Metaphor and Figure-Ground Relationship: Comparisons from Poetry, Music, and the Visual Arts

by Reuven Tsur

January 11, 2000


The gestalt notion "figure-ground phenomenon" refers to the characteristic organization of perception into a figure that 'stands out' against an undifferentiated background. What is figural at any one moment depends on patterns of sensory stimulation and on the momentary interests of the perceiver. Figure-ground relationship is an important element of the way we organise reality in our awareness, including works of art. Poets may rely on our habitual figure-ground organisations in extra-linguistic reality to exploit our flexibility in shifting attention from one aspect to another so as to achieve certain poetic effects by inducing us to reverse the habitual figure-ground relationships. This flexibility has precedent in music and the visual arts. Works by Escher, Mozart, Beethoven, Dickinson, Sidney, Shelley and Beckett are examined.


    There was an old joke in Soviet Russia about a guard at the factory gate who at the end of every day saw a worker walking out with a wheelbarrow full of straw. Every day he thoroughly searched the contents of the wheelbarrow, but never found anything but straw. One day he asked the worker: "What do you gain by taking home all that straw?" "The wheelbarrows." This paper is about the straw and the wheelbarrow, about shifting attention from figure to ground or, rather, about turning into figure what is usually perceived as ground. We are used to think of the load as "figure"; the wheelbarrow is only "ground," merely an instrument. Our default interest is in the act, not in the instrument.

    One of Anton Ehrenzweig's central claims in his seminal book A Psychoanalysis of Artistic Vision and Hearing (1965) is that the contents of works of art is best approached in terms of psychoanalytic theory, while artistic form is best approached in terms of gestalt theory. He has most illuminating things to say on these both with reference to music and the visual arts. While I am not always convinced by his application of psychonalysis to works of art, I find his discussions of gestalts and gestalt-free elements most compelling and illuminating. Gestalt theory has been systematically applied to the visual arts by Rudolf Arnheim (1957), to emotion and meaning in music by Leonard B. Meyer (1956). Cooper and Meyer (1960) applied it to the rhythmic structure of music. One of the earliest, and perhaps the most important application of gestalt theory to literature is Barbara Herrnstein-Smith's mind-expanding book Poetic Closure (1968). During the past two and a half decades I devoted much research to poetic prosody; I have found that many of the aesthetically most interesting issues regarding poetic rhythm, rhyme patterns and stanza form can be understood only through having recourse to gestalt theory (e.g., Tsur, 1977; 1992: 111-179; 1998; Tsur et al., 1990, 1991).

    One of the most interesting issues in gestalt theory, both from the perceptual and the artistic point of view is what gestalt theorists called "figure-ground relationship." The Fontana Dictionary of Modern Thought provides the following definition:

figure-ground phenomenon. The characteristic organization of perception into a figure that `stands out' against an undifferentiated background, e.g. a printed word against a background page. What is figural at any one moment depends on patterns of sensory stimulation and on the momentary interests of the perceiver. See also GESTALT.

Look, for instance, at this "droodle," entitled "Four Ku Klux Klansmen looking down a well, seen from below." Here, obviously, the four triangular shapes with the pairs of elliptical dots in them are the figures; the white space between them is the ground. If, however, you shift attention to this white space in the middle, KuKluxKlan Droodle you will discover that it has the shape of a distinct cross which, in turn, will become the figure, relegating the triangular shapes into the background.

    In this respect gestalt theorists discovered that some of the commonsense perceptual phenomena are not at all to be taken for granted as they would appear to the man in the street. The gestaltists not only brought to attention a most interesting phenomenon, but also laid down rigorous rules that govern the perceptual organisation processes that result in figure-ground relationships. The better the shape, the more it tends to stand out as a figure (and there are rigorous principles that account for what makes a shape "better" or "worse"). In the droodle, for instance, the "triangles" are well-differentiated, closed shapes; the pairs of dots further differentiate them. The title given further reinforces our "interest" in them. The white space in the middle yields a sufficiently symmetrical and closed shape to become figure once your attention is shifted to to it. Having discovered this white cross, you find it difficult to suppress it. It seems to me that the key-term for the perceptual distinction between figure and ground is "relative differentiation." Partially overlapping good shapes both in music and the visual arts may blur each other so as to form a lowly differentiated background. Irregularly distributed lines and dots make a rather poor shape; but, as we shall see soon, when they occur on a shape, they render it more differentiated relative to other similar shapes and tend to shift it in the figure direction.

    There seems to be a crucial difference between the visual and the other modes of perception. "It is difficult, if not impossible, even to imagine a visual figure without also imagining the more continuous, homogeneous ground against which it appears. But in `aural space', in music, there is no given ground; there is no necessary, continuous stimulation, against which all figures must be perceived" (Meyer, 1956: 186).

Due to the absence of a necessary, given ground in aural experience, the mind of the listener is able to organize the data presented to it by the senses in several different ways. The musical field can be perceived as containing: (1) a single figure without any ground at all, as, for instance, in a piece for solo flute; (2) several figures without any ground, as in a polyphonic composition in which the several parts are clearly segregated and are equally, or almost equally, well shaped; (3) one or sometimes more than one figure accompanied by a ground, as in a typical homophonic texture of the eighteenth or nineteenth centuries; (4) a ground alone, as in the introduction to a musical work--a song, for instance--where the melody or figure is obviously still to come; or (5) a superimposition of small motives which are similar but not exactly alike and which have little real independence of motion, as in so-called heterophonic textures (Meyer, 1956: 186).1

    In literature the figure-ground phenomenon is not at all easy to track; and sometimes, as in music, it does not seem to exist. The sociolinguist William Labov (1972) made an ingenious attempt to utilise this notion with reference to narratives. He pointed out that certain grammatical forms tend to relegate descriptions to a static ground, while some other grammatical forms tend to "foreground" the action of the narrative, which is the perceptual figure. One of the sad results of Labov's technique -- for which he certainly cannot be blamed -- is that quite a few linguists and literary critics who ignore the gestaltist origin of these notions "diagnose" figure and ground in a text by applying Labov's grammatical categories quite mechanically, without asking whether a ground does exist at all in a given text. Labov's own practice is far from "labelling"; it is, indeed, highly functional: it uses linguistic categories to trace the transformation of experience into story grammar. It should be noted that Labov uses these notions in a much wider context, in which figure-ground relationship is only one, and not the most important, aspect. He is more interested in means for evaluating experience in story-telling. I shall not elaborate here on this matter.

    Such theorists as Arnheim, Meyer and Ehrenzweig make illuminating comments on the dynamics of figure-ground relationship in their respective art media. I have elsewhere discussed at considerable length the figure-ground relationship in a passage by Milton resulting from an interplay between prosodic and syntactic gestalts and gestalt-free elements in the "world stratum" of the work, in an attempt to account for what Ants Oras (1957) described as perceptual "depth" (Tsur, 1977: 180-189; 1991: 85-92). In prosodic and syntactic structures too, good gestalts, strong shapes, tend to yield figures; where strong shapes blur each other or interact with gestalt-free qualities, they tend to blend in a ground. In this paper I am going to explore the manipulation of figure-ground relationships in the represented (or projected) world.

    To indicate the flexibility of such phenomena in painting and music, I will quote here two short passages on colour interaction from another book by Ehrenzweig.

The incisiveness of form, such as the comparative sharpness of its outline, or its pregnant shape, or the conflict or parallelism between superimposed or juxtaposed forms and so on, can be summed up as qualities of a `good' gestalt. We can summarize therefore that colour interaction between figure and ground stands in inverse proportion to the good gestalt of the figure (Ehrenzweig, 1970: 172).

He also indicates how the same notes in music may oscillate between ground and figure. Meyer suggested above that the musical field can be perceived as containing "several figures without any ground, as in a polyphonic composition in which the several parts are clearly segregated and are equally, or almost equally, well shaped." Ground enters into the composition as soon as the demands of the italicised portion of this fornulation are considerably relaxed:

To the extent to which a musical note is fitted into a clean melodic `line' it is prevented from fusing into harmonic tone `colour'; conversely a strong chord will temporarily fuse the loose strands of polyphony into solid tone colour so that the separate melodic lines disappear altogether. I have mentioned that the ear constantly oscillates between the harmonic fusion and polyphonic separation of the melodic lines; this conflict between `form' and `colour' belongs to the very life of music. A harmonically too luscious piece will soon lose its impact if it is not poised against a tough polyphonic structure (Ehrenzweig, 1970: 173).

This typically unstable relationship between figure and ground in painting and music must be borne in mind when we discuss those relationships in the verbal arts.

Figure and Ground in Escher

Escher: Liberation

    Figure 1. Escher: Liberation

Figure 2. Escher: Woodcut II, strip 3.

    Such artists as M. C. Escher deliberately experiment with the figure-ground phenomenon in visual perception. Escher himself described the organizing principle of Figure 2 (and similar drawings) as follows:

In each case there are three stages to be distinguished. The first stage is the reverse of the final stage, -- that is, a white object on a black background as against a black object on a white background. The second stage is intermediary between the two, and is the true, complete division of the plane, in which the opposing elements are equal (Escher, 1992: 164).

One may observe three points in Escher's experiments relevant to our problem. First, though according to gestalt theory the better the shape the more it is likely to be perceived as "figure," sometimes the same shapes may serve as figure or as ground. Additional differentiating devices seem to be at work. Second, the dots and lines on some shapes seem to serve as such "differentiating devices" that may turn them into "figure" (some of the fish have lines on one fin, and no lines on the other, which turns them half figure, half ground, emerging, as it were, from "nowhere"). In one instance in figure 1, one single dot within the area (indicating, as it were, the eye of the "bird") slightly shifts the shape in the direction of "figure." Third, with very little conscious effort we can almost freely switch from one organisation to another, sending figure shapes to the ground and vice versa. Now, as I said, William Labov imported this distinction into linguistic description. He collected stories in Harlem by asking informants for a "memorable fight." Very much in harmony with common sense and intuition, he pointed out that, in these stories, certain grammatical forms tend to foreground information as "figure," and some tend to relegate it into the "ground." It would be worth one's effort to investigate, in a separate study, how the resulting mental images do or do not preserve the gestalt rules for visual perception. But this is a relative matter, and poets may turn the distinctions backside forward. At any rate, labelling objects as "agents" or "instruments" in isolation has nothing to do with the issue

Figure and Ground (?) in Emily Dickinson

    In a forthcoming paper, "Poetry and the scope of metaphor: Toward a cognitive theory of literature," Margaret H. Freeman discusses several poems by Emily Dickinson. In relation to one of them she discusses the nature of Time in language and poetry.

    How do we understand time? It is commonly understood in two ways, depending on figure-ground orientation. That is, we can perceive time as a figure with respect to some ground, as when we say "Time flies when we're having fun," where time is seen as passing quickly across some given fun-filled space. Or we can perceive time as the ground for the figure, as when we say "The train arrived on time." Both these ways of looking at time come from a very general metaphor in our thought processes: the EVENT STRUCTURE metaphor.  . . . For example, a very common metaphor for time is TIME IS A HEALER. This metaphor for time depends on the EVENT STRUCTURE metaphor which entails EVENTS ARE ACTION, which in turn entails time is an object. The EVENT STRUCTURE metaphor is shaped by the notion of causality, in which an agent is understood to bring about an event. Thus we say "Time heals all wounds." But Dickinson rejects this metaphor:

They say that "Time
assuages" -
Time never did assuage -
An actual suffering
As sinews do - with Age -

Time is a Test of
Trouble -
But not a Remedy -
If such it prove, it
prove too
There was no Malady -
        Fascicle 38, H 163, 942 (J 686)

    Let us confine our attention to the following aspect of Freeman's discussion of time as an "EVENT STRUCTURE metaphor": "we can perceive time as a figure with respect to some ground  . . . or we can perceive time as the ground for the figure." In harmony with our foregoing observations we might add that we can also perceive time as a figure with no ground at all. At any rate, such a conception as Freeman's is most promising from the cognitive or the aesthetic point of view, because it offers terms with sufficient descriptive contents that allow us to make significant distinctions within a poetic text. That is precisely what she does in her following suggestion: "She rejects the idea of time as an agentive figure working against the ground of suffering and replaces it by reversing figure and ground. In the second part of the poem, it is suffering or "trouble" that is perceived as the figure against the ground of time."

    During the past few months we have been engaged in a lively correspondence on the application of cognitive science to poetry. On this issue I wrote to Freeman, among other things, this:

    This is not quite accurate. In both instances ("Time is a healer of wounds" and "Time is a Test of Trouble") we have, in M.A.K. Halliday's terms, the same types of "relational clauses," of the "attributive" kind. "The relation is one of class membership" (in these relational clauses the less general term is more abstract, unlike in "Marguerite is a poet," where the two terms are of the same order of abstraction). In such relational clauses it seems to be impossible to make distinctions in terms of figure-ground relationship. In "Marguerite is a poet," "poet" is not exactly ground for "Marguerite": we do not perceive Marguerite in the front and "poet" in the background; we percieve "Marguerite as a poet," as one figure. If, however, one insists that "poet" is ground for "Marguerite," then both "healer of wounds" and "a Test of Trouble" are ground for "Time." Whether ground or not, they only differ in the degree of activity they attribute to the agent "Time." But, in principle, if the text permits, very significant distinctions can be made in terms of figure-ground relationship.

Freeman gave the following answer to this comment: "The two expressions are not the same types of clauses. Time is a healer of wounds identifies Time as an agent; Time is a test of trouble identifies it as an instrument."

    According to the conception outlined above, there is no reason on earth why an instrument should not be granted the status of a figure. The question is not whether Time is identified as an agent or an instrument, but what kind of attention it attracts. Consider the following four sentences: "Time is a healer," "Time assuages," "Time is a Test of Trouble," "Time is a Remedy." In all of them Time is in the focus of our attention, while the various predicates attribute to it some kind (or degree) of activity. Consider the following two sentences: "Time is a healer," and "Time is a Remedy." They both attribute to Time the same kind of activity: it heals; but "healer" is said to be an agent, whereas "Remedy" is an instrument. In fact, dictionaries define "healer" as "one who or that which heals" -- that is, one can hardly tell whether the word suggests an agent or an instrument. "Time heals" expresses by a straightforward verb an activity that is expressed by nouns in the other two sentences; one cannot tell, however, whether it is an agent or an instrument; that is, whether it heals as a physician or as a remedy. What is more, "Time assuages" (which is after all the phrase used in the poem), may be perceived as "comforting, soothing, lessening pain" more as an ointment than as a person. Only one thing is certain: that in all these sentences, and especially where we have their cumulative effect in the poem, "Time" stands out as a figure; and I am not sure that there is a ground there at all.

Activities and Noun Metaphors

    I will not discuss sentence types beyond what I have already said regarding their contribution to the figure-ground phenomenon. I will point out only one thing. According to Labov, verbs that express straightforward activity tend to foreground their referents as figures, whereas static nominal predicates tend to relegate their referents to the background. In Freeman's foregoing examples, for instance, the things that are moving ("Time flies," "The train arrived") are said to be figure, whereas the things that are presented as a state are said to be ground (even though "having fun" may involve many straighforward activities). Now in metaphoric expressions one cannot take this for granted: noun phrases become effective means for expressing straightforward activity. One of the most fruitful insights of Christine Brooke-Rose in her A Grammar of Metaphor is that noun metaphors are much more effective in conveying figurative activities than verb metaphors.

In other words, whereas the noun is a complex of attributes, an action or attribute cannot be decomposed. Its full meaning depends on the noun with which it is used, and it can only be decomposed into species of itself, according to the noun with which it is associated: an elephant runs = runs heavily, a dancer runs = runs lightly.

Leaving adjectives aside for the moment, this means in fact two things. On the one hand, verbs are a more flexible element of language as far as meaning is concerned: that is, since they change their meaning slightly according to the noun with which they are used, they can also quickly extend their meaning and seem natural with each noun, so that an originally metaphoric use may rapidly cease to be metaphoric if the verb can be used in too many different senses with different nouns. On the other hand, when a verb is metaphoric, its adaptability to the noun is so great that its relationship to it is direct, and much stronger than its relationship to the action it is "replacing." And it changes, by implication, that noun into something else (Brooke-Rose, 1958: 209).

    The following quotations, by contrast, are from her discussion of the "genitive link," but she claims to them a wider validity ("In a general sense, all noun metaphors are the result of some activity").

All Genitive relationships are activity relationships, as I shall explain later. The body is called the hostel of the heart because the heart dwells or lodges there. The point about this type is that the metaphor changes another noun (heart becomes a lodger) and at the same time replaces an unmentioned proper term (body) (Brooke-Rose, 1958: 149).

With of in other relationships, I have constantly stressed its verbal element:  . . . I have found that of can most successfully express the complete identity of the two linked nouns when the metaphor can very easily be turned into a verb: if love burns, it is a fire, if we give love, it is, if death overshadows, it is a shade, etc. In a general sense, all noun metaphors are the result of some activity, but this has to be strongly felt when the grammatical link is artificial: The fir of love (Tr. I/436); the shade of death (Am. 17); in flames of pure and chast desyre (Am. 22); the fyre of loue (Am. 81); thy gift of love (Donne 10) (Brooke-Rose, 1958: 155).

    As I have suggested above, in the sentences "Time is a healer," "Time assuages," "Time is a Test of Trouble," "Time is a Remedy" the verb predicate and the various kinds of noun predicates alike attribute some straightforward activity to Time, and present it as figure in the focus of attention. We cannot know from the text, in what way Time tests troubles: whether it actively puts troubles to a test, or merely turns blue in bases and red in acids as the litmus paper. Nor can we know whether troubles are static as alkaline solutions and acids to be tested by some "litmus paper," or are more active. One thing seems to be quite certain however: that the testing Time is the figure, and troubles are merely ground at best or, perhaps, part of the figure=action. Our attention is focussed on Time. In Emily Dickinson's poem, "reified" Time is perceived as a figure, whether as an agent or an instrument. The figure-ground phenomenon is relational, and no labeling of isolated parts can illuminate it in any way.

Figure-Ground Reversal in the "Moonlight" Sonata

    This section is devoted to the problematic (or flexible) relationship of figure and ground in music. All through the opening movement of Beethoven's "Moonlight" Sonata there is a series of obsessive rising sequences of three notes, as in music excerpt 1. In the course of writing this paper I compared a wide range of performances of Beethoven's Sonata. Eventually I decided to quote here two of them, of unequal fame, by Alfred Brendel and Dubravka Tomashevich, a student of Rubinstein's, because they illustrate most clearly the contrast which I want to bring out: that the performer has considerable control over presenting the triplets as ground or as figure.

Music Excerpt 1   Compare Alfred Brendel's performance (a) to Dubravka Tomashevich's performance (b) of excerpt 1.

a.  Brendel playing excerpt 1 b. Dubravka playing excerpt 1

    Marcia Green drew my attention to a remarkable similarity between this passage and a passage in Don Giovanni: in the short trio of the three basses, Don Giovanni, Leporello and the dying Commandatore ("Ah, soccorso! son tradito!"), the orchestra plays exactly the same kind of repeated rising series of three notes. Here, however, it is deeply buried in the "ground," and even after repeated listenings I could only vaguely discern a dim um-pa-pa in the background, as in music excerpt 2.

Figure 3. The first two bars of the triplets in the Don Giovanni trio.

Figure 4. The first four bars of the triplets in the Moonlight Sonata

Music Excerpt 2.
Now, listen first to the Don Giovanni trio in Klemperer's recording. Second, listen to the triplets at the beginning of the same when the midrange is overemphasized. Third, listen to a piano extract of the triplets alone, played by Mira Gal.

Klemperer recording Same with mid-range emphasized   Piano Extract   

I will not quote here Green's interpretation of this similarity, because for me the most important part of the comparison is that Beethoven took a piece of ground music, that has a typically ground texture, and placed it in the focus of the sonata movement dominating for no less than six minutes the musical space. I had a long dialogue on this issue with Harai Golomb, professor of literary theory, theater studies, and musicology, who is certainly much more competent on music theory than myself, and I could not reach the ensuing conclusions without his insightful help. We both agree that Beethoven did not "imitate" the triplets from Mozart, and that this similarity does not indicate any significant relationship between them, either as composers or as persons (as Green would have us think). So, what is the point in pointing out the similarity besides the sheer piquancy of the comparison? Eventually I came to the conclusion that an analogy with intertextuality in literature may illuminate the issue.

    The juxtaposition of the two works foregrounds the different character of the two applications of the same technique. The similarity of Beethoven's triplets to Mozart's which in the "Moonlight" Sonata, in contrast to Don Giovanni, are in the focus of the listener's attention, foregrounds the difference between them. Golomb agrees that there is in the sonata a distinct, monotonous, repeatedly rising ta-ta-ta sequence. This sequence is exceptionally boring from the rhythmical point of view, resembling the typical " ground" texture in the Don Giovanni excerpt, and many other works.

    At the same time, the magic of the movement is due, he says, to tensions and resolutions in the harmonic structure of the whole, both in the sequence of triplets and the interplay of the various simultaneous melodic threads. There are three simultaneous threads in this movement, in, roughly speaking, the high range, the midrange and the low range. The afore-said triplets constitute the middle thread in this complex. There are the lower harmonic chords which, we both agree, generate a ground of tensions and resolutions, making a major contribution to the affective impact of the movement; and there is a higher sequence of longish notes, tam-ta-tammmm, which add up to a mildly rising and falling melody, which is the real figure of the movement. This melody, he says, though considerably diffuse, is more differentiated than the obstinately repeated ta-ta-ta series (as in excerpt 3). What is more, there is, from time to time, a "dialogue" between the highest and the lowest thread, skipping, as it were, the middle thread.

    Consider the following excerpt. In excerpt 3, as opposed to excerpt 1, all three threads are present.

Music Excerpt 3   Compare Alfred Brendel's performance (a) to Dubravka Tomashevich's performance (b) of excerpt 3.
a.  Brendel b. Dubravka

Now one thing appears to be quite certain. This dialogue does not turn the lowest thread into figure; it remains ground relative to the other two threads.

    In harmony with my argument in the present paper, I suggest that our comparison of excerpts 1 and 3 shows here, too, that figure-ground relationships are not determined once for all in all circumstances. As we have seen, "what is figural at any one moment depends on patterns of sensory stimulation and on the momentary interests of the perceiver."

    My point is that in the case of a musical performance, "the momentary interests of the perceiver" can be manipulated to a considerable extent by the performer, by rather evasive cues: in different performances different threads of the "patterns of sensory stimulation" may be foregrounded, by mild shifts of attention to and fro, as, in the visual arts, in, e.g., the Escher drawings. My own view of the passage may have been influenced to a considerable extent by Alfred Brendel's performance on Philips 438 730-2. In this performance, the middle thread is somewhat louder relative to the other threads than in some other performances. As a result, the highest thread (as well as its dialogue with the lowest thread, when perceived) is perceived as an intrusion into the "figure," the middle thread. This intrusion, in turn, will increase the sequence's tendency to reassert its integrity -- according to the gestalt assumption that a perceptual unit tends "to preserve its integrity by resisting interruptions." In this instance, the perception of figure-ground relationships can be further manipulated by the treatment of the longish notes of the highest thread. If their differentiation and connectedness into a melody is emphasized in the performance, they will attract attention as figure; if they are presented as more discrete notes, they will be perceived more as events intruding upon the middle thread, the rising sequences of three notes.

    My purpose here is not to offer a systematic comparative research of performances. What I want to emphasize is this: in Brendel's performance (more than in Tomashevich's), the highest thread is perceived more as a series of irruptions than as a melodic line. This is due to two features of the performance. First, in 3a the second thread is louder relative to the other two threads than in 3b; and secondly, Brendel performs the highest thread in a peculiar way. Compare excerpt 3a to 3b. In the highest thread, we hear twice a group of tam-ta-tammmm on the same note, followed by a slightly higher one. Owing to amplitude dynamics and Brendel's "pianists' touch," this higher note is perceived as exerting a greater effort to intrude rather than as contributing to a continuous melodic line. The result is monotonous and exceptionally dramatic at the same time.

Brendel envelope plot
Figure 5. The envelope plot of music excerpt 3
in Brendel's performance

Dubravka envelope plot

Figure 6. The envelope plot of music excerpt 3
in Tomashevich's performance

It is illuminating to consider the amplitude dynamics of the two performances in this excerpt. Figures 5-6 show the plot of amplitude envelope of the first tam-ta-tammm group in the two performances of excerpt 3. In these graphs, the lower window presents the wave plot display which shows a plot of the wave amplitude (loudness) as a function of time (in milliseconds). The upper window contains the envelope curve, representing the outer bounds of the soundwave's amplitude, created by connecting the peaks. Both windows present the relative amplitude of the sound. The greater the amplitude, the wider the plot (in the lower window), or the higher the plot (in the upper window).

    The three notes are of equal pitch, but, in Brendel's performance, each one of them begins with a distinct obtrusion of the amplitude envelope. In Tomashevich's performance, by contrast, the first two notes slightly fluctuate at a low level, and are followed by a third note of disproportionately great amplitude. Add to this that, although both performances are "overdotted," the duration of the middle note in Tomashevich's performance is shorter: 273 msec, as opposed to Brendel's 296 msec. As a result, in Tomashevich's performance the first two notes are subordinated to the third one. One perceives the middle note more as a "passing note," leading forward to the third note. This construction tends to merge the three notes into one melodic line. In Brendel's performance, by contrast, the three notes are perceived as more discrete; they have relatively greater perceptual separateness. The middle note is perceived not only as a note in its own right, but also as more grouped with the preceding one. Translating Lerdahl and Jackendoff's transformational terminology into plain English, backward grouping generates tension, forward grouping -- relaxation.

    Julian Haylock, who wrote the music notes for the sonatas on Alfred Brendel's CD, suggested, quite impressionistically, what the perceived effect of all this is: "The opening Adagio Sostenuto  . . . is quite unlike anything previously composed for the keyboard," and he speaks of "its dream-like texturing" which is, in this case, certainly, the artistic purpose of promoting a typical background texture to the status of a figure or, at least, of causing it to dominate a full-length sonata movement. According to Meyer, as we have seen, "the musical field can be perceived as containing a ground alone, as in the introduction to a musical work -- a song, for instance -- where the melody or figure is obviously still to come." It is the typical background texture pushed into the foreground throughout a full movement that is "quite unlike anything previously composed for the keyboard"; and this is also the basis for "its dream-like texturing" -- reinforced by its interplay with the other two threads, as discussed above.

Figure-Ground Reversal in Literature

    We have seen in Escher's drawings that they grant the perceiver considerable freedom to foreground certain shapes as figure or relegate them to an undifferentiated background. Such an "aspect switching" requires only minimal mental effort. Escher discusses at some length what kinds of shapes allow such flexibility of perception. He does not discuss the means by which he tilts the perceiver's inclination in one direction or the other. I have suggested that when the same closed area is repeated, lines or dots on it tend to bestow on it differentiation and induce us to perceive it as a figure; their absence, as ground. I have also suggested that the perceptual apparatus can easily overcome these "directive" means, by some conscious effort. Likewise, in Beethoven's sonata we have seen that the performer may manipulate the listener's perception of figure-ground relationships by connecting the notes of the higher thread into a perceptible melody, or leaving them as discontinuous, solitary events. Here the listener is more at the performer's mercy, and "aspect switching" requires greater mental effort. In what follows, I will consider three literary texts that exploit this readiness of human perceivers to switch back and forth between figure and ground. All three texts achieve their effect by inducing readers to reverse figure-ground relationships relative to their habitual modes of thought or perception.

    Consider the following poem by Shelley:

      A Song

A widow bird sate mourning for her love
  Upon a wintry bough;
The frozen wind crept on above,
  The freezing stream below.

There was no leaf upon the forest bare,
  No flower upon the ground,
And little motion in the air
  Except the mill-wheel's sound.

In aural perception, irregular noises are usually dumped into the background. But when Shelley ends his "Song" with these two lines, he turns into figure a percept that most commonly is dumped into the ground. And this is enormously effective here. I have elsewhere discussed this poem at some length. Here I will reproduce only part of my discussion of the last two lines. They have a rather complex function within the whole. Little as a part of the sequence There is no ... No... And little... suggests "none at all"; in this sense, "And little motion in the air" is one more item in the list of analogous items suggesting a theme of deprivation. In this sense, it seems to herald an unqualified statement that generates a psychological atmosphere of great certainty.

    The subsequent preposition except, however, makes a substantial qualification to this statement, substituting "a very small amount of" for total exclusion; that is, there is an exclusion from the total exclusion: a mill-wheel's sound. The relation of the mill-wheel to its sound is like the relation of a thing to a thing-free quality. What seems to be emphasised by this phrasing is that only the thing-free quality, the sound, but not the thing itself, the mill-wheel, is introduced into the description. This perturbation of the air becomes another item in the list of items with reduced activity; by the same token, it foregrounds the presence of the air, the thing-free quality par excellence pervading the scene. This shift of the meaning, qualifying the unqualified statement, performs a "poetic sabotage" against the determined, purposeful quality of the poetic closure, replacing the psychological atmosphere of great certainty with a psychological atmosphere of uncertainty, contributing to the emotional quality of the poem. This emotional atmosphere has been generated by the abstraction of certain qualities from parallel concrete items in the description. Both the emotional quality and the "poetic sabotage" of closure are reinforced by another aspect of the mill-wheel's sound, which I wish to point out through an idea borrowed from Joseph Glicksohn.

    Gestalt psychology speaks of figure-ground relationship. The mill-wheel's sound typically serves as ground to some aural figure. By forcing to the reader's attention a percept that typically serves as ground, the poem increases the emotional quality of the perception, and emphasises that there is no figure to be contemplated, reinforcing the quality of deprivation. Thus, the poem ends with "a ground alone, as in the introduction to a musical work  . . . where the melody or figure is obviously still to come." When it occurs at the end of a work, its lack of progress does not prepare for some thing to come as in the introduction to a musical work, but suggests some disintegration: the poem does not end, it passes out of existence, fades away.

    The next two examples can be regarded as displaying different degrees of one kind. Consider the following sonnet by Sir Philip Sidney:

Leave me, O love which reachest but to dust;
And thou, my mind, aspire to higher things;
Grow rich in that which never taketh rust,
Whatever fades but fading pleasure brings.
Draw in thy beams, and humble all thy might
To that sweet yoke where lasting freedoms be;
Which breaks the clouds and opens forth the light,
That doth both shine and give us sight to see.
O take fast hold; let that light be thy guide
In this small course which birth draws out to death,
And think how evil becometh him to slide,
Who seeketh heaven, and comes of heavenly breath.
   Then farewell, world; thy uttermost I see;
   Eternal Love, maintain thy life in me.


I have elsewhere discussed the light imagery of this sonnet at considerable length (Tsur, 1998). Here I will focus on just the third quatrain.

    Let us work out the internal logic of this image, in terms of mental habits and their manipulation by literary means. I will argue that the central device of this passage is a reversal of figure-ground relationship. But before discussing that, I wish to examine this passage in light of what Kenneth Burke calls "Scene-Act Ratio" and "Scene-Agent Ratio." In these Ratios "Scene" typically serves as ground to "Act" and "Agent," which are, typically, the figure. Burke proposed to analyse human motives and actions in terms of the "dramatic pentad": Act, Scene, Agency, Agent, Purpose.

Using "scene" in the sense of setting, or background, and "act" in the sense of action, one could say that "the scene contains the act." And using "agents" in the sense of actors, or acters, one could say that "the scene contains the agents."

    And whereas comic and grotesque works may deliberately set these elements at odds with one another, audiences make allowance for such liberty, which reaffirms the same principle of consistency in its very violation.  . . . In any case, examining first the relation between scene and act, all we need note here is the principle whereby the scene is a fit "container" for the act, expressing in fixed properties the same quality that the action expresses in terms of development (Burke, 1962: 3).

In the case of Sidney's poem, the scene and the act define the nature of the agent as well as his purpose: the Soul comes from heavenly breath and goes to (seeketh) heaven; according to Burke, this is a way to say in spatial and temporal terms that the Soul is (in the present) of a heavenly essence ("temporization of the essence"). George Lakoff and his disciples would speak here of the event structure metaphor "PURPOSEFUL ACTION IS A JOURNEYy"; the purpose of the action is expressed, very much in Burke's spirit, by the place   to which the journey leads. A more specific instantiation of this metaphor is "LIFE IS A JOURNEY."2

    Now consider this: in this poem, the purpose of the journey is presented by two different ends: "Who seeketh heaven, and comes of heavenly breath," and "In this small course which birth draws out to death." These two destinations have opposite implications. One presents "Life as full of meaning"; the other presents "Life as totally meaningless." There is all the difference if "this small course" leads to death or to heaven.

    Particular occasions of birth and death in everyday life are perceived as figures, and life only as ground, at best. But when we speak of Human Life, Life becomes the figure, only marked at its extremes by birth and death, which thus become ground. In Christian religious traditions Life is only a transient episode for the soul which "seeketh heaven, and comes of heavenly breath." Religious rhetoric frequently attempts to bring man to an insight into this truth by using paradoxical epigrammatic phrasings (such as "Whosoever will save his life shall lose it" -- Mark 8.35). Religious poetry may attempt to do this by a sudden shift of attention from the habitual figure to its ground, the markers of its extremes: Sidney gently manipulates attention from "this small course" to "birth" and "death," which are only meant to mark the extremes of life.3

     Now notice this. Purpose is not absent from the image

                  let that light be thy guide
In this small course which birth draws out to death,

it is only translated into a different visual terminology. In my paper on the cognitive structure of light imagery in religious poetry I discussed this poem at great length. I pointed out a wide range of meaning potentials in the light image, many of which are exploited in this poem. One of them is related to Lakoff's conceptual metaphor "KNOWING IS SEEING": Light gives instructions, shows the way. Another one is derived from the fact that the Light comes from an invisible and inaccessible source in the sky. Thus, these two lines do not express life's purpose by a place that serves as the destination of the journey; but this purpose is reintroduced by another conventional metaphor: light as knowing, understanding, or proper guidance.

    The same figure-ground reversal is brought to an absurd extreme in the following quotation from Beckett's Waiting for Godot:

Astride of a grave and a difficult birth. Down in the hole, lingeringly, the gravedigger puts on the forceps.

The tramp Vladimir sharpens Sidney's inverted image to absurdity: Man passes straight from the womb to the tomb, assisted by the gravedigger's forceps. In a world in which "God is dead," there is nothing beyond, and what is in between is meaningless and negligible. The emotional disorientation aroused by this understanding is reinforced by the grotesque image. In our everyday perception, birth is the beginning of life; death its cessation. What matters is life itself. Both in Sidney's and Beckett's image the two extremes, birth and death, or the womb and the grave become the figure; what is between them (life!) serves only to connect them. And the shorter the connection, the more meaningless life becomes.

To Sum up

    Figure-ground relationship is a very important notion of gestalt theory. Theorists of the psychology of music and the visual arts made most significant use of it. The significance of this notion in literary theory is rather limited. The most important attempt to import this distinction to linguistics and literary theory is William Labov's. Unfortunately, some linguists and literary critics regard Labov's work as a model for technical exercises rather than a source of insights into some significant part-whole relationship. The main point of this paper has been that such grammatical terms as "agent" or "instrument" are not foolproof diagnostic tools, and that figure-ground relationship is an important element of the way we organise reality in our awareness, including works of art. Here I have focussed attention on figure-ground relationships in extralinguistic reality as reflected in poetry. I argued that poets may rely on our habitual figure-ground organisations in extra-linguistic reality, and exploit our flexibility in shifting attention from one aspect to another so as to achieve certain poetic effects by inducing us to reverse the habitual figure-ground relationships. This flexibility has precedent in music and the visual arts. I have examined three instances from three literary masterpieces. An important concomitant of these close readings was to demonstrate that in most instances one may not only identify these reversals in the text, but may also suggest their effects. In Sidney's poem and the excerpt from Beckett the resulting "message" could be paraphrased in a straightforward conceptual language. But this is quite misleading. What is important here is not so much the "message" conveyed, but the insight resulting from the shift of mental sets. In Shelley's poem, the conceptual "message" diminishes to a minimum, and the main effect of the reversal is an intense perceptual quality that can only be approximated by such descriptive terms as "uncertainty, purposelessness, dissolution, wasting away."

This research was partly supported by the Israel Science Foundation.


1. Meyer uses the term "ground" within gestalt theory. It should be noted that in music theory this term has a technically defined, somewhat different sense. This difference of senses may be one source of the disagreement reported below between Harai Golomb and myself. [back]  

2. It is quite characteristic of present-day critical vogues that referees of my papers frequently suggest that in some place or other I might mention Lakoff's work; but so far they have never suggested Burke. [back]

3. The changing relationship between shapes and their edges as figure-ground relationship is well brought out by the following two locutions concerning geographic configurations: with reference to the U.S., the phrases "West Coast" and "East Coast" foregound the dry land between them as figure, the water being part of the ground. With reference to the Middle East, the phrases "East Bank" and "West Bank" foreground the water between them as figure, the dry land being part of the ground. For political reasons, the dry land of "the West Bank" has now become figure in its own right. [back]


Arnheim, Rudolf (1957). Art and Visual Perception. London: Faber & Faber.

Brooke-Rose, Christine (1958). A Grammar of Metaphor. London: Secker & Warburg.

Burke, Kenneth (1962). A Grammar of Motives and A Rhetoric of Motives (in one volume). Cleveland & New York: Meridian Books.

Cooper, C. W. and L. B. Meyer (1960). The Rhythmic Structure of Music. Chicago: Chicago UP.

Ehrenzweig, Anton (1965). The Psychoanalysis of Artistic Vision and Hearing. New York: Braziller.

----- (1970). The Hidden Order of Art. London: Paladin.

Escher, M. C. (1992). "The Regular Division of the Plain," in J. L. Locher (ed.), M. C. Escher -- His Life and Complete Graphic Work. New York: Harry N. Abrams Inc. Publishers. 155-172.

Freeman, Margaret H. (in press). "Poetry and the Scope of Metaphor: Toward a Cognitive Theory of Literature." To appear in Metaphor & Metonymy at the Crossroads. Ed. Antonio Barcelona. Mouton de Gruyter 1998.

Halliday, M. A. K. (1970). "Language Structure and Language Function," in John Lyons (ed.), New Horizons in Linguistics. Harmondsworth: Penguin Books. 140-165).

Herrnstein-Smith, Barbara (1968). Poetic Closure. Chicago: Chicago UP.

Labov, William (1972). Language in the Inner City: Studies in the Black English Vernacular. Philadelphia: University of Pennsylvania Press.

Lakoff, George (1993). "The Contemporary Theory of Metaphor," in Andrew Ortony (ed.,) Thought and Metaphor. Cambridge: Cambridge UP (second ed.). 202-251.

Lerdahl, Fred, and Ray Jackendoff (1983). A Generative Theory of Tonal Music. Cambridge Mass.: MIT P.

Meyer, Leonard B. (1956). Emotion and Meaning in Music. Chicago: Chicago UP.

Oras, Ants (1957). "Spenser and Milton: Some Parallels and Contrasts in the Handling of Sound," in Northrop Frye (ed.), Sound and Poetry. New York: English Institute Essays. 109-133.

Tsur, Reuven (1977). A Perception-Oriented Theory of Metre. Tel Aviv: Porter Institute for Poetics and Semiotics. 244 pp.

----- (1992). Toward a Theory of Cognitive Poetics. Amsterdam: Elsevier (North Holland) Science Publishers. 580 pp.

----- (1998). "Light, Fire, Prison: A Cognitive Analysis of Religious Imagery in Poetry." PSYART: A Hyperlink Journal for the Psychological Study of the Arts, article 980715. Available HTTP:

----- (1998). Poetic Rhythm: Structure and Performance -- An Empirical Study in Cognitive Poetics (in English). Bern: Peter Lang.

-----, Glicksohn, Joseph, & Goodblatt, Chanita (1990). "Perceptual Organization, Absorption and Aesthetic Qualities of Poetry." László Halász (ed.), Proceedings of the 11th International Congress on Empirical Aesthetics. Budapest: Institute for Psychology of the Hungarian Academy of Sciences. 301-304.

-----, Glicksohn, Joseph, & Goodblatt, Chanita (1991). "Gestalt Qualities in Poetry and the Reader's Absorption Style." Journal of Pragmatics 16 (5): 487-504.

© 2000, Reuven Tsur, all rights reserved.

To cite this article, use this bibliographical entry: Reuven Tsur "Metaphor and Figure-Ground Relationship: Comparisons from Poetry, Music, and the Visual Arts". PSYART: A Hyperlink Journal for the Psychological Study of the Arts. December 15, 2009. Available January 11, 2000 [or whatever date you accessed the article].
Received: February 1, 2000, Published: January 11, 2000. Copyright © 2000 Reuven Tsur