions allude would require a fair bit of prior reading. But the fact that the speaker utters nothing but questions, and that the questions, while containing sufficient information to be rhetorical, do not seem to permit the very answers that they seem to prompt — these facts should be evident to a reader attentive to mere syntactic and rhetorical conventions and only the vaguest and broadest cultural beliefs or traditions about creation and divinity. In the case of `The Snow Man', knowledge of the Romantic tradition of yearning for an unalienated unity with Nature only provides extra elaboration beyond the more basic point, which is that the sentence winds up not urging the reader to be like a snow man but rather says what being a snow man would be like, if it were possible (and it turns out to be, paradoxically, quite impossible). The reader who gets this point only need read the sentence as a whole rather than stopping too soon, and need think about the logical implications of what it would mean to urge someone to be `nothing himself' and therefore unable to perceive anything at all, including the poem making this supposed behest. Literacy in both of these instances requires, above all, that the reader read the entire poem through, all of its sentences, trying to see a pattern in the whole, and only then consider how each part contributes to that whole, rather than moving sequentially and making controlling inferences at every point from limited information. The procedure of accepting parts of the whole as dominant — without sufficient attention to possible irony — characterizes normal communication, rather than the reading of literary texts. A literary text is first and foremost a whole, an aesthetic pattern of effects, rather than a sequence delivering a message.
The reader must attend to the literary contexts that enable a literary text to deliver surprise by means of marked effects, and the reader must view the entire poem as a completed whole, a pattern whose function is to deliver that surprise. The surprises of literary texts form aesthetically satisfying patterns, and the reader's discovery of a satisfying pattern is, itself, the major pleasurable surprise that reading and interpreting a literary text provides. These two principles, context and pattern, can guide the new reader of literature from the outset. But it is the very nature of literary texts, because they draw upon literary contexts to condition markedness and form patterns of surprise, that reading leads to more reading, and more reading deepens and enriches the experience of rereading familiar literary texts, making them forever new. In a sense, reading has no beginning and no end. It has no end because the addition of more literary contexts will always enrich the experiences of markedness, pattern-formation, and surprise. And it has no beginning because one must always have some sort of prior knowledge, even if only rudimentary, to find pleasure in reading a literary text. Into this beginningless and endless labyrinth of texts, the teacher leads new readers, for literacy cannot be separated from pedagogy. The first step in this journey into the world of literary texts is to understand the difference between texts and normal communication, and how literariness conditions the proper approach to literature.
1. My reading of Blake's "The Tyger" follows, and develops, the one given by Bloom (1971:35-39).
Austin, Timothy R. 1981. Constraints on syntactic rules and the style of Shelley's 'Adonais': an exercise in stylistic criticism. In Freeman 1981a, 138-165.
Aviram, Amittai F. 1994. Telling rhythm: body and meaning in poetry. Ann Arbor: University of Michigan Press.
Blake, William. 1967. Songs of innocence and of experience. New York: Oxford University Press. (Facsimile reprint of Blake's 1789-94 ed., with intro. and notes by Geoffrey Keynes.)
Bloom, Harold. 1971. The visionary company. Ithaca: Cornell University Press.
Culler, Jonathan. 1981. Literary competence. In Freeman 1981a, 24-41.
Dickinson, Emily. The complete poems of Emily Dickinson. Ed. by Thomas H. Johnson. Boston: Little, Brown.
Epstein, E. L. 1981. The self-reflexive artefact: the function of mimesis in an approach to a theory of value for literature. In Freeman 1981a, 166-199.
Fairley, Irene R. 1981. Syntactic deviation and cohesion. In Freeman 1981a, 123-137.
Freeman, Donald C. (ed.) 1981a. Essays in modern stylistics. London and New York: Methuen.
Freeman, Donald C. 1981b. Keats's `To Autumn': poetry as process and pattern. In Freeman 1981a, 83-99.
Jakobson, Roman. 1960. Concluding statement: linguistics and poetics. Style in Language, ed. Thomas Sebeok, 350-77. Cambridge, MA: MIT Press.
— — -. 1968. Poetry of grammar and grammar of poetry. Lingua 21:597-609.
— — -. 1973. Principes de versification. Questions de poétique, ed. Tzvetan Todorov, 40-55. Paris: Seuil.
Kermode, Frank. 1960. Wallace Stevens. London: Oliver Boyd.
Keyser, Samuel Jay. 1981. Wallace Stevens: form and meaning in four poems. In Freeman 1981a, 100-122.
Kiparsky, Paul. 1981. The role of linguistics in a theory of poetry. In Freeman 1981a, 9-23.
Mukařovský, Jan. 1989. Standard language and poetic language. The Critical Tradition, ed. by David H. Richter, 860-868. New York: St. Martin's Press. (First published in 1948.)
Plato. 1982. Phaedrus. Plato in twelve volumes, Vol. I, ed. and trans. by Harold North Fowler, 405-579. Cambridge, MA: Harvard University Press. (Numbers and letters in brackets are the canonical form of citations to standard editions of Plato.)
Stevens, Wallace. 1954. The Collected Poems of Wallace Stevens. New York: Random House-Vintage.
Tarlinskaya, Marina. Rhythm and meaning: 'rhythmical figures' in English iambic pentameter, their grammar, and their links with semantics. Style 21(1):1-35.
This page last updated Friday, August 5, 2022 by the author — amittai dot aviram at gmail dot com. © 2001 by Amittai F. Aviram.