I know no better way to place a frame around the things I would like to say in this paper than to quote the opening words of a beautiful article by Roman Jakobson and Stephen Rudy - "Yeats' 'Sorrow of Love' Through the Years" -
Paul Valéry, both a poet and an inquisitive theoretician of poetry as an 'art of language,' recalls the story of the painter Dégas, who loved to write poems, yet once complained to Mallarmé that he felt unable to achieve what he wanted in poetry despite being 'full of ideas.' Mallarmé's apt reply was: "Ce n'est point avec des idées, mon cher Dégas, que l'on fait des vers. C'est avec des mots."
IT IS NOT AT ALL WITH IDEAS, MY DEAR DÉGAS,
THAT ONE WRITES LINES OF POETRY.
IT IS WITH WORDS.In Valéry's view Mallarmé was right, for the essence of poetry lies precisely in the poetic transformation of verbal material and in the coupling of its phonetic and semantic aspects.
What follows will be an exploration of the notion of "poetic transformation," one by a person struck by the contrast that Mallarmé was able to put so beautifully pithily. What exactly does it mean to write with words ?
To give you some idea of what this perhaps curious phrase might mean, I would like to examine the structure of Carl Sandburg's beautiful short poem Fog , viewing the poem as growing organically out of the particular kind of gesture we make with our throat and mouth when we say the title word. My aim will be to demonstrate that this gesture, or dance for our vocal tract, is a kind of verbal music, a one-syllable melody, which runs through the poem and provides coherences for it in just the same way that the famous da da da DOMMM of Beethoven's Fifth runs through that great symphony. I would also like to help you see why it can be valuable to work towards such a microscopic understanding of the music of a poem, for this music is not a background against which we are to contemplate the ideas of the poem. Rather, this musical structure, this phonetic architecture, is the idea. The idea and the patterns it makes on the page and in our minds are each other.
I hope that as the drama of this poem, this writing with the word fog , unfolds, it will become increasingly clear why poems in other languages, which might seem to be "about" fog, must necessarily link us to radically different experiences. Or to put it a bit more precisely, they must link us in such different ways that we can only wonder whether the "same" experiences are being evoked. For instance, any poems whose central words were brume, Nebel, neblina, or kiri, which are the words for "fog" in French, German, Portuguese, and Japanese, respectively, would be radically different, precisely because these other words offer such different melodic possibilities. To write a poem with any one of the five words for fog that we have mentioned, a great poet must be able to sense the musical possibilities of each word and link these to aspects of the image, concept, or experience of fog. When all of these aspects interconnect, harmonize, merge - "rhyme" - then a poem with the brilliance of Sandburg's arises.
I think that I have cleared my throat just about long enough. Let us look at the resonances that Carl Sandburg set in motion with his twenty-two words.
Fog
The fog comes
on little cat feet.
It sits looking
over harbor and city
on silent haunches
and then moves on.
Carl Sandburg
Before we start with the detailed process of what my poet-friend Jack DeWitt calls "ripping a poem to pieces," which may call up emotions in you ranging from puzzlement through boredom to outrage against such literary sacrilege, let me state at the outset that I love this poem very much, despite having already ripped it into as many pieces as I have been able to, and being bent on continuing the rippage. My stance towards you is that I want to share with you the joy that this verbal song evokes in me, and also the thrill of following a very elusive intellectual spoor. For we are at the same time confronted with the cameo beauty of Sandburg's words and with a puzzling fact.
I think that probably very many of us who went to school in the United States read this poem at one time or another in our schooling. We had to. This poem, like probably hundreds of others, was assigned - it is one of the famous American poems. Many of us dutifully plodded through our homework, and read at least some of the assigned poems, probably this one among them. It was a short one.
Now let me ask you to perform an experiment with me. Whether or not you have ever read the poem before, don't read it again just now. Cover it up with something, or look away from it. I have a question for you.
Are there any lines in the poem that you remember?
Moreover, if today was not the first time that you read this poem, did you, by any chance, happen to remember any of its lines even before rereading the poem? Are you one of the many people who, without any effort or intention at all, have the following line, out of all the thousands of lines that they dutifully looked at, while plodding through the evening's assignment, etched into your memory?
on little cat feet
My experience has been that of the probably by now hundreds of people who I have told that I am working on Sandburg's poem about fog, many, many have said something like, "Oh yes, 'The fog comes (in)/on little cat feet'. That one." The first line they don't always get letter perfect, but the second line, they have gotten down cold.
I submit to you that whether you were one of those unwitting memorizers or not, the existence of so many of them, in the face of a pretty generally deadening encounter with poetry at school, poses us a real riddle. Namely,
What did Sandburg do that was (and is) so right?
There is some magic in those four words, and I want to know where it comes from. I know that probably many of you will want to say something like, "Well, it is the rightness of the image, the aptness of the comparison. After all, cats and fog are alike in being able to move quietly and fluidly, and also in being shrouded in mystery. Sandburg just saw well." And I don't want to take issue with you here - for I agree entirely. Poets, like all artists, need great eyes - they have to be able to perceive similarities that others may have been asleep to. But that is not enough. For let me ask you how many of you misremembered the second line? How many thought it was either one of these?
on little cat paws
on small cat feet
I bet that it was very few indeed. These are, if you will pardon my bluntness, crummy poetry. And yet their crumminess cannot be said to derive from their poor imagery, for the image, and the comparison, the perception of hidden similarity, that give on little cat feet its memorability, are the same. Rather, what distinguishes Sandburg's indelible four words from my two lousy four-word "translations" of them is only the differences between their musics. For Sandburg not only had great eyes, he had great ears. And the secret of poetry is to be found somewhere in a country where "sight" and sound fuse. I would like to invite you to join me in a safari to that country.
**********
To start with, I would like to ask you to make an assumption about poetry that may initially strike you as very unlikely:
In the musical fabric of a poem, every sound has its
importance,
its very own part in the melody of the poem as a whole.
That is, I suggest that we see the individual sounds in the language of the poem as being a bit like the individual instruments in a symphony orchestra. And just as the strings have different kinds of music to make than do the brasses, and the brasses than the drums, so also the vowels in a poem will have different functions than will the fricatives (the sounds that make friction in the air - that hiss or hush or buzz - like f,v, s, z, sh, th), and the fricatives than the stops, like p,b,t,d,k, and g. The poet's job is to take the words that are involved in the experience that the poem evokes, and to listen to the kinds of sound colors that are provided in the little melodies that are made by the particular phonetic "instruments" that these important words contain. And then to paint with just these colors.
To give a very quick example, let us look at the English word fog. This word starts with the fricative [f], which is then followed by the long vowel that phoneticians write as [ø: ] (the colon is what phoneticians use to indicate that a vowel is long), with the syllable being closed by the stop [g]. If we compare the acoustic event of this word - a sequence of fricative-long vowel-stop - with the event suggested by the Japanese word for fog, kiri [kíDi] (which sounds more like what we would write as keedy than about anything else we could suggest in the English writing system), we see that the Japanese melody is very different from the English one. The Japanese one consists of two very short open syllables, which rhyme. By contrast, the English melody consists of one much slower, closed, syllable, with a "large" vowel (note how we have to open our mouths wide to say fog , while our mouths are nearly closed when we say kiri). The English syllable is brought to an abrupt finish by the closing stop [g].
Armed with these initial quite obvious observations about the music of the word fog, let us return for a look at the conceptual structure of Sandburg's poem. We see first that it has three clause-like "sections," A, B, and C, one for each of the poem's tensed verbs.
A. The fog comes
on little cat feet.B. It sits looking
over harbor and city
on silent haunchesC. and then moves on.
We can see that there are many ways in which these three sections resemble each other. Each has just one "finite," or tensed, verb in its first line, and each of these verbs is a monosyllable. The subject of each of these three verbs is the fog . In C, this is not true visually, for the subject of moves has been elided, but it is true notionally nonetheless. Furthermore, each of the three sections also ends the same way - with an on-phrase. The first two on-phrases are even more similar to one another, for each has, as the object of on, a plural noun phrase referring to a body part of the cat. The last on is followed by nothing - no cat body part. Why?
The cat has left!
That is also why the subject of the last verb has been deleted - when the fog/cat moves on, no part of it is left in the sentence.
This ABC sectioning is of course a structural analogy to one way of perceiving an experience of fog - as consisting of the fog's approach, then the state of being fogged in, and finally the fog's departure. The number of lines that Sandburg allots to each of these three parts of the fog experience is congruent with the following kind of subjective time for each.
Arrival |
State of being |
Departure |
of fog |
fogged in |
of fog |
Medium duration |
Longest duration |
Shortest duration |
2 lines |
3 lines |
1 line |
Most readers seem to agree on this ordering of the lengths of each of the three sections of an experience of fog - that the state of being fogged in is clearly the longest, since we never know when it is going to end, and that the arrival of the fog is typically slower than its departure. After all, we can say, "the fog lifted," while I have never heard anyone say, "the fog dropped."
What is relevant for our present concerns - namely, the way the music of a word interpenetrates with and creates the structure of any poem that it appears in - is evident when we ask how we perceive the durations of the three-part phonetic sequence in fog : fricative - long vowel - stop. Clearly, the vowel is the longest, with the fricative being longer than is the stop. So A-B-C "rhymes" with the three-part music of fog , as we can see if we map the phonetics of fog onto the above diagram of the subjective time of a fog experience.
Arrival |
State of being |
Departure |
of fog |
fogged in |
of fog |
Medium duration |
Longest duration |
Shortest duration |
2 lines |
3 lines |
1 line |
[f] |
[ø : ] |
[g] |
The poet slows down the middle (stative) section, B, not only by allotting more lines to it than to any other section, but also by putting most of the poem's long words in B. The poem begins and ends in the same way - with words of one syllable. And in the last line of the first section, we find just one bisyllable - little , which serves as a transition to the slow section, B.
Every line of B has a bisyllable at the end of it - and it is only this section that has long words at the ends of lines. Section B has all of the long words except the one in section A, which by chance (?) is little. And in B, there is just one line with more than two long words in it - line 4. This line is a special one conceptually, too: the two words harbor and city are the only ones in the poem that allude to the world of human beings.
An interesting question to ask ourselves is this: what would line 4 have been like if these two human-linked nouns had appeared in the reverse order? I think that we can agree that if Sandburg had written over city and harbor, we would have concluded that the fog was moving in the opposite direction - that it had originated inland, and was moving off out to sea. But in our poem, this is not the case. This fog comes in from the vastest thing on the surface of the planet - from the ocean, from the depths. It may arrive on "little" feet, but it grows tall. It rises high enough to cover the harbor - that human institution whose function is to serve as a haven for the brave boats that dare to traverse the Vastness. It continues to rise until it completely covers, and holds in its grip, humanity's proudest construction - the City. The city is where the illusion is strongest that there is safety in numbers, where we humans like to think that we have mastered the environment completely.
At the very end of line 4, the poem's longest line, whose seven syllables (and three bisyllables) have brought time to its slowest pace, the fog comes to us, who are held prisoners in our "magnificent" city. Our insignificance and powerlessness are emphasized by the smallness and quickness of city's two vowels. Line 4 is the only line that ends with a vowel, and since the next line begins with a vowel, the tiny last vowel of city can ease gradually into the next line's opening vowel, possibly suggesting a small, fading, unanswered cry
We are, at the end of line 4, in the grip of the fog, a presence that has been made increasingly animate, and even personified, by the sequence of words comes, feet, sits, and looking - a being who moves as fluidly and quietly as a cat, a being just as unpredictable and mysterious as a cat, a being with all the mythic power (and even menace?) of the Feline . . . . and time stops . . . .
*
In the next line, the mention of the adjective silent opens the possibility of the opposite of silence - could this vast being perhaps speak and judge us? We do not know. We are frozen into immobility. All we can do is to surrender into the waiting.
And then suddenly, its departure as rapid as its arrival was gradual, the fog lifts, is gone, is on its way to look over others, and hold them too in its thrall. The rapidity of the fog's departure is signalled by the shortness of the words in the last line, by the fact that moves has no audible subject, and by the shortness of the last on-phrase.
Experientially, fog can be seen to have a sonata structure - ABA. Our poem starts and ends with motion verbs (comes and moves), which sandwich a durative verb, a verb of stillness - sits, even though the word looking, which follows sits, tells us that there is no break in the activity of the fog. Paradoxically, this activity of looking, followed by the act of refraining from speaking, makes the center of the ABA structure the peak of agency, of animacy, of the visitor to our city. What is relevant for our concerns here is the fact that the phonetic structures of fog and of cat can also be seen as being of the form ABA - two consonants which sandwich a vowel. So the verbal melodies of fog and cat "rhyme" in yet another way with the structure of the experience of fog.
Now let us return briefly to the poem's six bisyllables. We find that they always come in pairs: the first two begin with [l], then we find a line with one beginning in [h] and one beginning in (phonetic) [s], followed by a line whose two bisyllables come in the reverse order - the one beginning in [s] preceding the one in [h]. These three bisyllable pairs always appear on adjacent lines, so arranged that if the first member of the pair appears early in the line, its "echo" in the next line will end the line, or vice versa. Thus there is a deeply chiastic, or crossing (literally, "x-like"), pattern that Sandburg sets up with his placing of the poem's long words.
The fog comes
on little cat feet.
It sits looking
over harbor and city
on silent haunches
and then moves on.(little-looking, harbor-haunches, city-silent)
There are no more bisyllables to link up, but our ears hear more similarities: the last two monosyllables of line 2 have the front vowels [æ] and [i:] followed by [t]'s, and the first two words of line 3 are monosyllables whose front vowel [I] is also followed by [t]. Furthermore, the [t f] sequence between cat and feet is echoed by the [t s] sequence between It and sits. Thus if we also link these pairs of adjacent [t]-final monosyllables up, we are left with the following, more symmetrical structure:
The fog comes
on little cat feet.
It sits looking
over harbor and city
on silent haunches
and then moves on.(little-looking, feet-it, harbor-haunches, city-silent)
But there are still more chiastic relationships to indicate, the clearest being the relationship between it sits and city , which I see as a sort of echo of the cat feet / It sits linkage between lines two and three. I think, however, that the first place that we hear phonetic crossing is between the last two syllables of the first line and the last two of the second line. Fog and feet contain the only two [f]'s of the poem, and comes and cat are the only two words that begin with [k].
"Hey! Waidaminit!" I can almost hear some readers saying. "Why should we care about all these formal patterns? What do they have to do with what the poem means to us?" I can sympathize with such objectors, and while I would not insist that such structures always must be linkable to meaning, I feel that in this poem, such a connection is possible. One interpretation of such crossing patterns, which make equivalent, in a certain sense, the orders of elements AB and BA, is interpenetration - the commingling of A and B - in our case, of fog and cat. While an essayist could simply say that fog and cats move like each other, a poet enacts this kind of perception, choosing crossing structures - which dissolve boundaries between entities previously felt to be distinct. Structures like this chiastic one come from a kind of universal symbolic alphabet, which I believe to be used not only in poetry, but also in other forms of art and in other rituals in a culture.
If I am right in seeing these crossing structures as being used to enact a kind of poetic fusion, then it is significant that the two lines between which there is the least evidence of this kind of crossing are lines 5 and 6 - just the place where the fog-cat leaves. So the fog is functioning as the agent of the chiastic structure, which we could show in its fullest form as below:
The fog comes
on little cat feet.
It sits looking
over harbor and city
on silent haunches
and then moves on.(little-looking, fog-feet-it-city-silent, harbor-haunches)
I believe, too, that Sandburg uses nasality as an index of motion. Note that each of the first four lines has just one nasal, and that these nasals are in syllables of steadily decreasing phonetic prominence. The first word to contain a nasal, comes, is highly salient. It is the first verb, and a verb of motion at that. It has full word stress, even more stress in the first line than does its subject, fog. This is a way of saying, in the language of poetry, that one thing that will be important in the poem is the activity , the motion, of the poem's topic. And what distinguishes this type of motion from other possible monosyllabic verbs of motion that the poet could have used (slides , creeps , roves , walks , etc.) is its direction: this fog comes to us .
There is a very tight patterning in the poem's first line which makes the first nasal [m] stand out even more visibly. For notice that the word fog has a labial [f] before its back vowel, and following the vowel, a stop consonant, the velar [g], made in the back of the mouth. By contrast, comes has the positions of velar and labial reversed: the velar [k] precedes its back vowel, while the labial [m] follows it. What makes this reversal particularly noticeable is the tongue-stopping phonetic sequence [gk], two velars, differing only in their voicing, coming between the first two content words. Thus line 1 gives a phonetic preview of the being stopped, and then being released, which will form a central part of the type of experience of fog that the ensuing poem will evoke.
The words fog and comes are linked by reversal in more than one aspect of their melodies: fog starts with the fricative [f] and ends with the stop [g], while comes starts with the stop [k] and ends with the fricative [z]. However, they are also linked by a pattern of repetition: the change of voiceless [f] to voiced [g] is echoed by the change of voiceless [k] to voiced [m] and [z]. Summing up all these relationships diagrammatically, we could show some of the ways in which the two stressed words of the first line are related as follows, where the double line links the repetition of the change in voicing, and the dotted lines indicate reversals:
f ø : g # k U m z
f | # | m | z | ||||
= | a: | U | = | ||||
= | = | g | k | = | = | = |
With the tightness of the linkings between these two words, I think it is fair to say that the [m] stands out fairly clearly. It makes for the first consonant cluster within a word, and it is the first line's finale. Then comes the next nasal: the [n] of on, a word whose vowel is in no way reduced, but one whose stress is clearly less than that of comes. The third nasal is the velar [°] at the end of looking; for the first time, the nasal is located in a suffix, not in the lexical part of a word. The preceding vowel, though having far less stress than the [U] of look, is probably most naturally pronounced as the unreduced high front vowel [I]. However, when we reach the and of line 4, the most normal pronunciation of its nucleus would be the reduced vowel [], not the unreduced [æ] of cat. Thus clearly, this syllable, as typically pronounced, has the least prominence of the first four to contain nasals.
Significantly, if it is correct to link nasality and activity, this lowest ebb of activity comes just before the place where we are - city. Immediately thereafter, as activity begins to pick up, we see that three out of five syllables in line 5 contain nasals, and then that all four of the last line's four words do. Notice that the last line's verb, the prototypical indication of motion, even starts with a nasal, the only word in the poem to do so. Since sounds at the beginning of their syllables are more prominent than are ones at the ends of syllables, this means that nasality attains its highest salience just as the poem awakes from its stasis and bursts into motion.
I have suggested that crossing structures can be one way of "saying" interpenetration, in the symbolic alphabet of art. But there are other ways, one of which being the use of alternating structures. Our poem manifests many kinds of alternation, one salient example being the number of nouns or pronouns per line: 1fog / 2cat, 2feet / 3it /4harbor, 4city /5haunches. Notice that this pattern starts with two singular nouns and a plural (all monosyllables - note that the first and last start with the poem's only [f]'s). Then, after the pronoun it , the only one in the poem, which marks the center of the alternation, there is a repeat of the sequence of two singulars and a plural, this time all bisyllables - with the first and last starting with the poem's only occurrences of [h]. Again, the locus of this alternating pattern is the first two sections of the poem. Like the locus of the crossings that we have tracked, we see that the locus of alternation in noun number is just that of the fog-cat itself. I think that above and beyond the brute fact of alternation, it is significant that it is nouns that are fluctuating in number, and the fact that the two numbers involved are 1 and 2. Are we perhaps justified in seeing this as a suggestion of the splitting and fusion of the two concepts of fog and cat? This may be stretching a point - I don't insist on it. As I said above, it is not always easy to be sure as to whether some identifiable aspect of a formal structure contributes to the interpretation of a poem or not.
I note in passing that the first five lines of the poem are set off as a unit from the sixth line by two more facts of distribution: lines 1 and 5 are the only lines in the poem which end in the suffix {-z}, and haunches is the only other word in the poem besides fog to contain the vowel [ø :]. Below, we will see further examples of the way Sandburg has sectioned off the first five lines, to form a poetic subunit.
To return to alternating patterns, there is another alternation, even more salient, which has as its scope the entire poem - the alternation in the number of words in a line. The odd lines each have three words, and the even lines each have four. This alternation, and the previously cited one involving the number of nouns per line, impose a different sectioning on the poem, a sectioning that groups the six lines into three "couplets" - lines 1 + 2, lines 3 + 4, and lines 5 + 6. Yet another alternation, a phonetic one, which has exactly the same scope as that of the number of words per line, concerns the length of the last vowel in each line, The odd-numbered lines have short vowels (comes [], looking [I], haunches []), while the even-numbered ones have phonetically long vowels (feet [i:], city [i:], on [a:]). Thus there are two distinct alternations which section the poem into three pairs of adjacent lines.
The previously mentioned ABC sectioning agrees with this "coupleting" in one respect: in establishing a boundary after the first two lines. But then, the coupleting "interferes" with the sectioning off of B - lines 3 - 5. The alternations break off the first two lines of B, drawing a boundary right after city . However, we have already seen that there are good reasons for thinking that the function of this longest line in the poem is to raise our waiting, our powerlessness, to the limit. Thus it can be maintained that on a deeper level, there is no interference here - the two sectionings are in no sense in competition. Rather, like polyrhythms in music, they play off on one another. Each contributes its own voice to the beauty of the poem.
This having been said, let me go on to one last sectioning, one which further reinforces the strength of the central couplet. In this third sectioning, the first line is linked to the last, the second to the fifth, and by elimination, the third to the fourth. The outer lines are linked, as I have already pointed out, by virtue of the fact that they are the only two monosyllabic lines. Furthermore, they contain the poem's only occurrences of voiced th, the sound, in the and then, and more saliently, also the only two occurrences of [m] - at the end of the first verb, and at the beginning of the last one.
However, the links between lines 1 and 6 pale beside those between lines 2 and 5. Being the poem's only two five-syllable lines, both of these lines begin with on, and each ends in one of the poem's two plurals (which are both body parts of the cat). And finally, these two plural nouns are modified by the poem's only adjectives, which might, by the way, be said to "rhyme" semantically, in light of the fact that silence may be seen as smallness of sound.
This then leaves us with lines 3 and 4, which I do not see any great phonetic similarities between, except for the already noted link between it sits and city. But they are solidly joined by virtue of their syntax: the five-word participial phrase looking/over harbor and city, the locus of the greatest animacy in the poem, is inserted between the verb sits and the on-phrase which is its object. Interestingly, this participial phrase contains the poem's only ambiguity: the over-phrase can be read as a directional object of a motional look (as in looking over them), or the conjoined noun phrase can be taken to be the direct object of the phrasal verb look over (as in looking them over).
The links between lines 1 and 6 are noticeable, and those between lines 2 and 5 links are compelling. This, when added to the syntactic cohesiveness that rules over lines 3 and 4, further increases the necessity of hearing the middle two lines as a couplet, thus once again heightening the strength of the tension at the end of line 4.
Let us look at one last type of alternation. Given the various alternating patterns that we have already seen, all of which "rhyme," at a deep phonetic level, with the A-B-A consonant-vowel-consonant structure of fog and cat, there is reason to take a careful look at the network of consonants that Sandburg establishes in composing the music of this poem. Before I do so, I want to make use of a distinction that phoneticians often draw among the various classes of sounds that can be found in the world's languages. The greatest distinction that appears in the domain of phonetics is that between the most musical, or sonorous, of sounds - the vowels - and the least sonorous ones - the stops (e.g., [p,t, çc (the first sound of chew), k, b, d, çj (the first sound of jaw), g, and so on]). Stops are given this name because in the course of making them, there is one point at which the vocal tract is completely closed, and no air emerges from the lungs through the mouth or nose. One degree more towards the free-flowing vowels is the next most sonorant (or less obstruent, to use a technical term) class of consonants, the fricatives (e.g., [f, s, " (the first sound of think), (the first sound of shift), v, z, (the first sound of this), çz (the last sound of rouge), and so on]). These sounds are called fricatives because in producing them, instead of stopping entirely the flow of air from the lungs, the tongue or lower lip is put so close to one part of the upper part of the mouth that a turbulence, or friction, in the airstream is caused. Phoneticians use the term obstruent to designate the class consisting of stops and fricatives together.
Returning to our poem, we see that fog and cat both begin and end not only with consonants, but with the subclass of consonants that I have just defined - that is, with obstruents. Let us therefore take a close look at all of the other obstruents of the poem.
The fog comes | th | f | g | k | z |
on little cat feet. | t | k | t | f | t |
It sits looking | t | s | t | s | k |
over harbor and city | v | h | b | s | t |
on silent haunches | s | t | h | ch | z |
and then moves on. | th | v | z |
For the first three and the fifth lines of the poem, a clear pattern emerges: each line has just five obstruents (for this purpose, I am treating [h] as an obstruent, though many languages treat it as a glide, like the semivowels [w] and [y]). Moreover, it is clear, in looking at line 2, that here we have another pattern of alternation to deal with: while the same sound, [t], appears in the odd positions in the set of five obstruents, in positions 2 and 4, we find different sounds - [k] and [f]. Thus line 2's obstruents are a clear instance of an abstract alternation schema which we can designate as ABABA. Line 3 presents us with a slightly different variant of the same basic schema: what the odd positions here share is the fact that they are all stops, while the even ones share the property of containing fricatives - here, the same fricative, [s]. The fifth line has fricatives appearing in its odd-numbered positions, with its two stops showing up in its even positions.
So far, so good. Now what of line 1? In order to be able to perceive the same kind of ABABA pattern that has come to light for lines 2,3, and 5, we will have to introduce one more phonetic distinction - that of voicing. Phoneticians classify the two sounds [f] and [v] as both belonging to the same class - that of the labio-dentals, since both are made with the lower teeth touching the upper lip, producing a characteristic friction at that point. However, since the two sounds are clearly different, they observe that during the production of [v], the vocal chords, the two bands of muscle in the throat just behind the Adam's apple, continue to vibrate, while this is not the case for [f]. They call all such consonants as [v] voiced, with consonants without this characteristic buzz of the vocal chords being called voiceless. In English, all obstruents (except [h] - a reason for not thinking that this sound is an obstruent) come in voiced/voiceless pairs: [b/p, v/f, d/t, z/s, çj/çc, çz /, g/k, etc.]. Now when we turn to line 1, to try to find out what the odd positions have that the even ones do not, we can see that it is precisely the feature of voicing.
I have saved until last lines 4 and 6, which raise several problems. First of all, line 4 has six segments that should normally count as obstruents (I will include [h] in this list, as I would like it to be included in the algebra of line 5): over harbor and city. The only way that I can see to make this line conform to the generalization that there are just five obstruents in each line is to say that the [d] of and does not count, since it is swallowed up in the [s] which immediately follows it. This does no violence to the phonetic facts; it is very difficult to articulate line 4 so that both a [d] and an [s] are audible. However, even if we take this tack, thus preserving, on the phonetic level, at least, the generalization that there are only five obstruents per line, what can we say about the further generalization to the effect that the obstruents manifest an ABABA alternation? What [h] and [s] have in common is that they are both fricatives and voiceless, two properties which no other segments in the line have. While the [v] of over is a fricative, it is a voiced one; and while the [t] of city is voiceless, it is not a fricative. But it is clear that this conjunctive characterization of the even-numbered obstruents of line 4, though it is technically accurate, is a more complex one than was necessary for any of the other lines. Thus line 4, even though it may be seen in such a light as to remain within the ABABA framework of alternating obstruents, only makes it by a whisker, a fact which is highly congruent with the intensity of the poetic experience that this line evokes, as I have called attention to above. Line 4, then, is a good example of the kind of paradox that poetry delights in being: something that is simultaneously the same and different.
It is time to conclude our discussion of the alternation among the obstruents within a line. Line 6 has four obstruents, but I would like to suggest again, as was the case in line 4, that the [d] of and can fuse with the following sound, the [th] of then, so that we can say that line 6 manifests only three obstruent positions, in the A-slots of our abstract ABABA schema. All of line 6's obstruents are voiced, as are the three odd-numbered ones of line 1, which suggests the return to the beginning that is independently conveyed by the return to monosyllabicity, the return of the [th] and [m], and the return of a motion verb. However, as we have seen in many ways, just as it is the first five lines which are the locus of the alternation in noun number and also of the chiastic linkings between lines, it is also in these first five lines that we see the find the full ABABA obstruent alternation - all that remains by the time we come to line 6, when the cat-fog goes, is the wisp of an echo. I think that it is even likely that the dropping out of the B-slots in the ABABA pattern in line 6 is one last way of "saying," in this language of poetry that we have been exploring, that the fog/cat has left. For note that what occupied the B-slots in lines 1 and 2 were the pairs of sounds [f]-[k] and [k]-[f], respectively - the initial and most salient sounds of the parts of the fog/cat. The alternation in order helps the fusion of the two parts along.
There is one final family of patterns which are woven into the fabric of the poem, as they are into the fabric of its central word, fog. This is a feature of the poem which we could refer to as its threeness, or, as long as I am making up words, let me suggest instead a new verb. I will say that Sandburg threes this poem, that this threeing is a primary source of poetic coherence. We have already mentioned many examples of sets of three elements in the discussion above - let us have a brief review.
Threeings
Syntax: 3 sections, each consisting of 3 basic parts - a subject referring to the fog, a monosyllabic verb in the present tense, and an on=phrase (the central section has an adverbial looking-phrase inserted in it)
Experience: arrival of fog - state of being fogged in - departure of fog
Phonology: 3 segments of fog, corresponding in their length relationships to the subjective durations of the 3 parts of the fog experience. Furthermore, each line manifests 3 obstruents of the same type, with all lines except the last having two more obstruents interspersed between the first 2 and the last 2 of the set of 3.
Words per line: the odd lines have 3 apiece, with the even ones having 4, which then gives 3 couplets.
Onion-skinning: first and last lines are linked by virtue of being mono- syllabic; second and fifth are linked because of having only plurals, only body parts, only adjectives, etc. And the third and fourth lines are joined by the syntax: the adverbial gerundive phrase looking/over harbor and city, the peak of the fog/cat's animacy, interrupts the clause that begins on line 2 and ends on line 5. So the poem consists of a set of three nested pairs of lines:
1 6 2 5 3 4
To conclude our preliminary look at the way Sandburg's whole poem is a masterful expansion, or a kind of holographic echoing, on many levels, of the single three-sound word which serves as its title, let us look at the poem one last time, with its three alternative sectionings and its chiastic crossings indicated.
Fog
A.
1. The fog comes
2. on little cat feet.B.
3. It sits looking
3. over harbor and city2. on silent haunches
C.
1. and then moves on.(little-looking, fog-feet-it-city-silent, harbor-haunches)
I have tried to show how the acoustic structure of the word fog, with a basic ABA melody laid down by its sequence of obstruent-vowel-obstruent, harmonizes both with the basic motion-stillness-motion perception of an experience of being fogged in that is given by the three-part sectioning ABC on the level of syntax, and also with the durational structure of 2 lines-3 lines-1 line mirroring the durational relationships among the three phonetic elements of the word fog. In addition, alternation flowers in many ways throughout the poem, as we have seen above, with a suggested interpretation that it can be one way of letting fog and cat fuse.
I am afraid, though, that I have probably lost most of you, unless, you, like Jack DeWitt and me, just can't help ripping poems apart to try to see where the magic in them comes from. For I have gone into a lot of detail, more than many people will perhaps want to concede might be tolerable, let alone necessary. After all, I hear them saying in my mind, isn't this a rather minor poem? Why all this fuss about it?
To try to answer these possibly only straw people in this imaginary dialogue, let me remind them that I find this poem still extremely beautiful - I think more so than when I started working on it about six years ago. The more ways I see in which this poem's 22 words become interconnected, the more radiant it becomes for me, and the better I think my chances are at explaining the incredible memorability of this work of art. It can't be just the admittedly masterful metaphoric linking of fog-cat, for that wouldn't explain why people don't confuse line 2 in their minds with such rotten lines as my on small cat feet or on little cat paws, which would preserve the image/metaphor intact, but lose on the music. I think that this poem gets its power from the way the aptness of the image interlocks with the multiple musical coherences that are generated by the acoustic-articulatory-percep-tual event unleashed by a pronunciation of the word fog. This interlocking is made to shine even more brightly in the unguessed depths of our inner selves by the numinosity, the heart-in-the-mouth grippingness, of our vulnerability, which we feel when we encounter, all alone, the Vastness. Sandburg "says" this encounter for us, in the universal language of poetry, through the fusion of catfog, fogcat, cfaotg, . . . We have no way to spell such experiences in English, except by a wonder like the jewel network of these twenty-two words.
I offer this poem, and the preliminary look at the kind of richnesses of musical texture that reward the hopelessly addicted ripper-to-pieces of poems, as a kind of meditation for all of us who gather here in this volume to celebrate the teaching and thinking of our friend Jim Jenkins, who has delved into many similar depths of cognitive and developmental structure in his work. I am glad to see workers in the broad field of cognition, where Jim has pioneered, turning, from many and varied starting points, to the study of metaphor, which surely plays a fundamental role in the process that Mallarmé refers to as "writing with words." But I want to invite cognitive scientists to remember that mere metaphor, metaphor not even linked as fragmentarily to verbal music as I have tried here to sketch a beginning for, will never lift us to see the profoundest truths of the song of human creativity, which I think is what we are all really longing to understand. If seeing the beauties of this song is what our goal really is, the magic of "writing with words" cannot be something we come to last, as an afterthought, but must instead become a foundation stone for a new, fuller, and deeper understanding of the mysteries of mind.
Bibliography
Jakobson, Roman, and Stephen Rudy, "Yeats' Sorrow of Love through the years," Poetics Today, Cf. p.-15
**********
I believe, too, that Sandburg uses nasality as an index of motion. Note that each of the first four lines has just one nasal, and that these nasals are in syllables of steadily decreasing phonetic prominence. The first word to contain a nasal, comes, is highly salient. It is the first verb, and a verb of motion at that. It has full word stress, even more stress in the first line than does its subject, fog. This is a way of saying, in the language of poetry, that one thing that will be important in the poem is the activity , the motion, of the poem's topic. And what distinguishes this type of motion from other possible monosyllabic verbs of motion that the poet could have used (slides , creeps , roves , walks , etc.) is its direction: this fog will approach us .
There is a very tight patterning in the poem's first line which makes the first nasal [m] stand out even more visibly. For notice that the word fog has a labial [f] before its back vowel, and a stop consonant made in the back of the mouth, the velar [g], following it. By contrast, comes has the positions of velar and labial reversed: the velar [k] precedes, while the labial [m] follows, its back vowel. What makes this reversal particularly noticeable is the tongue-stopping phonetic sequence [gk], two velars, differing only in their voicing, coming between the first two content words. Thus line 1 gives a phonetic preview of the being stopped, and then being released, which will form a central part of the type of experience of fog that the ensuing poem will evoke.
The words fog and comes are linked by reversal in more than one aspect of their melodies: fog starts with the fricative [f] and ends with the stop [g], while comes starts with the stop [k] and ends with the fricative [z]. However, they are also linked by a pattern of repetition: the change of voiceless [f] to voiced [g] is echoed by the change of voiceless [k] to voiced [m] and [z]. Summing up all these relationships diagrammatically, we could show some of the ways in which the two stressed words of the first line are related as follows, where the double line links the repetition of the change in voicing, and the dotted lines indicate reversals:
f ø : g # k U m z
With the tightness of the linkings between these two words, I think it is fair to say that the [m] stands out fairly clearly. It makes for the first consonant cluster within a word, and it is the first line's finale. Then comes the next nasal: the [n] of on, a word whose vowel is in no way reduced, but one whose stress is clearly less than that of comes. The third nasal is the velar [°] at the end of looking; for the first time, the nasal is located in a suffix, and not in the lexical part of a word. The preceding vowel, though having far less stress than the [U] of look, is probably most naturally pronounced as the unreduced high front vowel [I]. However, when we reach the and of line 4, the most normal pronunciation of its nucleus would be the reduced vowel [], not the unreduced [æ] of cat. Thus clearly, this syllable, as typically pronounced, has the least prominence of the first four to contain nasals.
Significantly, if it is correct to link nasality and activity, this lowest ebb of activity comes just before the place where we are - city. Immediately thereafter, as activity begins to pick up, we see that three out of five syllables in line 5 contain nasals, and then that all four of the last line's four words do. Notice that the last line's verb, the prototypical indication of motion, even starts with a nasal, the only word in the poem to do so. Since sounds at the beginning of their syllables are more prominent than are ones at the ends of syllables, this means that nasality attains its highest salience just as the poem awakes from its stasis and bursts into motion.