Thursday, September 3, 2020

The Inattentive Reader*


Henri Matisse, "The Inattentive Reader," 1919 (image from tate.org.uk).

    The silent reader motif in painting (sometimes depicted either as a female imbued with faint corporeal sensuality, or as an individual in serene isolation whose whole attention is absorbed by the text) appears to function as a cognitive or semiotic means to divide the visual field between what is seen and what is read. More importantly, though, is the rhetorical fusion it hides while demarcating the boundaries between the body of the text from the body of the reader via its detour into the erotic or the sensuous. Reading becomes for the beholder the enjoyment of rediscovering the body of the reader as distinct from the body of the text and the world. It is as if it were in the gesture or posture of holding the textual body that the sensuous energy of the reader's body is generated and invested with desire. Reading or the pleasure of the text arises in that scene where the desire for a body becomes its paramount object of focus. 

    To maintain this achievement at its highest notes, however, the reader's body must keep contact with the textual body. That tactile bridge represents the material hypostasis required in the generation of the physical in the world of the image. The loss of this tactile bond (the reader letting go of the text) dissolves the spell that maintains the eroticized lectural bond: the text being read becomes an inert object, and desire finds nothing to feed on but drab elements of the medium (here given to us by Matisse through the color choices, the posture and eye direction, and the sexless clothing). The shimmering light that held up the rapt attention on the readerly face or body dissolves simultaneously with the disappearance of the tactile connection with the textual body. The eye wanders away from the page and the text is left idle on the table, disrupting the symbiotic bond that conferred the simulacrum of lifelike energy to the readerly body as the ontological extension of the textual body. 

    The painterly contact between the model and the text--existing in principle as signs on the canvas and on the page--feeds both ends of their conceptual possibility, pushing us to vacillate between the two poles of the simulacrum, that is, between "essence" and "appearance," so that the semiotic discourse of perception that splices the sign between the physical and nonphysical can function in language as identifiable categories. They are, hence, not so much irreconcilable opposites as the medium in which reading and desire become functional metaphors of one another.


The self-generation of textuality can be seen in the binary separation it enacts between itself and its referent. To exist as an object apart, it must name its other, thereby maintaining its status as a perceptible referentiable object. This primary ergonomic dichotomy can be seen via the deictic functions which orient the imaginary space of reading. Various reality effects depend on such markers which name the outsides of the text, while at the same time index the text "itself" as the site where such citation happens. More obvious self-references are locutions such as ''at the time of this writing.'' In general, the spaces and temporal dimensions it creates, its proleptic-analeptic axes, extend for the reader a plane of existence between what is read and what is mentioned. We can even look at narrative codes similar to what R Barthes listed in S/Z less as linguistic or literary "devices" than as ergonomic resting points or "landmarks" where both text and reader become locatable coordinates in the groundless space of representation.

    There is much work to be done in the elucidation of this self-generative logic where textual and non-textual coordinates are set up by reading motions as projections of its own prosodic traces. This ergonomic notion dilutes the over-emphasis on structurality or literary form and answers the question what audience it would be a structurality for. Encoded in the design feature of textuality are the means by which the dichotomy text-nontext is made and kept perceptible. It is a border that can become thematized whenever a metafictive or metalingual tactic is set in motion. Like the mirror stage, a reader is drawn into the deictic network of the interminable play of coordinative reference where binary values exchange places in an irresolvable dialectic. This could be seen as the chiasmic spiral where desire and semiosis animate each other in an exchange logic without closure.

Wednesday, July 22, 2020

Notes for "r-p-o-p-h-e-s-s-a-g-r"

   --In piece no. 13 of No Thanks, as it is often elsewhere with Cummings' penchant for tmesis, simultaneity and seriality are placed in tandem, as if language was forced to confront the very irresolvable dialectic informing its temporary movement in phonographic solution space. Why do we need a solution space? It’s less a question of need than what is simply an encounter, a disjunctive encounter, the arrival of the gap and the metaphoric dimension of the sign as a stopgap measure, the solution which simply puts the gap in sharper focus or perpetuates it, prolongs it, by plotting one (re)solution after another.

   --Paradigmatic vs syntagmatic crowding, as if slots for letters had more space for just one, so that a polyphonic stream is filtered to derive the normal word morphology and syntax. Thus simultaneity demands a compromise in serial space where several signals can coexist and must be parsed into their common morphological grouping. Syllables, affixes, phonetic groups, syntax groups, or a unilinear sentence or IPs, no longer command primary aesthetic lodging in syntagmatic chain, but must compete for priority with the virtual crowd in the paradigmatic background. As if Cummings was illustrating an algorithm of linguistic online processing, the pre-utterance cognitive neural mass going through the neat grid of the Symbolic order. Everything is free grapheme, only ASGrs set the conditions to bind them. This is why the iconic is the grammatical and vice versa for Cummings (cf grammatical metaphors). In solution space, all graphemes are free until a regime of signs grids them.




    --The in/famous grasshopper poem can be seen less as a mimesis of the kinetic or a visual onomatopoeia or isomorphism between graphic prosody and motion, than the provision of the space of reading motions where the recombinatory logic of semiotic notation is foregrounded and reactivated. In short, a diagram of lexicogenesis in reading motions. We are returned to a metaprosodic stance where verbal and graphic equivalences are being invoked for reconstitution in their pragmatic and informational equivalences or parallelisms. It is moving closer to the simulation of interactivity as the spacing of reading motion, returning us to the productive moment where segments and supersegments are reconstructed out of nonsegmentarity, where linearity and nonlinearity become metaprosodic options for the reconstruction of the writing system in graphematic solution space. A writerly grammatogeny, instead of a readerly one. Tinkering with the writing system’s solution spaces in a more explicit manner, above the implicit infrathin differences we introduce with each deformance, by reconfiguring its forms, order, disposition, direction, appearance, relations, values, operations, functions, organization, categories, combinations, tactics, and so on.

    --Running against the monocular unilinear voice. Polyphony. One line at a time is too limiting. Multi tasking. An utterance is a crowd, parsed through an algorithm of priorities. Parentheticals as the compromise in syntagmatic chain or line, emphasis on orthogonal paradigmatic field, the simultaneous vs the serial. Metrical grid is dominantly serial syntagm. Cummings grid overlays it with the paradigmatic mass coexisting pt by pt in synchrony with the overmarked line. Line within a line, multiple syntactic processes running side by side, criss crossing, interacting, commenting on and in dialogue with each other. Such multiple articulation interlacing and implicating each other pushes the limit of the glottographic resources of the graphematic solution space. Cacophony is managed by recourse to nonglottographic prosody, letting the eye pick out the multiple interlacing syntax via scriptly signatures.

   --The graphemic morphotactic arrangement of < grasshopper > would seem to be a commentary on the eventuality of the arrival of standard orthography and lexeme, representing the direction which looks like the telos of becoming in which linguistic elements and beings coincide in an isomorphic mapping format. Three timelines intersect in the space of reading, the event being divisible among the syntactic grammar of Subject, Verb, and Object (spectator, process, spectacle). The directionality of writing as a linear narrative frame against which the process of arrangement happens, a sort of slow motion frame of literacy. The grasshopper poem is diagrammatic mapping where the aesthetic entities we call word, line, sentence, syllable, CV, directionality, page, stanza, paragraph, letters, punctuation, syntax, morpheme, phoneme, and so on emerge as perceptible objects in the dynamic space of reading.

Wednesday, May 6, 2020

Visual Prosody, E. E. Cummings, and Solution Spaces

The Shadow of Speech: Visual Prosody, E. E. Cummings, and Solution Spaces



I. Glottography, prosody, and solutions spaces


From this theoretical conception it follows that written representations do not relate to the phonetics of a language, as phonetics is situated outside the language system and pertains to the realm of language use. Instead, written representations are related to phonological representations. The relation between these two kinds of representations, however, is not derivational but rather indirect: a regular written representation of a specific word is element of the graphematic solution space of the phonological representation of that word. To put it differently: a word has to be spelled in a way that allows the regular recoding of its phonological representation. If the graphematic component of a writing system supplies more than one such spelling option, as is the normal case in natural writing systems, a set of systematic-orthographic constraints may reduce the number of options. This reduction does not necessarily leave one fixed spelling of the word but usually a larger set. Conventional orthography in the end decides which of these options the correct one is. This means that conventional orthography cannot be fully reconstructed as a theoretical system, but only partially.

Martin Neef (2015) 1

As Sproat (1996) observes, ‘the primary purpose of writing is not phonetic transcription, but the representation of words and morphemes’. Different writing systems represent those words and morphemes through different systems of representation and English uses twenty-six letters derived ultimately from the Roman alphabet. Each spelling is a complex piece of data which can contain more information than just a string of phonemes. English spelling frequently contains morphological and etymological information, connecting spellings to related words within English and/or in the language from which the word was borrowed. Further intermediary levels of phonological information may even be encoded, including syllables, stress and even foot structure, although the recent research on this is too new to have reached any consensus. All of this information is contained within an elaborate set of correspondences mapping from spelling to sound (and back) and much work has been done in recent decades— to understand the complexity.

— Des Ryan (2016) 2


     As a complement to the American computational linguist Richard Sproat’s idea that writing has a primary purpose, or a hierarchy of choices between the phonographic and the logographic principle in writing systems, we turn to an article which tangentially connects E. E. Cummings to the subject of this essay, and where the historian of science and linguistic scholar Malcolm D. Hyman differentiates between the “glottographic” and “nonglottographic” functions of writing. In general, “Glottographic writing represents linguistic content that can be spoken, whereas the content of non-glottographic writing is non-linguistic. We read glottographic writing, but verbalize non-glottographic writing.”

In a passage that has been quoted to the point of tedium, Aristotle declares that written words are the signs of spoken words, which in turn are the signs of psychological states (De interpretatione 16a4–6). That, of course, is a statement about glottography. Nonglottographic writing lacks the mediation of spoken language; its notations are themselves signs of psychological states....

Although natural language syntax comprises hierarchical relationships (represented by trees in generative grammar), speech is realized as a linear sequence of words (or of morphemes). We expect a glottographic writing system to encode speech as a string of graphemes. This string must be accommodated to the two- or three-dimensional surface upon which it is inscribed (e.g., by making provision for line-breaks); but the dominant principle is that graphemes are arranged in an order that corresponds to their licensing elements in the temporal chain of speech. (Though the letters of an e.e. cummings poem sprawl drunkenly down the page, their arrangement is determined in large part by a code that exists in addition to the glottographic one – How to reproduce the arrangement orally?)3

     Here, we will need to place our focus on the term “function,” which is an operation that bifurcates into the phonographic and the logographic principle, instead of viewing a pre-existent substance called “writing” whose nature we already know in advance outside of or before writing. This is the main reason Hyman goes to explain language in term of dynamically interacting subsystems where glottographic and non-glottographic functions figure as the constituting operations (specifying also how a notational technology involving physical graphic marks could be set apart to become recognizable as “writing” and not as “speech.”) In other words, if we get rid of the idea of a global system of language divided between parole and écriture, and view everything as a dynamic operation of differentiation wherein various subsystems emerge and play, then what gets indexed as speechly or scriptly does not need any labelling as either “writing” or “speech” in terms of idealized domains. Instead, they are pragmatic functionalities which bifurcate into an encoding for vocal/phonic and/or visual-graphic (re)production. Such bifurcation is merely an adaptation of a double-encoding for the two major sensory modes utilizing two media channels: the Ear for sound (or mechanical waves), and the Eye for light (or electromagnetic waves).

     Hyman, therefore, counsels against using these functions as typological or as structural criteria because “the binary classification of entire writing systems as glottographic or not is overly schematic.”4 Giving the English writing system as an outstanding example of where the two functions coexist to produce a mixed operational modality,5 Hyman also raises an interesting point concerning these functions:

What is the form of glottographic writing? In a superficial sense, it is a line: left-to-right (for English), right-to-left (for Hebrew), top-to-bottom (for Mongolian), boustrophêdon (i.e. alternating direction: left-to-right, then right-to-left, etc.) in the case of several palaeographic traditions. Sometimes writing follows a different curve in order to accommodate itself to its material: thus an inscription around the rim of a plate will follow the circumference. Only rarely does glottography transfigure itself completely, into something like Apollinaire’s Calligrammes, or the calligraphic rendition of sūra 105 of the Qur’an (al-Fīl) that takes the form of an elephant.6

     Although it is doubtful that such an occurrence where glottographically-associated elements (like the alphabetic system) become nonglottographic is a rare phenomenon, found as if mostly in the realm of Concrete poetry and other avant-garde styles from Apollinaire to Cummings (belonging to a long tradition, as noted by poetics scholar Richard Cureton in 1986), Hyman at the least makes us aware that the functions are pragmatic operators in the sense that they are dynamic principles of continuous “inherent variation.”7 Given sufficient ingenuity, therefore, a glottographic instance can be made nonglottographic or vice versa, and this is what we see often happening in the visual prosodic employ of typography for which the avant-garde since Futurism, Lettrisme, modern advertising, New Media font technology, and, of course, E. E. Cummings are well-known. Even from Hyman’s “spatialized” characterization for glottographic writing (“right-to-left,” “top-to-bottom,” “a different curve”), we could already sense the importance of overtly visual cues or graphic prosodies which accompany, corroborate, or translate the glottographic deployment of sign-forms. (We should note, however, that it would be misleading to attribute—as is traditionally done— temporal seriality exclusively to speech and “spatiality” to writing. It would be more accurate to say that, while speech acts are temporally constrained by linearity, script acts are capable of various temporal modalities and are, therefore, multilinear.)

     It is understandable then that Hyman has reservations about the “usefulness” of the two terms as typological and structural determiners. However, we can easily infer from below how a shift in the way the terms are handled transfers their importance to become metaprosodic operators that negotiate the indexical terrain between verbal and visual prosody.

In light of the diversity of form and function in real-world instances of writing, I propose that the typological model of pure glottographic and non-glottographic systems is unhelpful. Rather, we may conceive of writing as a system of systems. Among the subsystems commonly are a numeric subsystem and a glottographic (lexical/morphosyntactic) subsystem. Interspersed with these are other subsystems, such as punctuation and indexical marks. Additional subsystems are present in texts that deal with specialized domains....

It is also evident that writing that is (or purports to be) glottographic may serve – as we learn from Greek nonsense inscriptions on vases or Japanese T-shirts with messages in dubious English (or non-English) – other functions: e.g., to communicate prestige directly or to connote cultural capital. Even writing that straightforwardly notates spoken language is overdetermined, in the sense that it can perform other functions besides. In conclusion, then, we should perhaps better view glottography not as a type of writing but rather as a function of one subsystem within the system of writing.8

     Since the multimodal sign is already the harbinger of this metaprosodic operation which modulates the form of its function and the function of its form, it is always potentially open to exploitation for visual and phonic cueing. The variational mappings that these functions or principles encode, in both their disjunctive or conjunctive modalities, can therefore account for the complex interactions between speech- and script-oriented textuality from the featural to the discursive levels of the linguistic hierarchy.9 We can witness its footprint in diverse textual effects in such modes as the “stream of consciousness” technique of the modernist novel and in the so-called “free indirect style” which blurs the border between narrated and reported speech. More generally, the “silent reading” we routinely experience today in literary and everyday context of printed texts (like textbooks or newspapers) can be accompanied by something called “implicit prosody.” As advanced by Anglo-American linguist Janet Fodor, the implicit prosody hypothesis (IPH) assigns a “default prosodic contour” on linguistic strings, eliciting “not only information about syllabification of individual words but also information about their metrical structure.”10


Silent reading and implicit prosody

     It is no accident that modern literary history after the ascent of print technology became the privileged site of silent reading as the negotiatory space between the oral and the literal. The introduction of spaces between words in the age of manuscripts,11 the eventual massive standardization of linguistic characters and layout—together with the political rise of monolingual states and grammar, the advent of a reading public with leisure time and education, and the dramatic increase in printed texts in a capitalist economy—all worked together with the institutionalization of literature in shaping the dialectical relationship between the verbal and the visual in the history of reading and language studies. Even if we disagree with literary culture scholar Elspeth Jajdelska (2007) when she claimed that the “contrast between ‘the reader as a speaker’ and ‘the reader as a hearer’” in the situation of silent reading is more absolute than the opposition between oral versus literate or spoken versus written (where gradations in scale are in force),12 we can appreciate how her study of silent reading in the context of literary history, by framing the relationship between the verbal and the visual in oratorical and rhetorical terms, indicates the contrastive role of speech-like prosody in the stylistic redesign of modern narrative texts toward silent or graphic reading motions. This leads her to consider specific scriptly matters such as punctuation and syntactic grouping as a means by which speech-oriented prosodic rhythms are stylistically differentiated from a more graphic or scriptly prosody where the narratorial signature emerges as a literary “voice” set apart from (narrated or reported) “oral” speech.

     The stylistic changes (co-extensive with the manipulation of graphemic and syntactic procedures) in written prose due to the ascendancy of the silent reading model where the reader is a hearer is summed up in passages such as below:

I suggested that there was a significant increase in the number of fluent silent readers at the end of the seventeenth and beginning of the eighteenth centuries, and that this increase led to important changes in prose style and genre because fluent silent reading assumes a model of the reader as a hearer rather than a speaker. I suggested that writers using a model of the reader as a hearer found it easier than their predecessors to introduce longer and more frequent pauses into their writing by means of punctuation, and that greater use of pauses promoted an increase in the number of sentences in the form of autonomous units bounded by pauses. Sentences of this kind represented a significant departure from the previous long strings of loosely connected clauses, by means of which the reader as speaker had been allowed to continue speaking without a pause until a topic boundary was reached.13

     I would be inclined to interrogate, however, Elspeth Jajdelska’s claim that the reader-as-speaker model has been largely eclipsed today (rather than saying that it is the metaphysical idea of “speech” which has become more difficult to locate beyond stylistic overmarking), even if what she says could be seen as a corroborating instance of Richard Cureton’s statement about the “rise” of visual prosody in the 20th century:

Silent (and fluent) reading in modern industrialized societies is now so widespread that the older model of reader – that of the reader as a speakerhas become almost invisible. The only place for this model might be in scripts for film, television, or theatre, and even there the ubiquity of silent reading means that dramatists feel obliged to give their readers copious stage directions. In the early modern period, such directions might have been rendered unnecessary by the prevalence of a uniform skill of reading aloud. The metaphor that ‘reading is hearing’ is ubiquitous today, as common in the scholarly worlds of linguistics and literary criticism as it is in ordinary speech.14

     Furthermore, since what we are dealing with here are idealized or analytical models, we need to check them against actual practice where the four corners of literacy skills—namely reading, writing, speaking, listening—are multimodal and multisensory functions that run parallel to each other in all linguistic activities, even if one of them seems to be in “silent” mode. We are already familiar with the poststructuralist idea of the iterative nature of language, and how the image of the polyvalent sign is implicated in all linguistic performance—from production to reception, as a recitation and remediation of cultural memory.

     This is the reason why we need to add a complication to the unilinear modeling of a silent reading which simply overlays a phonic prosodic contour over the graphic surface as a cognitive means of parsing the textual signal. This is not to say that, as multimodal functions, literacy skill domains are not indexed as such by pragmatic context; it is only to question any absolute status in the divide among them and to direct our attention to the pragmatic operation which overmarks them as such and such type of action. Secondly, it is also to underline the antecedent or precedent textual memory behind any speechly utterance where the speaker is already someone who is thinking or reading aloud, as someone who is irreducibly in the act of recitation.15 That is, if a phonic model is involved in the implicit prosody of the silent reading of texts, it is because a graphic model already informs the phonic model of speech recitation in a sort of feedback loop. Following the blurring of the essentialist division between spoken and written language toward what Geoff Hall calls the turn to “discourse,”16 we can then view both “mirror image” prosodies as its facets, as the operation of the stylistics of discourse indexing pragmatic information on top of such linguistic binaries. The disjunctive/conjunctive rapport between verbal and visual prosody, therefore, is less an essentially static opposition than a dual signature made palpable by stylistic overmarking. Because of this, it even happens that raising the scriptly gradient of a mark makes it all the more adapted to bear the speechly prosodic signature, as in the case of italics, comic book visual typography, or in the New Media development of “prosodic fonts”17 (which recall the agile typographies of E. E. Cummings’ visual prosody).

     The phonographic solution space which must always be invoked in all reading/speaking motions follows, then, a feedback loop, or, more accurately, a spiral. Such a dynamic can clarify what the Belgian linguist Caroline Féry noted as the “puzzle” in Janet Fodor’s implicit prosody hypothesis and its role in the processing of syntactic structure.

Fodor...took a more general perspective on the issue of implicit prosody. The puzzle she was confronted with is the following: if prosody in reading must be projected on the basis of the lexical string and the syntactic structure assigned to it, as proposed by the syntactic hypothesis (30): how could prosody ever influence the assignment of syntactic structure to the lexical string? In such a view, prosody should only redundantly interpret the established meaning, as its only function is to represent sentence structure.

(30)  Syntactic Hypothesis: The prosodic structure of a sentence is immediately determined by the syntactic structure. Different prosodic structures emerge automatically when the syntactic structure differs.18

     To resolve the question of whether it is the prosody which assigns the syntactic structure or if it is the written syntax that dictates the prosodic contour,

Fodor proposed that sentence comprehension involves both a syntactic and a prosodic parser operating in parallel. At the same time, she went a step further than Bader in proposing that the prosodic parser is always active, also when it is not necessary to disambiguate a syntactic structure with prosody. The role of the prosodic parser is more general. The result of the syntactic parser... is provided with prosodic phrases. This prosodic parser is active in silent reading, even though no real prosody can be heard. Instead, a default prosodic contour is projected onto the stimulus, which may influence syntactic ambiguity resolution. ‘Prosody is mentally projected by readers onto the written or printed word string. And–the crucial point–it is then treated as if it were part of the input, so it can affect syntactic ambiguity resolution in the same way as overt prosody in speech does.’19

     From this formulation, it might be concluded (wrongly) that the prosody-syntax interface mapping is direct, exhaustive or definitive in the model of the full isomorphism between the phoneme and the grapheme in idealized orthographic writing. (This is a question addressed by the notions of the transparency of signs in interpretation.) Up until now, I have not given much attention to the definition of “prosody” itself, in the same manner that no extended definitions were given to the plethora of terms that I have mentioned, like “syntax” for example. It is true that definitions are notoriously difficult and, if we heed Wittgenstein’s famous admonition (“What should we gain by a definition, as it can only lead us to other undefined terms?”), would force us to remain within the limits of what can be considered sufficient for discussion, specifically in relation to the functional nature of the prosodic-syntactic “parser” working in parallel in silent reading motions.


The residue of prosody

     In the case of “prosody,” then, that of Anglo-American linguist Robert Ladd seems to be the most thorough in historical and theoretical content, even if at the end of it all he still remarks that “One justifiable conclusion from this list of possible definitions is that the current sense of ‘prosody’ really is incoherent.”20 Caroline Féry, after giving a working definition of “intonation” and “prosodic structure,” gives a shortlist of the history of “prosody” as a phonological term in 20th century linguistics:

Even though the terms ‘intonation’ and ‘prosodic structure’ are straightforward, this is not true for the term ‘prosody’.... The term has a Greek origin and until the twentieth century was used to denote the musical accompaniment of the segmental parts of the words and sentences (or verses in poetry). It is only with Trubetskoy (1939) that the term entered linguistics and denoted other properties besides word-level accentual features. According to Trubetskoy, prosodic features refer to all rhythmic-melodic aspects of speech. Consequently, the term has been used as a synonym for suprasegmental phonology for several decades, until Bruce (1977), Liberman (1978) and Pierrehumbert (1980) separated tonal structure from metrical structure, that is, the accentual properties of speech. The next step leading to the definition of ‘prosodic structure’ as used in this book was the development of the prosodic hierarchy by Liberman (1975), Selkirk (1984) and Nespor and Vogel (1986).... In sum, prosody is characterized by intonational features, and intonation is realized within prosodic domains. For this reason, since the overlapping of intonation and prosody is intrinsic to the organization of speech, an overlapping of the terminology is unavoidable. As a result, the terms ‘intonation’ and ‘prosody’ are often used interchangeably. The shared meaning of intonation and prosody includes what is sometimes referred to as ‘suprasegmental phonology’, and thus tonal structure, pitch accents, phonological boundaries, duration and intensity. The terminological confusion is acknowledged in this book, and no attempt to escape it or correct it is provided.21

     In the passage below where the quote from Ladd was taken, we can see how the various attempts by sophisticated theories to define prosody always seem to have involved a series of persistent dualisms, something that Ladd builds upon for an interesting commentary of his own concerning the relationship of prosody with the alphabetic writing system:

One justifiable conclusion from this list of possible definitions is that the current sense of ‘prosody’ really is incoherent. In that case, the expansion of the term among the ancient grammarians and in modern linguistics is essentially an artefact of alphabetic writing: prosody is a grab-bag of things that are hard to write with a string of symbols. Yet the phenomena in the prosodic grab-bag do seem to fall into two main clusters. One involves phonetic properties that are often thought of as running in parallel with the segmental string—elements such as pitch that are phonetically quite independent of segmental articulation, and elements that apply to stretches of speech longer than individual segments—while the other involves hierarchical structure and syntagmatic relations. The oppositions discussed in the previous section can be separated into two groups along these lines: source vs. filter, non-verbal vs. verbal, suprasegmental vs. segmental, and slower vs. faster periodicity are all about parallel phonetic streams and long domain properties of the segmental string, while the syntagmatic/paradigmatic distinction (and its forerunner prosodic vs. inherent) involve the hierarchical structure of the string itself.22

     As Ladd explains in the following quotations, prosody appears to be an “artefact of alphabetic literacy.” In linguistics that is still current today, the use of the term “suprasegmental” is responsible for the “impression that prosody is little more than the residue of segmental transcription” (my emphasis). Since it is composed of “ordered strings of atomic elements” involving the “segmental idealization” of language as in IPA transcription, alphabetic writing is “intrinsically poorly adapted to representing certain phonological features that are generally taken to be ‘prosodic’.” Among the main prosodic features alphabetic writing seems to be ill-equipped for in representing are indications of duration and pitch. Apart from linear seriality and diacritics for length, Ladd notes that nothing systematically represents “gradiently variable temporal features that have other linguistic functions,” such as those marked by “transcriptional abstractions like boundary symbols in IPA transcription and punctuation in ordinary alphabetic writing.” Tones which move concurrently with segmentals “[appear] to be a phenomenon apart” in alphabetic writing history “where phonemic distinctions of tone are unusual” and are bound to be less represented. Beyond the core features of pitch and duration, investigations of the global phonological properties of language occasioned by the speech-centered linguistics of the 20th century have radically expanded the domain of prosody.

In less than a century, then, the modern linguistic terms prosody and prosodic appear to have undergone a development that is parallel to that of Ancient Greek prosōdia. Starting from a specific reference to word-level tonal or accentual features, both the classical and the modern terms expanded to include a range of other phenomena not normally indicated in writing, including in particular, features of quantity and cues to the grouping of words into phrases. One possible interpretation of this parallel is that modern scholars have rediscovered a valuable insight into the organization of phonology, and that there is some natural unity to the range of things grouped together under the expanded meaning of the terms. Another is that the expansion of the meaning simply reflects the biases induced in classical times by alphabetic literacy and in the modern era by IPA transcription: anything not written with consonant and vowel letters must by definition be something else.23

     The noted expansion of the prosodic domain seems to be directing us toward a reorientation of conceptual priorities. We should probably stop looking at prosody as the artefact of the alphabetic writing system, and consider instead that the more compelling order of events should make the alphabetic writing system the residue of the dynamics of prosody. Such an inversion of the conceptual order between the “source” and the “filter” should not be seen as the revival of the phylogenetic priority of speech over writing since the dialectical relation involved is ultimately that of the continuous against the discontinuous in the segmental nature of the alphabetic system.24 From a philosophical point of view, the prosodic waveform in general dips back and forth into the dynamic field of unbroken temporality. By its presence in the universe, it is human subjectivity which introduces a break, a schism, in the continuum of things. The anthropic presence, therefore, marks an event, or the arrival of the event as such. Previous to it, nothing has taken place, that is, nothing in terms of the punctual by which any event can be grasped. Language and subjectivity embody those punctual events since they register the moment of discontinuity, the break or the (anthropic) gap which happens (like Lucio Fontana’s “tagli” works). The spectrogram called “consciousness” parses the continuous into the discontinuous, into variously segmented signals processed by the multisensory organs of the body.25


Lucio Fontana, #8 Noi, dopo i tagli, (c. 1960).

     Whether or not everything that was parsed gets encoded in an alphabetic or syllabic notation system or another grammatological system is one question; and whether or not such multichannel parsing is cognitively necessary in the production of redundancy and salience is another. As a continuum subtending physics before its parsing into the phonic, graphic, and other media channels, the prosodic stream is only potentially glottographic and nonglottographic in the punctual moment which opens and closes segmental solution spaces; that is, in a cycle of resolution and dissolution of segmental perception. (I am tempted to ask this early: won’t that be the same continuum that chromatic music and avant-garde poetry are attempting to signal within the anthropic break, whose parallel visual effort would be a move toward the fuzzy, the aesthetics of the informal, the retreat of perspective and the line, the use of abstraction and the asemic, the replication of hyperforms, and the reconfiguration of phonographic solution spaces in visual prosody?)


Segmental solution spaces

     We need to make more explicit, then, the relationship among prosody, the alphabetic notational system, and segmental thinking, and its connection to the causal bind of Fodor’s IPH for silent reading. Interestingly along this line, the American linguist Alice Faber, echoing somewhat the causal bind in Fodor’s IPH, puts forward the idea that segmental thinking is the epiphenomenon or after-effect of alphabetic writing, conditioning a certain global bias in the description and analysis of language systems:

In fact, segmentation ability as a human skill may have been a direct result of (rather than an impetus to) the Greek development of alphabetic writing. Thus, the existence of alphabetic writing cannot be taken eo ipso as evidence for the cognitive naturalness of the segmentation that it reflects. Given this conclusion, it is necessary to pose the question of why virtually all linguists have fallen prey to what Ladefoged (...) has, in a comparable context, referred to as a phonemic conspiracy. That is, we as linguists feel that, because we can describe linguistic systems in terms of phonemic segments, we must do so. That we are influenced in this practice by the structure of the alphabets that we use in our ordinary lives is obvious. As far as I know, every technical linguistic tradition that refers to segments arose in an alphabetic milieu or was influenced directly by such a tradition. There is no segmental analytic tradition not supported in this way by an orthographic tradition. In contrast, the indigenous Chinese linguistic tradition, described by Halliday (1981), has as phonological primitives syllable initials and finals, that is, onsets and rhymes. This analytical division is not supported by the logographic Chinese orthography, a lack which strengthens the force of the analysis.26

     The general bias for segmental thinking which alphabetic writing occasions can thus be seen in the postulation of the prosodic-syntactic parser working in parallel in the IPH of silent reading motions. This appears to animate the abstract elaboration of prosodic structural domains (mora, syllable, foot, prosodic phrase, and intonation phrase) projected along equally abstract syntactic grammatical domains (NP, VP, PP, Predicate, Clause, and Sentence): they are clearly metalinguistic extensions of segmental logic to higher linguistic bundles. But more significant is the fact that, once again, placed externally or orthogonally to both these formal structures, are the informal prosodic physics of intonation and melody, intensity and duration, pitch and rhythm, and so on (the so-called prosodic “features”). In effect, the segmental solution space of the alphabetic system keeps reasserting itself in this dichotomy because this is how it should work as a notational system. Like the conceptual order which induces the view that prosody is an artefact, the causal puzzle in Fodor (Is syntax the projection of prosody or prosody the projection of the syntax?) would be a symptom that the segmental and sub-tonal alphabetic system is alive and well in both linguistic theory and practice.

     Fodor’s prosodic-syntactic parser works, therefore, in accordance with the segmental logic of the alphabetic solution space. Since its operation leads to the attribution of elements as either prosodic or nonprosodic and segmental or suprasegmental, the elements overmarked thus bear such and such trait less as structural properties than as the index of the operation of the segmental solution space in silent or not so silent reading motions. This is certainly how I would read the British linguist Anthony Fox’s observations below:

Although the terms 'suprasegmental' and 'prosodic' to a large extent coincide in their scope and reference, it is nevertheless sometimes useful, and desirable, to distinguish them. To begin with, a simple dichotomy 'segmental' vs. 'suprasegmental' does not do justice to the richness of phonological structure 'above' the segment; ..., this structure is complex, involving a variety of different dimensions, and prosodic features cannot simply be seen as features which are superimposed on segments. More importantly, a distinction can be made between 'suprasegmental' as a mode of description on the one hand and 'prosodic' as a kind of feature on the other. In other words, we may use the term 'suprasegmental' to refer to a particular formalization in which a phonological feature can be analyzed in this way, whether it is prosodic or not.

"The term 'prosodic,' on the other hand, can be applied to certain features of utterances regardless of how they are formalized; prosodic features can, in principle, be analyzed segmentally as well as suprasegmentally. To give a more concrete example, in some theoretical frameworks features such as nasality or voice may be treated suprasegmentally, as having extended beyond the limits of a single segment. In the usage adopted here, however, such features are not prosodic, even though they may be amenable to suprasegmental analysis.27

     There is in Fox’s comments a provision for a de-essentialized notion of the prosodic or nonprosodic, that is, as the function of a pragmatic operation where what is segmental is differentiable from the suprasegmental only on a formal level of analysis. If we return to the remark by Alice Faber, if there is no physical basis for segmentation, it certainly arises in the context of reading given the phenomenological fact that reading involves discrete visual graphic marks separated by spaces. However, such segmentation is arguably idealistic not only on the level of phonemes but also on the level of graphemes, since “the segments identifiable in an acoustic waveform, to the extent that their boundaries are well-defined, are not isomorphic with the segments of phonological analysis or of alphabetic writing.”28 Inasmuch as the continuous sound stream is hardly a basis for discrete units, as Neef explains, segmentation is but a dynamic specific to the solution space where “written representations are related to phonological representations,” or where the “regular written representation of a specific word [serves as the] element of the graphematic solution space of the phonological representation of that word.”29

     The segmentation by which the phonic-sonic stream is subjected is, therefore, embodied in its representational moorings in alphabetical solution space involving graphemic elements and the overmarking of the prosodic difference between the verbal and visual signatures. In the alphabetic solution space, the polyvalent sign is itself the diagrammatic expression of this “primary” reading motion, and the vacillation between the arbitrariness and the motivation of the sign is a debate around the redundancy mapping motions in the production of the interpretant. What characterizes the operation of the solution space, then, in silent or not so silent reading motions is the role of analogy or equivalence relations where one medium attains a coefficient of the real by being placed in a bifocal and bilateral relationship with another medium. Such an associative juxtaposition frames the other as semiotically different so that an equivalence or diagrammatic relation could be installed across the divide, between two abstract domains containing the phonemic and morphemic principles. One hears what one reads, and reads what one hears: such cross-talk between media channels would be the dialectical tension informing the metaprosodic overmarking of the speechly against the scriptly signature realized redundantly in alphabetic solution space. In this sense, the glottographic and nonglottographic are not diametrical but dialectical pairs, which implies that there is a perennial discourse (Latin discursus, “running back and forth”) between the prosodic and the non-prosodic, the qualitative difference between them being the disjunctive function of the segmental space instigated by the anthropic gap.

     How the graphematic solution space then decides which segment is henceforth glottographic (speechly) or nonglottographic (scriptly) without reference to the continuous sound stream (as “floating” abstract signifiers) would require the metaprosodic processes of channel or medium separation, designation, and diagrammatic correlation in which the phonemic principle is cyclically reassessed through the lens of the morphemic principle in the guise of equivalence gradients such as iconicity and analogy or via structural substitution principles and deep transformation algebras. The German linguist Gerhard Äugst, in an article evaluating phonologically-dependent and autonomous models of orthography in the codification of “meaning,” in giving us an idea on how to visualize the graphematic solution space and its complex operation, has equally pointed out the same co-referentiality involved in the diagrammatic correspondence operation at work in the mapping between the phonemic and the graphemic principles.

As the phonemes even in functional structuralism, but even more in generative phonology are established only with reference to morphemes, a morpheme appears - in the spirit of the double code theory - on the signifiant side of the sign as a phoneme schema, from which on the one hand a sequence of phones depending on the phonic and morphophonemic context is realized. On the other hand, the grapheme schema is derived from it through the working of phoneme-grapheme-correspondences. This graphemic schema then is also realized following allographic and morphographemic rules as a sequence of graphs.30

     Since this is not the place for the discussion of a more exhaustive technical modeling of the segmental solution space of the alphabetic system, I merely alluded to the works of these scholars to outline the metaprosodic dynamic by which the ambiguous or double coding of the sign into the speechly and scriptly signatures is executed as formal operations of that solution space. That such a solution space—overmarking the tonal and the formal, the prosodic and nonprosodic, or the glottographic and nonglottographic—continually informs the production and perception of language in an alphabetical culture should reorient us toward understanding the pragmatic bias animating not only the metalingual but also the poetic functions of everyday discourse.



 
II. The rise of visual prosody


In the twentieth century there has been a steadily increasing interest in visual prosodies. The long (but relatively peripheral) tradition of the ‘shaped’ poem (the 'pattern' poem, the acrostic, and related visual forms) has been reviewed, redefined, and placed in a more central position in poetic theory and practice. Many developments have generated this new interest in visual form: The canonization of ‘free’ verse as a prosodic standard, the influential theorizing of the great moderns (Pound, Williams, etc.), the development of ‘concrete’ poetry as a worldwide movement, the development of the post-modern aesthetic of the self-reflexive text, the rising influence of Derridean deconstructionism in the critical academy, the emergence of semiotics as an academic discipline, and general, cross-disciplinary interest in the influence of visual form in our contemporary, print-dominated societies — to name a few.

—Richard Cureton (1986)31

The notion of a silent letter is one that assumes letters actually have a sound, independent of the ones we ascribe to them. Albrow writes: ‘All letters are “silent”, but some are more silent than others’.

—Des Ryan (2016) 32


     Beginning at least with the experimental typography of Futurist poetry, the graphemic signifier as a visual artefact regains its ambiguous role in solution space, problematizing the unilinear signifier-signified polarization by which the glottographic is materially or formally fleshed out from the nonglottographic.33 After Futurism, many stylistic features of Concrete poetry—arguably one of the most important poetic movements of the 20th century after free verse—not only exploit the iconic or pictorial potentials of unconventional typographical layout or exhibit the semiotic spaces of visual prosody but also underscore the pragmatic operation where what gets assigned as the linguistic “medium” or “message” is an index of the discursive process of the phonographic solution space. 34 (Traditionally, the graphic signifier is reduced to being the “transparent” or neutral medium whose only possible inert or passive value is speech notation or logical representation; that is, to have the paradoxical status of a non-entity representing both the figure of silence and loquacity.) Even if E. E. Cummings is not considered to be a full-blown practitioner of Concretist poetics, he has made use of its signature semiotic procedures of visual prosody in various ways, ranging from “geometric-iconic”35 pieces worthy of the Concretist name down to the reversal or reconfiguration of the overmarked role of the visual and technological dimensions of the graphic signifier and its bibliographic surface as semantically- or prosodically-subordinate constituents of higher phonetic, syntactic, and formal hierarchies. As poetics scholar Richard Cureton declares, by way of providing some rationale to any research interest in the American poet’s variational stylistic, E. E. Cummings’ oeuvre is a good source for exploring the semiotic spaces of visual prosody, an area about which Cureton lists challenging aesthetic and critical questions whose serious investigation remains to be pursued.36

As we begin to entertain these questions, the poetry of e. e. cummings should attract increasing attention. While I am not sure how one would argue for this, one might claim that Cummings made more intensive and extensive use of visual form than any other poet in literary history. In Cummings' poetry one finds all of the major visual prosodies: pure concrete poetry, the 'expressive' free verse line and stanza, 'objective' or 'arbitrary' form, the 'shaped' or 'pattern' poem, parenthetics, acrostics, and many others of his own devising. And he experimented with these forms in almost every conceivable combination with one another and with other more conventional prosodies. In this area, as in others, Cummings is the ultimate 'cubist' poet, spinning a bewildering array of original forms with his 'kaleidoscopic play' (Steiner). These forms provide us with an unprecedented reservoir from which we can learn about visual prosody.37

     Whether or not the relative lack of scholarly attention on visual prosodies noted by Cureton in his 1986 essay on Cummings’ may have already been amended by now after three decades is a matter whose extended consideration must be revisited on another occasion. Some comparable terms we can encounter today are “visible language,” “silent poetics,” “visible word,” “deaf poetics,” “pictorialist poetics,” “visual enactment,” “visual form,” “graphical prosody,” “poesis of space,” “the poem on the page,” or “typewriter poem,” intersecting here and there with studies on Vispo, Concrete, and postmodern poetry.  When Cureton asked how different or similar the operations of visual prosodies might be versus other prosodies, what impresses us immediately is the plural form it comes with: there isn’t just one type of it. Half-seriously we can ask: if one is already confusing, why complicate matters further by adding another to the list? As Robert Ladd and other language scholars have indicated above, the term “prosody” wavered both in ancient and modern times between a limited and expanded linguistic conception. It referred not only to a poetic organizing principle but also to linguistic aspects or constituents for which that principle is the organizational frame. In the case of poetics, it is primarily equated with a metrical scheme, but not limited to classical foot (length or stress) prosody. In linguistics, it would mean phonological suprasegmentals, divided between prosodic domains and features. Thus, in Cureton, we have the mention of these two prosodic streams, but this time mixed in with visual prosody. The question that would come to mind, then, is what specifically they are prosodies of, and why do we have a number of them, beyond the obvious answer that they represent different traditions or approaches.


The metrical object

     From the rough (non-hierarchical) pie figure below, many possible referents behind a prosodic notion or principle are possible, intersecting in various relational or hierarchical modalities depending on the analytical perception of poetic organization. Hence, from metrical poetics and modern linguistics, we have these prosodic grounds which function like master signifieds. Nonetheless, positing them as grounds or base brings us back to the same spot: what are they speeches or structures or rhythms of anyway? The literary and linguistics scholar Christoph Küper, for example, explains somewhat circularly how the core problem of metrical theories involves the “role of the linguistic givens in different metrical types” since “metrical entities are realized by linguistic entities” and the “linguistic entities represent or count as metrical entities.”38 Yet, the presumed linguistic entities beg the question of what they are linguistic entities of to be capable of spawning a prosodic dimension. This critical question, in my opinion, would be a defining literary historical factor both in the conflicts among various currents attempting to model the metrical constitution of any prosody in general, and in the fluctuating notoriety, prominence , or place of visual prosody in poetic practice in particular.



     The British metrical scholar Kristin Hanson, surveying the generative theories of meter, reflects on the quandary of defining one problematic notion in terms of another equally problematic concept:

In the theory of Hanson and Kiparsky (1996), meter is claimed to be just such a stylization of the phonology of rhythm. This “stylization” is implemented pretty much in the way proposed by Halle and Keyser (1966), through an underlying structure (or template) and realization constraints (like correspondence rules) defining mappings of the phonological structure of language into the template, thus maintaining the formalization of meter as involving a kind of second-order experience of language. This stylization [metrical prosody] and hence this experience are specifically “of the phonology of rhythm” [suprasegmental prosody] insofar as both the templates and the realization constraints are formally like what is found in the phonology of rhythm; it is therefore in them that developments in the phonology of rhythm show their influence on metrics.

Obviously it would not be possible to spell out here the entire content of what it means to be “like what is found in the phonology of rhythm,” even if it had been fully discovered and agreed upon and articulated by phonologists, which of course it has not been; and in any case Hanson and Kiparsky (1996) formalize only some aspects of how whatever the phonology of rhythm may be shapes templates and realization constraints.39

     To speak of a “second-order experience of language” makes us ask, of course, what a “first-order experience” of the same might be like and if anyone has experienced it, without returning us to the traditional divide between “language” and “metalanguage,” “poetry” and “prose,” or between “special” and “ordinary” or “artistic” and “natural” language. Such dualisms always seem to haunt the discussion of metrical and prosodic theories, as we can glean even from the eminent metrical scholar Marina Tarlinskaja:

One of the cornerstones of metrical analysis is to differentiate actual stressing from the abstract metrical scheme. By “actual stressing” I do not mean variants of performance by different modern actors (the number of such variants is almost endless), but a “neutral” oral rendering based on what a speaker knows about grammar, phonology, and the meaning of words and phrases in his language. The distinction between actual stresses and the metrical scheme was the subject of heated debates in Russia in the early twentieth century. The scholars were perplexed: how is it possible that a poem contains so few “perfect” iambic lines, and yet the reader knows that the text is iambic? So they came to the conclusion that there must be a model for all lines, both “perfect” and “imperfect,” a general scheme, a meter. We discern the abstract scheme, the “metrical sound keyboard,” from the sequence of lines in a poem.40

     The stylized or abstract metrical scheme (or “verse design” in Roman Jakobson’s terminology) which represents the second-order experience of language would certainly make no sense without a first-order experience. The determination of what such first-order language experience might be (as a natural rhythmic phonological base or as a “neutral” performance of a generic or idealized speaker in the above examples, or as generative, grammatical, musical, or physical principle, and so on) as a stable frame of reference is thus imperative if any prosodic or metrical scheme desires to be the formal measure of the first-order experience of the language. The literary scholar and critic Derek Attridge, for instance, in his classic The Rhythms of English Poetry (1982), when he takes issue with “the reification of the foot” in classical prosody, opposes its scansion of the line against the tension afforded by “the normal pronunciation of English,” a pronunciation which should be “a true reflection of the reader’s experience,” and not following the “ghostly divisions” of foot prosody. He argues for a more faithful prosody that is based on the temporal characteristics of the “underlying” rhythmic properties of spoken English itself. Not surprisingly, though, we will read later on that, far from being based on the empirically varied way English is pronounced around the world, such rhythmic properties of English which form the backbone of his notion of a temporal poetic prosody can only be found in an idealized speaker (and not in so many “different modern actors”).41 Furthermore, the essentialist appeal to the rhythm of the language itself is combined with psychological and physicalist notions, making us ask what exactly rhythm is a movement of in the first place. These multiple loyalties appear to arise out the analytical frame itself which codes linguistic and paralinguistic phenomena into the segmented causal grammar of Subject, Verb, and Object. That is, it simultaneously locates the basis of rhythm in the (equally ghostly) ideal speaker’s actual performance of the “inherent” rhythmic structure of the language, which in turn auto-generates the rhythm the ideal speaker knows intimately and then produces competently. This is why Attridge counters the classical proposition that an abstract metrical scheme is the “external” source of the poetic rhythm of language, blurring practically, it seems, any divide between the first- and second-order of language experience.42

     The dilemma which any analytical metric frame encounters is, therefore, shaped by the need to define the criteria marking off the “metrical” from the “non-metrical” and then explain how the language of form is derived from the form of language.43 It does not usually proceed the other way since language in its “natural” state is already regarded as a self-evident given. Yet, this question seems to concern less the essential definition of what verse could be against what is not—an important question, no doubt, but which mutes a more pressing matter—than the creation of a cognitive support from which language as a “natural,” first-order experience, in part or in whole, could be posited as a referentiable object. Even if we take rhythm to be the key determining field in which metered and unmetered language could be tested—a likely reason why it occupies the central hub of prosodic investigation— the question leads us back to the metaprosodic operation that unpacks the information design conditions of the possibility of the perception of rhythm as the interplay of metered and unmetered streams (prominence and boundaries). This amounts to saying that the perceptual ground of rhythm for both “metered” and “unmetered” language is triangulated by a passage through the glottographic and nonglottographic functions of the phonographic solution space.

     In a general sense, insofar as any metrological approach would like to establish a common measure by which the perception of rhythmic unities and their order could be made possible, all could equally be called “meters.” By providing for the “formal description of the rhythmic intuitions of the experienced reader,”44 metrological frames are cognitive solution spaces where rhythm as a readable information design system is the perceptible or “surface” manifestation. In Cureton’s brief but substantial elucidation of fifteen theoretical approaches to (English) rhythm,45 one would not certainly miss the greater emphasis on the glottographic principle in the formulation of metrological frames. (Reuven Tsur’s “cognitive poetics,” based on the informational view of semiotic encoding; and David Gil et al.’s prosodic approach where all textual rhythms could be attributed a suprasegmental function, are the interesting exceptions from the metaprosodic angle being developed in this paper.) The glottographic orientation of metrological frames in the analysis of rhythm, that is, the use of “speech” as a phonological or phonetic constant, is possibly the most obvious choice given the predominantly phonocentric grounding of many poetico-linguistic conceptions in the history of literature and also modern linguistics. This is possibly due to the seriality of speech offering itself as the most basic line of reference in terms of the temporal unfolding of language. All other pragmatic and cognitive motions range along its line, so that divagations are measured against its fictional path, even though the straight forward moving line it draws is a huge shadow hiding so many detours and invisible returns.


To measure by a shadow

     Apart from the writings of Jacques Derrida documenting the predominance of the phonocentric signified in the history of writing (in the Occident at the least), contemporary language scholars have also noted the phonovocalic mimesis of poetics and modern linguistics, especially in the studies of metrical prosody.46 This is not to say that the glottographic principle has no part in the conception of linguistic rhythm, only that the use of speech as the primary ground of measurement would be—as Dr. Samuel Johnson in 1755 had said it, “to accommodate orthography [writing] better to the pronunciation [speech], without considering that this is to measure by a shadow, to take that for a model or standard which is changing while they apply it.”47 The use of “natural” speech (the first-level language experience) as the target of prosodic or metrical frames (of second-level language experience) is certainly not just a major given in the linguistic theories of meter but also a conception many poets struggle with, something we can easily recall in, say, William Wordsworth in his attempt to write using “the language really spoken by men,” putting forward the discourse paradigm view that “there neither is, nor can be, any essential difference between the language of prose and metrical composition.”48 The dissolution of the distinction between the first- and second-order experience of language, between the empirical and the formal, is actually a double-edged sword, either dissolving the notion of a “natural” ground and implying that such a ground is itself a metaprosodic invention; or, fortifying the view that there is only discourse, and that speech/writing or metered/unmetered language have no essential difference between them because they are merely indices of a larger operation.49
   
     Whichever side we take concerning this long standing debate, one thing remains abundantly clear: the glottographic basis of metrological frames has taken for granted the passivity of the nonglottographic or visual aspects of written language as a neutral and transparent signifier. In other words, the question of metricality focused the application of the principle of prosodic markedness exclusively on the glottographic plane of the phonographic solution space. Even if we concede to what the conversation scholar Deborah Tannen has observed concerning the poetic nature of conversation and all of discourse, we still would find such claim 1) to preserve the mysterious existence of poetic markedness and extend it back to the first-level of language experience grounded in speech conversation; 2) and to leave unanswered the mechanism involved which allows us to say—following her citation from Paul Friedrich, that “the reality is not poetry vs. nonpoetry but more poetry versus less poetry.”50 The appeal to the capacity of “more poetry” to fire up the imagination and move emotions seems to be an insufficient criterion when simple expletives would undoubtedly qualify under the same definition. Although a major advance in the demystification of the ordinary/special language divide, the proof she puts forward grounds itself on strictly phonolinguistic shapes and effects.51 Even so, it is significant that she traces the poetic qualities of ordinary discourse to its rapport with the prosodic structure of literary discourse as presented in the visual format of printed texts.

I argue that rather than distinguishing poetry from ordinary talk, line and verse structure is the way written poetry captures in print the rhythmic chunking of talk. In other words, line breaks in poetry encourage readers to perceive the poetic discourse as they perceive all spoken discourse—in spurts. As Chafe (1980) has demonstrated at length, language in oral discourse emerges not in a steady stream but in small chunks segmented by prosody, intonation, pausing, and discourse markers such as “and,” “but,” and “yknow.” Thus, when any oral discourse is transcribed, its comprehension is facilitated by transcription in (poetic) lines rather than undifferentiated (prose) blocks.52

     It would appear that we have simply returned full circle to the problem of first- and second-level language experience, this time in the context of an implicit prosody traversing the organization of the perception of discourse, but a prosody linked to cognitive processing facilitated by information design afforded by the aesthetic format of “poetic” transcription (segmentation by chunking). The persistence of this divide can be attributed to the continued use of the categories of “spoken” and “poetic” discourse (or “prose” and “poetry”) as if the prosodic markers which united them both were the sole property of the speech act itself as first-level language experience (when we are never in the presence of naked speech, but socio-cultural discourses and the material technology of their propagation). The passage above, though, implies that this distinction is actually a cognitive and pragmatic operation having to do less with whatever “speech” might be in all its eventuality than with the physiological and cognitive conditions of discourse production and perception; that is, in the line of cognitive poetics, as information design, so that prosodic chunking or segmentation pragmatically accords with the aesthetics of perception (following the Greek term aisthêtike more than its modern sense of a theory of beauty). This aesthetically-oriented prosodic process is not accomplished on thin air but must be worked out on extensible media (sound, image) whose condition of informationality requires redundant encoding or cross-mapping among the senses.

     But it would be mistaken to swing around and view the implicit prosody at work in discourse perception as the sole property of the script act or the graphic medium. This would only reverse the medium nativism, instead of considering the speechly and scriptly signatures as the result of a metaprosodic operation where both are the products of a continuous application of prosodic markedness in phonographic solution spaces (as stereotyped, for example, by the dual but unilinear and exclusivist overmarking of Sr vs. Sd in the Saussurean view of the sign). This process renders the graphemic aspect neutral or unmarked, without a voice of its own, silent, passive, and transparent, even when, or even more so when, overlaid with an implicit prosody. At the same time, in this muted state, strangely it becomes semiotically marked— primarily under a phonocentric regime of signs attributed with subjects of discourse53—as the glottographic signifier of an enunciative stream more prosodically sonorous and present, transcending its determination as an image triangulated in graphematic space.



Eugen Gomriger’s “Silencio” (1954) could be considered the perfect manifesto of visual prosody, if ever there was one. It generates not the silence of graphemes subordinated to a syntactic or poetic line, but the prosodic articulation of silence, fulfilling via paradoxical subvocal performance the silence it was deprived of by not fulfilling it. By not being silent, it is not allowed to be what it is, and thus gets silenced much more effectively.

     Tannen’s transformation of an “undifferentiated (prose) block” of text into a series of lineated (“poetic”) segments or chunks illustrates how fundamental the role of the graphemic field is in visualizing the implicit prosody at work in discourse. As a general application of the IPH syntactic and prosodic parser, the translineation she follows simultaneously undermarks “prose” as a grey field of writing, and overmarks the newly segmented graphical layout as an exemplary notational representation of “speech” and “poetic” rhythm. To quote only a portion relevant for our discussion:

(1)
And uh-and I was going into the city, from Queens? And I was standing in a very crowded car. And I remember standing I was standing up, and I remember holding on to the center pole, and I remember saying to myself there is a person there that’s falling to the ground. And that person was me.

(2)
And uh-and I was going into the city,
from Queens?
And I was standing in a very crowded car.
And I remember standing
I was standing up,
and I remember holding on to the center pole,
and I remember saying to myself
there is a person there
that’s falling to the ground.
And that person was me.54

     After identifying ease of reading as a cognitive virtue of the second format (resulting from the use of the “line breaks as cues to the segmentation of ideas... to slow the reader down”), Tannen reaffirms the prosodic and syntactic connection between verse and oral discourse in the way the former “captures in print the larger episodic units that are characteristic of oral discourse,” and how its “breaks” and “line spaces” are correlates of “discourse markers [such as] ‘um’ and ‘and uh’ as well as pauses and hesitations.” Beyond the anaphora, lexical refrains, and the syntactic parallelism which Tannen elsewhere identifies as literary devices widely used in conversations, we can notice, too, how the manner she fleshes out the glottographic from the (less glottographically obvious) “prose” block above can equally be parsed into a strophe of potentially five hexameter plus one pentameter and trimeter lines:

(3)
And I was going into the city, from Queens?
And I was standing in a very crowded car.
And I remember standing I was standing up,
and I remember holding on to the center pole,
and I remember saying to myself
there is a person there that’s falling to the ground.
And that person was me.

     Thus, “hidden” in the graphic layout of (1) and (2) is a possible draft of a more symmetrically visual version of the same passage, a version which follows the “serious” style of classical stanzaic formats favouring full-stop lines and canonical syntax. It can be argued, therefore, that the manner by which the graphemic elements were arrayed on the display surface contributes to perception that (2) is more colloquial than either the prosaic block of (1) or the neoclassical patterns of (3). The discourse filler < uh > and the enjambed prepositional phrase < from Queens? > could be seen as the stylistic markers which enhance this perception. The predominance of the trimetric line in (2), however, actually follows the scansion of the “undifferentiated” prose block into the pattern of the “Common Meter” used mostly in ballads and conversational parts of classical dramatic and narrative poetry. That is, what we are given as conversational is actually a long standing feature of literary writing.



A comparative parsing outline of the first two lines of v. (2) showing “smooth” segmentation alignments between the syntactic and the prosodic grouping as mapped in graphematic solution space. The subordinate and consecutive positionality of the additional complement (“from Queens?”) is signalled via its relocation on a new print baseline, reinforcing the prosodic rise in the intonation demanded by the shift to the rhetorical interrogative within the same syntactic string. The semantic “transparency” results from the simple parallelistic or redundant mapping among the syntactic, speechly, and scriptly realizations of the utterance. This map is, hence, “utopian.”


     This is not in conflict with Tannen’s objective to demonstrate the poetic qualities of discourse. She does show us the stylistic continuity between the poetic and non-poetic registers of language, but she traces the ultimate source of this unity to a linguistic typology of speech rhythms that did not emphasize the role of the visual dimension in the transformation of mute marks into a glottographic image. As a site of quotation, the speech act itself is a script act,55 deploying the image of language as it has been imagined and re-imagined through various media. The prosodic parsing or chunking of (1) into trimetric lines (here credited with carrying the speechly signature) is only a specific organization of signs on visible media. This is to say that it is ultimately no one’s speech in particular but an instance of what Deleuze and Guattari would have meant when they said that all language is “indirect discourse.”56 (It is not an individual speaking, but a regime of signs deploying its locutionary registers where the subject of enunciation is both medium and message.)  As a codified style, the textual parsing into version (2) is not speech itself, but the representation in visual format of a specific stylization of speech. In this particular sample, we can see how the speechly signature is attributed to be the property of upper syntactic and textual organization in mimetic or isomorphic relationship with the phonetic and prosodic dimensions of the verbo-acoustic stream, instead of being viewed as the reiteration of the graphematic prosody of the glottographic function of the writing system.

     Even though we would think that the default value of phonetic writing is glottographic, present in either prose or poetry, it would appear that this must also be made more explicit by increasing the phonovocalic gradient of one region of writing over another. In the graphematic world of silent letters, some must be made more silent so that others could speak. That is, for phonetic writing to look or sound scriptlier in some iteration, but speechlier in another, a metaprosodic operation must be enforced in which the syntactic and the prosodic dimensions of language could be posited and set off against each other in a dialectical relationship of sameness and difference, formality and informality, or vocality and subvocality. This would entail the designation of some aspects for the syntactic or the prosodic which could be used as linguistic and pragmatic Archimedean reference points (that is, as first-level language experiences). Caroline Féry, for example, gives us an idea of the quandary involved in the mapping of the diagrammatic relationship between intonation and punctuation—that graphic subsystem mediating between the strange worlds of syntax and prosody:

With the exception of punctuation as a rough indicator of higher prosodic phrasing, intonation is not part of the written versions of language, and even when it is present in the writing system in the form of commas, semicolons, full stops etc. it does not necessarily impose a concrete intonation pattern; rather it sometimes gives an approximate idea of the syntactic structure. More specifically, a full stop does not tell a reader whether the sentence ends with a falling or a rising pitch. In the same way, a comma does not necessarily indicate a break in the flow of speech. The reader may choose to ignore punctuation in places. Thus, with implicit prosody the only exception... intonation is concerned exclusively with spoken language. In this it differs from other components of grammar.57

     Hence, the question which preoccupied the American linguist Wallace Chafe—in his study of how “punctuation units” (the supra-lexical form of orthography) in the grammar and syntax of written texts matched with the “auditory imagery of intonation” (the locutionary form of orthoepy)—is not a superficial problem of language.58 (He calls it the “covert prosody of written language,” echoing Fodor’s IPH, but possibly some years ahead of her.) Instead, it represents the articulation of the quintessential problem of diagrammatic mapping among the various subsystems of language such as orthography, orthoepy, phonetics, morphology, phonology, grammar, syntax, numerals, typography, graphematics, bibliographic codes, semantics, pragmatics, and verse prosody (a listing which courts controversy, of course). It won’t be surprising then to read that prosodic-syntactic boundary markings oscillated depending on text, context, and circumstance, 59 and that punctuation units (in written texts) could be longer than intonation units (in spoken or recited texts). Concerning the first observation, Katy Carlson gives the current thinking to explain the syntax/prosody asymmetry in a formulation which pushes the subsystems to their conceptual limits:

Does this mean, then, that a listener can posit a syntactic boundary whenever (and only when) a prosodic boundary occurs? It seems not, although much research has attempted to predict the position and strength of prosodic boundaries in speech from syntactic structure Selkirk (1986) presented one of the simplest predictive systems, with prosodic phrase boundaries predicted at the end of any syntactic phrase. Other researchers have been pessimistic about the general predictability of prosodic phrasing, with Gussenhoven (2004) pointing out that phrasing can be affected by numerous factors outside syntax, such as the position of accents and the phonological length of lexical items and syntactic phrases. Recently, Watson and Gibson (2004; Watson et al. 2006) developed a system of predicting prosody that acknowledges that most boundaries are optional and that they are affected by factors like the phonological length of phrases. Overall, it is clear by now that it is rare for a single prosodic phrasing to be the only way to phrase a particular syntactic structure. Instead, there are many possible prosodies for any given sentence with different features and consequences.... Is the prosody of a sentence predictable from its syntactic structure? Not entirely, although prosodic boundaries often do align with the ends of syntactic phrases and clauses.60


The divagation of the line

     In Carlson’s expression of the usual problematic rapport between the continuous and discontinuous in the parsing interface between the prosodic and syntactic domains, we can at least note that it is formulated in terms of its visualization as a LINE. Despite the tendency to treat it as a natural or concrete phenomenon, the line is an abstraction representing the diagrammatic field where various prosodies intersect and struggle for identity and dominance. Essentially invisible, the abstract line is made manifest by the dialectical rapport that tugs and pulls on the textual body, with one pole carrying the syntactic attribute of the prosaic thread running unhampered across all of space; and the other pole bearing the prosodic signature punctuating syntax with its periodic intervals, carving up space into discrete segments and bracketing the textual body from the page. In fact, this is what (undifferentiated) prose is: it postpones the rhetorical intrusion of the page and tunes down the segmented appearance of speechly discourse, able to transect prose only under a grammaticalized set of punctuations such as the indented line and inverted quotes—a break of air in the grey sea of scriptio continua. This differential bibliographic manoeuvre makes us forget that all linguistic segments are punctuations, given that their discrete inscription divides the continuity of surfaces into an alternating series of figures and gaps.61




     Through the justified bracketing of the textual body, the typographical line becomes possible as a referential object, a metaprosodic operation that we can spy in Man Ray’s infamous “poem” under erasure. The British scholar and critic Richard Bradford, in his important study of the visual prosody that he traces as far back as John Milton’s blank verse, provides us with the following commentary:

Man Ray’s ironically titled ‘Lautgedicht’ (1924) involves an ingenius dismantling of the thesis that what we see on the page is a model for vocalization or an accurate record of how poems work. Even without Man Ray’s droll provision of a title we would know instinctively that what we see is a poem, of sorts, for the simple reason that it triggers our recognition of the way that the typographic layout of verse foregrounds its stylistic fingerprint, the line.62

     The line as defined is, of course, consequently overmarked by various means and measures on many fronts, often via a set of redundant operations which overcode (or thwart) it for the sake of both visual (nonglottographic) and verbal (glottographic) recognizability (pausing/spacing, rhyming/punctuating, syntax group/stress group, soft/hard enjambment, etc.63). In Chafe’s experimental survey (bringing us to his second observation), this directly leads us to the notion of length, where the intonation unit is given a mean span of five to six words, while the punctuation unit with more or less nine.64 This measurement is then correlated with the limitations imposed by memory on the cognitive apparatus, interpreting the segmental chunking operated on the line as the visual and verbal punctuations of the brain’s processing power. The entry on “Prosody” in The Princeton Encyclopedia of Poetry and Poetics connects this length to the neoclassical measures of poetic discourse:

More recently, there is agreement that information is sectioned into “chunks” because the brain appears to process it that way, in both the making and receiving modes.... The chunks present the way the speaker is thinking about the information’s importance as foreground and background. A chunk may have up to five or six prominence peaks, a number which appears to correspond to the brain’s processing limit. This limit matches that noticed by Chomsky and Halle, and it matches the upper number of peaks in poetic lines in the world’s langs.—besides the pentameter, the line has a strong tendency to group into three and three (*hexameter), three and four (*fourteener), etc. But in poetry, these chunked linguistic groupings coexist with the rhythmic groupings, working sometimes within, sometimes across the established rhythmical set of a poem. The cross-structuring creates potentials for vocal realizations that open up a multivalent, polysemantic verbal design.65

     As defined by the mean length of the pentameter, the neoclassical line (henceforth “p-line”) is the goldilocks zone designating the diagrammatic utopian field of the mapping motions among the subsystems of language. This is where the grammatical (syntax) and the rhetorical (discourse) aspects of language could have the greatest incidence of segmentational alignment, maximizing the redundancies required in the production of informational and aesthetic salience in both iconic and structural modes. Thus, although grammatical and rhetorical segments are distinguished since ancient times by traditional punctuations and pauses respectively,66 the correspondences between them are brought into focused alignment by the pentametric length because of the way it reflects the cognitive threshold of oral and visual reading. Both the 6° parafoveal vision and respiratory cycle average span have been observed to converge around the length of five to six words, with a possible limit of nine to account for predictive reading afforded by syntactic or lexical memory.


A graph representing the mean span of reading, with the point of fixation indicated by the eye icon (from Schotter et al).67

     As a mean span, the p-line is, therefore, an idealized measure, but is encoded as a specific graphic and phonic length. To speak of the specific complexes of the line in graphematic terms, we can adopt with some revision the term “abstract supragrapheme” (henceforth ASGr) from the Swiss linguist Peter Gallman. In his article where he provides a preliminary non-exhaustive taxonomy of graphic means,68  he refers to the graphematic line as a modification of the boustrophedon into a series of broken line rows. He explains that the “linearity” of writing gets “partially broken up” in this new arrangement, opening the possibility of building two-dimensional ASGrs like “text-blocks” (presumably held together by other abstract cohesion principles). However, Gallman did not specify what was responsible for the emergence of the “broken up” state of the line and what kind of operations it could perform as an ASGr. That is, apart from equating writing directionality with the line concept, Gallman also neglected to mention how the line became perceptible if it were itself a type of ASGr. If we follow his own characterization of what ASGrs do, it would involve identifying the mechanisms which would allow the emergence of the LINE as a perceptibly marked (aesthetic) object:

While for concrete supragraphemes it may be easy to keep grapheme and supragrapheme apart as two graphic means, abstract supragraphemes can only be understood as instructions to transform unmarked graphemes into marked graphemes. 69

     To give an example of this process, a “marked” grapheme can be obtained by applying the ASGr “capitalization” on a minuscule and, with the additional “overlay” of a second ASGr “italicization,” transforms the resulting majuscule into an italicized grapheme: < y > becomes < Y >, then finally < Y >. (Such a unidirectional process cannot but lead to the establishment of formal substances carrying integral identities which resist reconversion. In principle, nothing prevents envisioning ASGr conversion to go in the reverse direction, with the consequence of dissolving the fine distinction—within the writing system—between concrete and abstract, basic/unmarked and overlaid/marked graphemes or supragraphemes.70) For the matter of the line, we can enumerate some possible ASGrs which would allow it to become perceptible as a signifying structure different from those which avoid or evade from it (a random cluster of letters, a one-word advertising, free verse, Concrete poetry, collage, nonlinear or Chance poetics, or Book art). Moving between the linguistic frame and the thresholds of typographic and bibliographic codes, some of the potential ASGrs may include Punctuation (since it marks up a turning point and converts the line into a hyper-punctuation unit71), Capitalization, Directionality, Length, Leading, Spacing, Justification, figural Orientation, and Layout, all of which must work in concert to solidify the visibility of the linguistic line as an abstract form segmented by the syntactic-prosodic parser in graphematic solution space. How these graphematic aspects are handled would reflect the varying stances concerning the verbo-visual rapport licensed between the grammatical (syntax) and the rhetorical (prosody) in the history of speaking, listening, writing, and reading.


Fearful asymmetries

     To think that there are many moving parts that must be coordinated to push the line into visibility as a cohesive ASGr, it will not surprise that the constant historical and literary re-evaluation of its complexes would lead to the various metaliterary and metalinguistic debates and innovations in the history of poetic discourse. E.E. Cummings’ output, which covered a good deal of four decades, would be an exemplary record of the divagations of the line given the stylistic range and variety of forms with which and through which he worked.

     A portion of the critical approach to Cummings usually focuses on the deviational or idiosyncratic aspects of his output,72 an initial or impressionistic assessment which can strike any reader, but which becomes more difficult to maintain given an understanding of 1) the difficulty of delineating the grammatical core and norms of language73 and 1) the observed prevalence of grammatical scaffolding in his compositions.74 To see these two tendencies as diametrically opposed, or to limit any analysis to the grammatically “transgressive” appearance of Cummings’ oeuvre, would be to miss the whole point that language practice is NOT limited to the confines of a grammatical system. Although the deviationist accounts can provide a good description of the surface “idiosyncrasies” of his poetry, they fall short in supplying a convincing motivation for his practice, limited as they are to humanist or romantic concepts like artistic freedom, a desire to be unique or to deform language, or a rebellion against conventions. The exclusive equation of language with grammar (seen in the positioning of canonical grammar as the governing “deep structure” of his pieces) is perhaps one of the most enduring results of the schematic conditioning done within political regimes of literacy.75

     Going beyond these notions, the central question that language practice must confront is the existence of persistent asymmetries that inform the cross-mapping motions among the various semiotic subsystems involved in the phenomenon of language as a dynamic ecological field of complexity.76 The Cubist angle mentioned by Richard Cureton is a step toward this direction with its interrogation of unilinear perspectives. In writing, this would be any monopolist metric based on the presumed linearity of speech prosody or phonological rhythm, or word-order grammar. It presumes that writing is an isochronic and isomorphic map of a self-evident ground of rhythm, instead of the field of conflict among warring tribes of posited rhythms envisioned to be encoded by ''passive'' graphic means.

     As someone educated in the Classics and as a writer worth his salt, E. E. Cummings also worked with and against traditional literary forms such as the sonnet and its pentametric as well as stanzaic organization. It was a poetic archetype he did not simply adopt as an ahistorical form.77 In the 1935 collection No Thanks, “generally recognized as Cummings' most experimental volume,”78 we can see how the “iconic” sonnet form became a wandering shadow in the din of other major prosodic streams such as free verse, Concrete poetics, and other avant-garde movements which proliferated during the modernist period. The in/famous piece # 13 of the collection, “r-p-o-p-h-e-s-s-a-g-r,” is possibly one place among many where Cummings’ interrogative stance in relation to the traditional prosodies of poetic discourse would find its most paradigmatic execution. If the sonnet wandered in and out like a pale shadow in Cummings’ poetic landscape, it would be due to Cummings’s “unmaking and remaking [the sonnet’s] lineation and prosody, calling attention to the visual referent of the set form”79 whose mythical stability rested on the integrity of the line.


The three-body problem of prosody

     I argue that it is exactly this assumed diagrammatic coherence of the line as a poetico-discursive or aesthetic object that E. E. Cummings’ work is revealing to be a very precarious entity. It is indeed a literary commonplace that the relative stability of the line has always been recognized as the interplay of the regularity of schematic meter and the irregularity of linguistic variations. Yet, this “counterpoint” or “tension” between the form of language and the language of form as the classical “double pattern” depended on an axis around a phonolinguistic center that has for the most part ignored the role played by the graphematic dimension of writing as constituting a prosodic force all on its own.80 For Richard Bradford, as well as for Richard Cureton, 81 the primary innovation of free verse relies on the decentering of this axis to enable an opening toward a three-body prosodic problem involving classical meter, (segmental & suprasegmental) phonolinguistics, and graphematics.82

     E. E. Cummings began his literary career early, as evidenced by the composition of pieces dated down to his adolescent years. He might have achieved notoriety with his “meta-sonnets” and other experimental visual pieces, but he has also trained himself in the Petrarchan format (see “On souls robbed of their birth-right’s better part”) or translated Horace’s Odes during his Cambridge Latin years.83 In an early piece (“IF”) where Cummings appears to be testing the flexibility of the pentametric line, we will notice how a somewhat innocuous indentation disturbs the prosody of the p-line, even if framed by the classical ASGrs of Capitalization, Rhyme, Punctuation and syntactic Stops,  and where enjambed lines obey line breaks of syntax groups and prosodic phrasing. This perception, of course, depends on the accuracy of the editorial edition of the piece, if the indented section indeed were there as Cummings intended them to be. As an early piece, it reflected Cummings’ initiation into the notion of the poetic LINE, and the piece “IF” is one of those pieces where its literary hegemony continued to exercise itself.


“IF” by Cummings in print in George J. Firmage’s edition

     After scanning the piece, we would be first struck by how unCummingslike it is, given the supersymmetrical or utopian diagrammatic alignment of the graphematic, prosodic, syntactic, and syllogistic planes. Schematically, the argument format is If a = b & c = d, then x, but y, because z, forming the skeletal frame of the piece in both rhythmic and intellectual directions. In this way, all planes are in a macro-iconic relationship, a multi-parallelism which is reinforced by the lower level equivalences arrayed among the phonic, lineal, stanzaic, phrasal, and logical composition of the elements of the piece. In general terms, the iconic84 and structural (or analogical and grammatical) principles of linguistic assembly coincide and do not enter into any conflict. It would seem, then, that as a student of language and literature, Cummings wanted to satisfy all the demands of the various prosodic and syntactic subsystems by attempting a utopian mapping between the glottographic and nonglottographic planes.


     Each of the three stanzas of “IF” is actually a quatrain if the last two stichs of each stanza were combined in the same line, giving us four pentametric lines per stanza, all laid out with the “ideal” caesura placement after the fourth syllable. This rigid graphic alignment is avoided perhaps to introduce a minor variation in an otherwise textbook grade exercise in poetic composition. Yet, it is a graphic and typographical division whose possible consequences for reading the lines have not yet been a part of Cummings’ understanding of the semiotic potentials of print layout and of visual prosody in general. (This will come later.) Notice how the recognizability of the p-line is redundantly reinforced by overcoding it with ASGrs:

   1) The use of capitalization at the beginning of each line does not just signal the beginning of each graphematic line but also the beginning of the syntactic Sentence and the new prosodic Utterance

   2) The consistent rhyme scheme provides the orthoepic and orthographic buffers which signal the intonational break of the prosodic contour that returns for all the pertinent lines, thus indicating or imposing a notion of their isometrical and harmonic nature

   3) The use of grammatical punctuation or “syngraphemes” (Peter Gallman) at the end of some of his lines indicates Cummings’ literate dependence on Grammar in overmarking the line break instead of allowing justified space to play this role, as if the line as an aesthetic figure might evaporate if not overcoded in this manner

   4) The number of distinctively-rendered lexemes and graphemes in the first two stichs, being mostly eight, conforms to the p-line’s goldilocks zone where the physiological limit of reading and speaking are most coincident, unifying the periodic mean rate of visual scanning with vocal performance, and satisfying the redundant IPH mapping between the syntactic and prosodic domains

     At this point in his career, it will be understandable if Cummings had not yet explicitly reflected on the semiotic role of the printer’s baseline (and of typography) as a visual device in the inscription of the line form as a poetic element of composition, and simply assumes its passive role in the regulation of meaning (cf. Stéphane Mallarmé’s Un coup de dès first published in 1897). The indented grouping of the middle three lines per stanza probably would seem to have no value beyond signalling a minor rhetorical shift in the argumentation and the subordinate role of its lines to the main thread of the argument. The return to the initial alignment for the sixth lines reinforces its status as an ASGr of Conclusion. Its position at the end of each stanza gives it the highest significance, measured against the weight of the main premises of the first two lines and nullifying their truth value. Cummings’ employ of the visual dimension is here limited to the ASGr of paragraph Blocking found in prose. The indentation did not perturb the alignment of the iambic pattern and has, therefore, no greater prosodic function than an expository one. The use of the ASGr Italicization also echoes the common practice of overmarking graphemes visually for emphasis, and we probably cannot credit Cummings here for communicating the multiple layers of the use of such ASGr. Considered in all these values, the ASGr of Italicization is the paradigmatic figure of markedness, able to convey semantic, referential, phonetic, logical, grammatical, pragmatic, metalingual, deictic, and discursive functions by embodying all the glottographic and nonglottographic polyvalency of the sign.

     The stylistic breakthrough in Cumming’s writing came, according to the British Classicist J. Alison Rosenblitt, 85 at the end of his Harvard years:

For Cummings’ readers, the licence which he takes with capitalization is one of the most immediately recognizable features of his poetry, along with his irreverent punctuation, his experimentation with the placement of text on the page, and his unique syntax. How far he pushed his distinctive techniques varies across the different volumes of poetry, but the essential breakthrough came during Cummings’ final (graduate) year at Harvard in 1915–16 and during the months spent just afterwards in New York, in the company of his Harvard friends.

Cummings’ first volume of poetry, Tulips & Chimneys / Tulips and Chimneys, contains work from both before and after this stylistic breakthrough. The change, coming hand in hand with a change in diction and register, is obvious, from lines such as:

Thou aged unreluctant earth who dost
with quivering continual thighs invite
the thrilling rain thy slender paramour
to toy with thy extraordinary lust

to:

your little voice
                       Over the wires came leaping
and i felt suddenly
dizzy


     Rosenblitt traces Cumming’s “desire for novelty” and his “liberation from conventional English capitalization, punctuation, and syntax” not only through his Classical training in Greek and Latin but also through his contact with the “literary context of early modernism” such as Cubism and Imagism.86 Although “liberation” may be too strong a word to characterize Cummings’ stylistic “breakthrough,” we can at least immediately notice how the p-line of “Thou aged unreluctant earth who dost” was abandoned for the rhetorical, typographic, and syntactic innovation of the cited piece starting with “your little voice.” In its use of enjambment, it deploys a key signature technique of free verse poetics.87 Here, the visual rhythm is starting to gain traction, and meter is less invoked as an ordering principle. Even if syntax remains in the background, it does not serve as an ASGr operator of the p-line. He is less dependent on grammatical punctuation, and uses Spacing like indents as an iconic marker of separation, distance, intensity, and movement (“Over,” “leaping,” “suddenly”). Capitalization no longer simply Overmarks the initiation of a line, but visually accentuates a major shift in emotive content. The typographically and visually isolated single word “dizzy” is able to communicate a major prosodic pause without the aid of any metrical, syntactic, or grammatical punctuation. Here, Cummings maintains his affinity for end-line or last line phrasal or logical stress, using a preceding setting to highlight a dawning moment, as if the telos of rhythm was to arrive at a prosodically marked resting point, before launching into another.

     The “destruction of the pentameter” (R. Bradford) would arguably be most evident in one of Cummings’ best known pieces, “l(a,” usually just referred to as the “leaf” poem, published in 1958 under the collection 95 Poems. What would immediately strike any English reader would be its vertical typographical format, the absence of horizontal baselines for reading, and the rarity of lexicographic objects. In short, if we adhere to the habitual left to right (LTR) Directionality ASGr of reading, it would appear that there is not much to “read.” The most concession to the LTR reading motion would be 1) the lexeme or free morpheme “one” standing by itself among incoherent free-standing glyphs; 2) and the bound morpheme of “state” {-ness}. Prosodically, the poem does not marshal metrical, phonological, morphological, or syntactic support as its main semiotic carrier of signification and formal organization. It deflates the LTR directionality which dominates many writing systems in the world, including English, and instead reorients us toward a top-to-bottom (TTB) direction like Japanese or Chinese. Once this new directionality is taken, the reader will be able to recombine the decomposed glyphs to form a renormalized LTR syntax. Such a renormalized syntax and directionality is, however, not the hermeneutic objective of the leaf poem. Instead, they form the semantic and metaphorical “backstory” or fabula of the poem whose graphematic format would be the syuzhet. In Cummings’ experimental work where the graphematic primes are disentangled from being atomic components of the upper hierarchies of the writing subsystems, it could be said that the symbolic and prosodic functions of the latter subsystems are “shelved” into a fabula as a form of metaliterary, metalinguistic, and metaprosodic allusion.


Tabular summary of the asymmetries between the fabula and the syuzhet poles of Cummings’ l(a.
Typographic plane

l(a

le
af
fa

ll

s)
one
l

iness
Syntactic plane
The Sentence here is not the aesthetic object and semantic support. The absence of a Sentence pattern following the LTR reading directionality is enhanced by the lack of grammatical punctuation and majuscules. The typographical layout does not mirror a standard Syntax, now revealed to be a type of ASGr for overcoding glyphs into a visual signature of syntactic strings. The typographic order has only < one > as a familiar readable lexeme. The renormalized a leaf falls / loneliness or loneliness / a leaf falls converts the morphologically disjunctive typographic series into a syntactically hierarchical Sentence on the reading surface, requiring a different mode of aesthetic attention that wipes out the iconic potency of the typographic layout in which loneliness is not just an accidental attribute but an indissociable element of the falling leaf. In the technique where “visual merger renders thematic merger,” 88 the emotionally generic word “loneliness” harbours within its formal structure its concrete analogy, metaphorical image, and objective correlative, “a leaf falling.” This agglutinative effect would be lost if the two parts of the poem were graphically separated as independent words.
Prosodic plane
The layout does not dictate any “instruction” graph for vocal or subvocal performance in IPH unless it employs a conversion via Syntax and Word reconstruction ASGr to manufacture a fully-decked Intonational Phrase. However, this won’t be isochronic with the typographical map which follows a fusional or intercalated TMESIS sequence. Only < one > and possibly < iness > could be mapped for a conjoined phonetic-semantic valence for vocalic performance (and do you say ī-ness, one-ness, or ĭ-ness?). The rest following the syllabic format—C(V, CV, VC, CV, CC, C), VCV, C, VCVC—have no semantically significant phonetic value by themselves (as freed or dissociated graphemes, morphemes, or phonemes in English). Apart from < one >, not any of the CV bundles has a phonetically valid lexemic existence. The grapheme set < iness > can be vocally reproduced, but it is not a regular “word.” Prosodically, however, it has the role of an ASGr of Conclusion and stabilizes the piece by providing a metaphoric ground state framed against the metaphoric vertical motion of the “leaf.” It is the telos of the diagram, the final resting state iconically overcoded by its terminal position, its horizontal reading direction, and its promotion of the morphological suffix to a lemma. By its visual placement at the end of an iconic motion, it is made to carry, subvocally, the whole burden of the Conclusion.
Metrical plane

The syllabic primes, being mere dissociated phonemes, have no intonational or suprasegmental value since they are not part of any segmentally prosodic or lexical string. Their stresses can only be “acquired” in a segmental, lexemic, syntactic, or poetic LINE assembly. Thus, the poem does not invoke the sound-sense parallelism dear to neo-classical versification. In the place of sound or speech iconicity, it invokes a form of visual self-reference in which its smaller constituents form the micro-image of its larger constituent. That is, the iconic numeral < 1 > which the macro-geometry suggests is replicated by the letter-grapheme < l > in minuscule and spelled out as the singular, self-referring, solitary lexeme of the entire poem, < one >.
Lexicographic plane
Following English reading directionality, we activate a tension between lexicosynthesis and lexicoanalysis as we scan along the graphemic pattern. At the start, we are reminded of interjections, or aphasiac speech. The tension, held up by the typographic distance and proximity along the glyphic bundles, invites both treatments, without settling into any. Thus, apart from < one > and < iness >, the tension never resolves into a stable lemma or morphemic set, free or bound.
Graphomorphemic plane
In lexicoanalysis, the lexemes are broken up into constituent glyphs, suspending their letteral and syllabic affiliations, and attaining a level of independence to become perceptible as visual aesthetic artefact. Cummings invites us to work with the lower level elements of the linguistic hierarchy.89 Not one is morphologically intact, except in < one > and < ness > as suffix of adjectival state, reinforcing the semantics of isolation by their relative state of well-formedness. On the graphematic level below the lexical level, none of the glyphic groups have any morphological value. They are simply graphic integers, and they are able to acquire some semantic charge only by being operated on by the TTB, and not LTR, ASGr of Directionality. The iconic/analogical function of language assembly takes precedence over the structural/grammatical. The front-staging of the graphematic downplays or “mutes” the glottographic function so that the nonglottographic function can become aesthetically hyperactive.


     Cummings’ suppression of 1) the p-line as the utopian meeting point of rhetorical and grammatical systems; 2) the set of lexicographic, morphological, and phonological objects as the main bearers of semantic content; and, 3) the role of verbal prosody as the only medium which conveys such semantic content, places the burden of signification on the graphic medium and its decomposed glyphs (part of the lowest and least regarded levels of the linguistic hierarchy in both cultural and philosophical terms). Yet, instead of being an iconoclastic or simply anti-grammatical procedure, it raises the iconic coefficient of the disassembled glyphs by the suspension of ASGrs that overmark them as segmental glottographic or lexical primes. Divested of such segmental affiliations that constitute the elemental base of upper phrasal structures in the writing system, they shake off their phonetic, morphemic, lexical, and syntactic roles. No longer participants of upper signifying bundles that were supporting glottographic and referential subsystems, they recover an ideographic and pictographic charge which replace the semantic, symbolic and metaphorical properties of those bundles. On top of this however, they are placed in a diagrammatic kinetic window where their decomposed status intersect with the ontological rhetoric of material objects, a rhetoric which enables them to be iconic embodiments of themselves or to be literal versions of their own metaphors.

     The central dynamic which supports this transmutation is no longer the glottographic and its authoritative Voice or Ear but the ASGr of Directionality and Orientation, a pure movement of visual reading informed by the iconic modeling in graphematic space of the phenomenon of universal gravitation. Here, the disassembled graphemes are transformed into quasi-objects with a physical dimension capable of being affected by physical forces. The rhetorical result we get is the diagrammatic syntax of things, not the linear syntax of language. The geometric metaphor of the Line gives way to the kinetic metaphor of Directionality.90 By pushing the metrical, syntactic, lexical, morphological, and phonological codes into the background as an inactive fabula, Cummings solicits our aesthetic focus on the “liberated” graphematic components, combines their multilevel self-iconic functions with the TTB directionality and spatial orientation of reading motions, then couples this “fractal” self-reference with the kinetic modeling of gravity to invest them with the aesthetic aura of physicality. In effect, the metaphorical basis of signification is shifted away from the symbolic and phonolinguistic function of a semiotic system toward the ergonomic dynamics of reading motions. The thingness of the glyphs would be an overlay translation of the movement of reading projecting its own aesthetic image across the prosodic surface. By laying out the macro-geometric icon of the piece as a pictorial element of the page, then transecting it with the punctual diagram of the kinetics of falling bodies, Cummings renders the spatio-temporal semiotic axes of (silent) reading motions perceptible.



Notes


1. Martin Neef, “Writing Systems as Modular Objects: Proposals for Theory Design in Grapholinguistics, Open Linguistics 2015; 1: 708–721. DOI 10.1515/opli-2015-0026. Italics added.

   Beatrice Primus gives an explanation of the unavoidable asymmetry in the inter-systemic mapping of linguistic components:

Orthographies have been criticized for mapping spoken language imperfectly. But functional imperfection is a natural trait of language. This means that a mapping between a A and b B, where A and B are two distinct linguistic levels that stand in a functional relationship, is not always a one-to-one correspondence or a complete structural isomorphism. One of the most obvious sources of functional imperfection is that linguistic units and representations serve various functional constraints that may compete with each other. Thus, for example, phonological features do not only correspond to phonetic features, but also have a lexical distinctive function. Furthermore, phonological features are motivated intrinsically by phonological rules. These multiple functions lead to functional imperfection in relation to the phonetic system: not all phonetic features are mapped onto phonological features, and some phonological features are hard to identify phonetically. It comes as no surprise that the phonological function of a writing system is never perfect. The source of functional imperfection lies in the fact that a writing system represents different aspects of spoken language and that the representation of spoken language is not its only function.

(Beatrice Primus “A Featural Analysis of the Modern Roman Alphabet,” in Written Language and Literacy, Vol. 7 (Issue 2): 235–274, 2004. https://doi.org/10.1075/wll.7.2. Emphases mine.)

   Another approach which explains this asymmetry comes from Martin Neef. Neef’s inquiry about the requirements for a theoretical modeling of the abstract nature of writing systems and their relationship to the abstract language system in general and to speech in the phonological and phonetic sense in particular appears to depart from the Saussurean notion that the language system and writing system are independent of each other. He views “the language system [as] a module of the writing system.” In addition, for him, “trying to derive spelling from speech is a pointless venture [because] spoken language consists of continuous sound streams that hardly give a basis for the derivation of spellings with discrete units.” The reason is that the graphemic or graphematic notation is not phonetic but phonological, that is, a relationship between abstract objects.

   Closer to Martin Neef’s concept of the graphematic field as a “solution space” for the representation of unsegmented phonic-sonic stream is the portrayal of Manfred Kohrt of that field as abbreviations of statements of rules, a characterization whose emphasis is less on the abstract entities which supposedly populate the field than on the equivalence operations which happen in it.

As a matter of fact, however, what is represented by such a symbolization are by no means simple units of a special substantial kind; instead, we are confronted with abbreviations of statements of rules that connect items of written language with those of spoken language.

(Manfred Kohrt, “The Term 'Grapheme' in the History and Theory of Linguistics,” in Gerhard Augst, ed., New Trends in Graphemics and Orthography (Berlin and NY: de Gruyter, 1986), p. 92.)

   In this paper, this solution space will be described generally as “phonographic,” that is a mapping operation between phonic and graphic signatures. If grammatology is the study of writing systems, graphemics or graphematics is the study of the scriptural aspect of the abstract system of graphemes notated by the written medium. The phonographic solution space is the abstract mapping between the phonemic and morphemic principles, and as graphemic/graphematic to refer to the material inscription or translation of such an abstract operation on a display medium such as the Page or Screen.

  Furthermore, the phonographic solution space can be seen as analogues of Jerome McGann’s (2007) “environment of thinking,” or of Stephane Mallarmé’s “espacement de la lecture,” or again of Manuel Portela’s (2013) mediatic spaces of reading motions. If we add in Deleuze & Guattari’s continuous variation as its pragmatic condition, such a solution space will always remain open-ended and populated by heterogeneous elements resistant to the metaphorical ambition of analogy and equivalence (or isomorphism, iconicity).

2. Des Ryan, “Linguists’ Descriptions of the English Writing System,” in The Routledge Handbook of the English Writing System, eds. Vivian Cook and Des Ryan (NY: Routledge, 2016), pp. 41 ff.

3.  Malcolm D. Hyman, “Of Glyphs and Glottography,” Language & Communication, Volume 26, Issues 3–4, July–October 2006, https://doi.org/10.1016/j.langcom.2006.03.001, pp. 244, 236. Italics added.

   The distinction between reading and verbalizing seems to apply as operations equally to both functions. Hyman, of course, possibly meant that a certain work of paraphrasing is involved when nonglottographic signs are “verbalized,” including some amount of translating. In verbalizing a nonglottographic mark, we again associate it with pronounceable words, something done with glottographic marks as a regular process of reading. Numerals and mathematical equations were given as examples, but some more graphic types like road signs were included. This certainly opens both processes of reading and verbalizing to all types of signs, especially when “primarily” glottographically-oriented graphemes are functionally given diagrammatically iconic functions like in Concrete poetry or E. E. Cummings (as in the famous falling leaf piece).

   This observation reconfirms the ambiguous encoding of linguistic signs: letters, words, syntagms, and texts blocks are animated by both glottographic (phonographic) and nonglottographic (logographic) principles. As Martin Neef (2015) points out, since “the Roman script is not inherently alphabetic but only in the context of a specific writing system..., the distinction between alphabets, adjads, and syllabaries is located on the level of writing systems and not on the level of scripts.” Put differently, it is only through the glottographic or nonglottographic operation that a script becomes phonographic or logographic, and that this process involves a metaprosodic mapping of visual and verbal signatures. This implies that the distinction between what is prosodic or not is an operation of self-differentiation traceable to the (meta-) prosodic operation itself.

   In the first passage quoted, Hyman appears to restrict the nonglottographic signs to nonlinguistic “content,” something which appears to place it outside of writing in a Saussurean way. However, it is made clear that, since the nonglottographic is not linked to vocal performance, its content is already psychological: “its notations are themselves signs of psychological states.” In short, nonglottographic signs function ideographically (since mental states are also ideas) in the same way spoken signals do in Saussure.

   The difference between the phonographic and logographic principles is illustrated by American linguist Alice Faber:

Orthographies differ in whether the bulk of the symbols in the system, taken in isolation, are susceptible of semantic interpretation. The English symbol < d > has no inherent meaning, while the symbol <  +  > does. Systems in which symbols like the former predominate are phonographic while systems in which symbols like the latter predominate are logographic. Logographic systems code for the most part morphemes, as Sampson notes, and phonographic systems code units of sound.

(Alice Faber, “Phonemic Segmentation as Epiphenomenon: Evidence from the History of Alphabetic Writing,” in Pamela Downing, Susan D. Lima, and Michael Noonan, eds., The Linguistics of Literacy (Amsterdam: John Benjamins, 1992), p. 118)

4. Hyman, op. cit., p. 241.

5. Hyman, on the hybrid nature of the English writing system:

Few would hesitate to classify our writing as glottographic. Yet this writing contains various elements whose relation to the spoken language is quite unclear. Foremost among these are the Indo-Arabic digits, which are often identified as ideograms.... These, of course, correspond to different words in different spoken languages.... More importantly, even in English the same numeric notation may be verbalized in various ways (e.g., 1941: ‘nineteen- forty-one’, ‘nineteen-hundred-and-forty-one’, ‘one-thousand-nine-hundred-forty- one’). Nor is it clear when we manipulate numbers arithmetically that we need to verbalize them at all. As for our system of punctuation: sometimes it has phonetic correlates (such as a pause or particular intonational contour), sometimes it serves for delimiting syntactic boundaries (without any evident phonetic correlates), and sometimes it is generated by rules of the written language grammar that do not correspond to anything in spoken English....

Yet the alphabet is used for a variety of ancillary functions entirely separate from the coding of spoken language. Alphabetic characters are used to order, classify, and rank.

(Hyman, pp. 241-242. Italics mine)

   The writing systems scholar Peter T. Daniels has the following comments on English orthography:

The much maligned orthography of English itself is usually recognized by linguists to be a pretty good system, not in need of reform, precisely because it is not phonemic, but morphophonemic. Actually, though, for the basic vocabulary, the several hundred or so of the most common words, it is best to regard English spelling as primarily logographic: regardless of how many spelling-to-sound regular correspondences there are, and how much information about derivational morphology and etymology is packed into the spelling of a word, the spellings must be memorized. The pronunciations of cove, love, move cannot be predicted, no matter how justified the different realizations of -ove are by either history or underlying form.

(Peter Daniels, “The Syllabic Origin of Writing and the Segmental Origin of the Alphabet,” in Pamela Downing, Susan D. Lima, and Michael Noonan, eds., The Linguistics of Literacy (Amsterdam: John Benjamins, 1992), p. 100. Italics added.)

6. Hyman, p. 237. Italics added

7. The phrase comes from Gilles Deleuze and Felix Guattari’s “pragmatics,” which is a principle of general chromaticism whereby the variations are themselves the system: “It is the variation itself that is systematic, in the sense in which musicians say that "the theme is the variation."”

This is what we are getting at: a generalized chromaticism. Placing elements of any nature in continuous variation is an operation that will perhaps give rise to new distinctions, but takes none as final and has none in advance. On the contrary, this operation in principle bears on the voice, speech, language, and music simultaneously. There is no reason to make prior, principled distinctions. Linguistics in general is still in a kind of major mode, still has a sort of diatonic scale and a strange taste for dominants, constants, and universals. All languages, in the meantime, are in immanent continuous variation: neither synchrony nor diachrony, but asynchrony, chromaticism as a variable and continuous state of language.

   The idea of a continuous variation means that any so-called linguistic units are in perpetual state of variation as if they resided in a virtual plane of time without being unreal.

Not only are there as many statements as there are effectuations, but all of the statements are present in the effectuation of one among them, so that the line of variation is virtual, in other words, real without being actual, and consequently continuous regardless of the leaps the statement makes. To place the statement in continuous variation is to send it through all the prosodic, semantic, syntactical, and phonological variables that can affect it in the shortest moment of time (the smallest interval).... This is the standpoint of pragmatics, but a pragmatics internal to language, immanent, including variations of linguistic elements of all kinds.

(Gilles Deleuze & Felix Guattari, op. cit., pp. 93, 97, 94. Emphases added.)

   The citation from Richard Cureton is from his article, “Visual form in E.E. Cummings' No Thanks,” July 1986, Word & Image 2(3):245-277, DOI: 10.1080/02666286.1986.10435349. I will return to Cureton’s excellent scholarship on E. E. Cummings in the later parts of this essay.

8. Hyman, pp. 245-246. Italics added.

   The reading directionality which characterizes writing systems is clearly a simultaneous cueing for both oral and visual processing. How such visual prosody dovetails with what Jerome McGann calls “bibliographic codes” and what Johanna Drucker terms “diagrammatic writing” will be taken up in the next parts.

9. The American linguists Wallace Chafe and Deborah Tannen have noted “the inextricability of speaking and writing even in those modes of discourse that seem most exclusively a matter of writing and reading, and the inherently social nature of all discourse.”

(Wallace Chafe and Deborah Tannen, “The Relation between Written and Spoken Language,” Annual Review of Anthropology, Vol. 16 (1987), pp. 383-407, http://www.jstor.org/stable/2155877, pp. 396, 398.)

   As Geoff Hall, quoting Chafe and Tannen, has argued in his promotion of the “discourse”” as a broader category subsuming the domains of speech and writing, “there is no single feature or dimension that distinguishes all of speech from all of writing,” and “Distinctions between orality and literacy on the one hand, and spoken versus written language on the other, do not suffice to characterize real discourse . . . the relationship of literary to conversational language [is] . . . closer, and distinctions between them foggier, than had previously been thought.”

(In Geoff Hall, Literature in Language Education (NY: Palgrave MacMillan, 2005), p. 66.)

   The above remarks on the blurring of the distinction between speech and writing on the discursive and generic levels should not be taken to mean that all graphemic features (must) have speech correlates, only that the division is not absolutely clear-cut in real practice, and that the ambiguous encoding of the sign is reiterated on various scales of the linguistic hierarchy. Since even glottographic forms could be assigned nonglottographic functions and vice versa (as Hyman had noted with mathematical formulae), the contexts in which this happens are metadiscursive and involve bidirectional scriptly-speechly transcription and translation in instances of discourse. That is, the difference resides not on the level of essential identity but on stylistic overmarking.

10. Once again, this result indicated that the phonology of a word—here, its stress pattern, unconfounded with any other lexical factors—is included in a reader’s representation of a word. This supports the claim that a veridical representation of a word is required before the eyes move on in a text. Together, the results reviewed in this section support several conclusions. Phonology is deeply involved in silent reading. While it is not likely to require explicit subvocalization, the phonological representation is not impoverished. It includes not only information about syllabification of individual words but also information about their metrical structure.... Phonology may have to be viewed as a critical part of the veridical representation of a word that a reader requires before moving ahead in a text....

As part of a broader analysis of how prosody affects language comprehension, [Janet Dean] Fodor advanced the implicit prosody hypothesis: Implicit Prosody Hypothesis (IPH): In silent reading, a default prosodic contour is projected onto the stimulus, and it may influence syntactic ambiguity resolution. Other things being equal, the parser favors the syntactic analysis associated with the most natural (default) prosodic contour of the construction.

(Charles Clifton, Jr., “The Roles of Phonology in Silent Reading: A Selective Review,” in Lyn Frazier & Edward Gibson, eds. Explicit and Implicit Prosody in Sentence Processing: Studies in Honor of Janet Dean Fodor (Springer International, 2015), pp. 161-176. Italics added.)

11. The landmark scholarship for the origins of silent reading in the 7th century is by Paul Saenger, Space between Words: The Origins of Silent Reading (Stanford: Stanford University Press, 1997).

   Shafquat Towheed and W. R. Owens, in the introduction to the three volumes on the history of reading they edited, write:

In the world of manuscript production before the printing press, and in the centuries before mass literacy, the relationship between reading and listening, between the oral and the aural, and between text and the spoken word was much more evident than it is today. Some of the most important scholarship in the history of reading in the last few decades has brought to the fore this particularly complex relationship between orality and text in the era before the widespread availability of print. Paul Saenger has demonstrated the close relationship between the rise of monastic silent reading and the development of spaces between words in medieval scribal manuscript production, while Armando Petrucci has argued that the shift from a largely monastic to an increasingly secular (and often solitary) engagement with books gave rise to new intensive humanist reading practices.

(Shafquat Towheed and W. R. Owens, eds., The History of Reading, Volume 1 International Perspectives, c.1500–1990 (Hampshire: Palgrave Macmillan, 2011), pp.3-4. Italics added.)

12. I cite fully the passages where she argues for “the absolute nature of the contrast” between ‘the reader as a speaker’ and ‘the reader as a hearer’ models:

At this point, it is important to distinguish the contrast between ‘the reader as a speaker’ and ‘the reader as a hearer’ from more familiar binaries such as ‘oral versus literate’ and ‘spoken versus written.’ The terms ‘oral’ and ‘literate’ have been widely deployed in the fields of literature and anthropology, particularly since Walter Ong’s influential Orality and Literacy (1982). Ong’s work implies (though does not make explicit) a scale running from oral to literate, on which a work of prose might be situated, for example, more towards the ‘oral’ end or more towards the ‘literate’ end (Ong himself talks of ‘residual orality’ and ‘internalized literacy’). In this model, a written text can be described as ‘oral’ if it exhibits features more commonly associated with communication in oral societies (see, for example, Ong’s work on Tudor prose, 1965). More recently, scholars like Fox (2000) have questioned the starkness of the oral-literate contrast, by describing societies in which the oral and the literate not only coexist but interact in complex ways. The terms ‘spoken’ and ‘written’ are more narrowly associated with the field of linguistics. Here too, earlier research emphasized the differences between speech and writing, whereas more recent work illustrates the degree of variation within speech and writing, which complicates the relationship between the two (see, for example, Biber, 1988). Both sets of binaries, then – ‘oral versus literate’ and ‘spoken versus written’ – are founded in an opposition which has been softened by subsequent research.

The contrast between ‘the reader as a speaker’ and ‘the reader as a hearer’ is not like these other contrasts, in part because it cannot be expressed on a scale. A model reader can either be speaking the text or (inwardly) hearing it. But it is not possible to carry out the roles of speaker and hearer simultaneously.... The contrast between two models of reader is therefore rather more precise than the contrasts between oral and literate or speech and writing. The transition from one model to the other might be prolonged and take place in a society of mixed and interacting degrees of literacy. But the model of the reader as a hearer can come into being only at a point when there is a critical mass of very accomplished silent readers – enough to support the publication of texts based on this model. Its appearance in history, then, can potentially be identified more precisely than the beginnings of the broader ‘literate society.’

(Elspeth Jajdelska, Silent Reading and the Birth of the Narrator (Toronto:University of Toronto Press, 2007), pp. 8, 6-7)

13. Jajdelska, op. cit., p. 166. Italics mine.

14. Jajdelska, p. 8. Italics added. Reading as hearing… is ubiquitous in ordinary speech. Won’t this blur the solid distinctions she had laid down, if ordinary speech is also informed by silent reading motions?

   The term “visual prosody” will be discussed in the later sections. See note # 55 below.

15. See Deborah Tannen, for example, on the idea of a “poetics of conversation” in her article entitled “Ordinary Conversation and Literary Discourse: Coherence and the Poetics of Repetition” (in The Uses of Linguistics, ed. Edward Bendix, Annals of the New York Academy of Sciences vol. 583, 1990). Her work with Wallace Chafe has systematically eroded the boundaries between writing and speech, orality and literality, ordinary and special language.

   I will not go to mention more instances where absolute divides between an oral and written register are difficult to maintain, a division upon which the distinction reading as hearing and reading as speaking could be neatly ascertained, except as generic, stylistic, pragmatic or functional overmarking: George Philip Krapp’s “eye dialect,” online type chat, Eleanor Berry’s notion of “ritual speech” in 20th century American poetry, etc. If they will be alluded to, it will be in the stream of the discussion of E. E. Cummings’ metaprosodic negotiations.

16. See Geoff Hall (2005).

17. The quest for a prosodic font can be traced all the way back to the typographic experiments of the Italian Futurists and the avant-garde. Today, when display technology is more advanced, the multimodal realization of visual marks has become a possibility. An example of current work on this line is Tara Rosenberger Shankar’s Prosodic Font: the Space between the Spoken and the Written (1998). As she sums it up in an article,

The proliferation of speech recognition as input to Computer Mediated Communication (CMC) systems opens up new possibilities for the design of typographic forms. Designers can use the musical expressiveness of the speaking voice to shape letterforms in real time. Letters formed by speech are more representative of the emotional and contextualized person speaking than are fonts now. Prosodic Font is an object-oriented font that assumes a dynamic, temporal form. It emulates the tonal and rhythmic motion in the speaking voice. Preliminary user testing results show that people are able to identify Prosodic Fonts as representative of particular prosodic variations.

(Tara Rosenberger Shankar & Ronald L. Macneil, “Prosodic Font: Translating Speech into Graphics”, February 2001, DOI: 10.1145/632716.632872. Italics added.)

   Two comments must be put forward here. First, the psychologistic grounding of such prosodic font concept presumes a universalist and transparent notational technology which translates the speech signal into unambiguous psycho-emotive values; second, it elides the fact that every notation is already prosodic even if the technology involved before the New Media had been fundamentally analog (with “the tonal and rhythmic motion in the speaking voice” characterizing what we understand as the posited “referent” of poetic prosody). Martin Neef’s comment on determining the writing system from purely phonetic grounds can be extended to the design of prosodic fonts:

If such a speech-based approach is considered seriously in grapholinguistics, the input of the derivation is not taken to be the result of some phonetic measuring but a (broad) phonetic transcription. Such a transcription, however, is not a representation of spoken language in an empirical sense of phonetics but, as argued above, merely a different kind of writing system.

(Neef, op. cit., my italics)

18. Caroline Féry, Intonation and Prosodic Structure (New York: Cambridge University Press, 2017), p. 306-307. Italics added.

19. Féry, op. cit., p. 307. Emphasis mine.

20. Robert Ladd, Simultaneous Structure in Phonology (Oxford: Oxford University Press, 2014), p. 74.

   The British linguist Anthony Fox says pretty much the same thing:

The description and the definition of these features have always been something of a problem for linguists; for many years, and especially in the formative period of modern linguistic theory in the second quarter of the twentieth century, the study of these features suffered from relative neglect. With some exceptions, phonological descriptions were based primarily on 'segments'—vowels and consonants—and prosodic features were either ignored or forced into an inappropriate segmental mould. In recent years this imbalance has been redressed, and several phonological theories are now available which are not merely more sympathetic to prosodic features but are even largely based upon them. However, it remains the case that there is no universal consensus among phonologists about either the nature of prosodic features themselves or the general framework for their description, and it is difficult to obtain a clear picture of the field as a whole.

(Anthony Fox, Prosodic Features and Prosodic Structure: The Phonology of Suprasegmentals (Oxford University Press, 2000), pp. 1-3. See the discussion of the problem with prosody’s distinction from paralinguistic features in pp. 9-10.)

21. Féry, p. 6. Italics mine.

   It might be useful to note down below her working definition of “intonation” and “prosodic structure” from the same page for later discussion. The terms, however, may not be as “straightforward” as it appears for many other linguists.

(1) Intonation

Intonation is the tonal structure of speech expressed by the melody produced by our larynx. It has a phonetic aspect, the fundamental frequency (F0), and a grammatical (phonological) aspect. As for the term ‘prosodic structure’ (a definition follows), it will chiefly refer to the phonological domains arising from syntactic structure, information structure and pragmatic roles.

(2) Prosodic Structure

Prosodic structure refers to the parsing of continuous speech in hierarchically organized [...] prosodic domains – mora, syllable, foot, prosodic phrase and intonation phrase.

   The relationship of these (somewhat abstract and still controversial) prosodic domains with the (similarly abstract and controversial) figures and categories of syntax remains a topic investigated in linguistics. This relationship is also, of course, reflected in the poetic question of the rapport between, or the hesitation between, sound and sense, as Paul Valéry is quoted to say. Or, is it the hesitation between the sound of sense and the sense of sound as mediated via the notational technology of scripts and writing systems, even if inexhaustively?

22. Robert Ladd, ibid. Italics added.

23. Ladd, pp. 63-65. Emphases mine.

   The definition given below from the 2010 volume entitled Experimental and Theoretical Advances in Prosody: A Special Issue of Language and Cognitive Processes edited Duane G. Watson, Michael Wagner, Edward Gibson reflects such an outlook of prosody (my emphasis):

Prosody is the rhythm, stress and intonation of speech, which encodes information that is not encoded by the syntax or words of an utterance. Prosody is critical for parsing speech, constructing syntactic structure, and building a representation of the conversational discourse model, among other linguistic functions.

   This seems rather an absolute characterization of the divide between the prosodic and nonprosodic, but it illustrates the traditional dualism involved in the conceptualization of what is properly syntactic and what is not without accounting for pragmatic factors beyond the structural or formal ones. This duality is what Deleuze & Guattari have summed up as the two “treatments” of language in the history of linguistics:

There are not, therefore, two kinds of languages but two possible treatments of the same language. Either the variables are treated in such a way as to extract from them constants and constant relations or in such a way as to place them in continuous variation. We were wrong to give the impression at times that constants existed alongside variables, linguistic constants alongside variables of enunciation: that was only for convenience of presentation. For it is obvious that the constants are drawn from the variables themselves; universals in linguistics have no more existence in themselves than they do in economics and are always concluded from a universalization or a rendering-uniform involving variables. Constant is not opposed to variable; it is a treatment of the variable opposed to the other kind of treatment, or continuous variation. So-called obligatory rules correspond to the first kind of treatment, whereas optional rules concern the construction of a continuum of variation. Moreover, there are a certain number of categories or distinctions that cannot be invoked, that are inapplicable and useless as a basis for objections because they presuppose the first treatment and are entirely subordinated to the quest for constants: for example, language as opposed to speech; synchrony as opposed to diachrony; competence as opposed to performance; distinctive features as opposed to nondistinctive (or secondarily distinctive) features. For nondistinctive features, whether prosodic, stylistic, or pragmatic, are not only omnipresent variables, in contrast to the presence or absence of a constant; they are not only superlinear and "suprasegmental" elements, in contrast to linear segmental elements; their very characteristics give them the power to place all the elements of language in a state of continuous variation—for example, the impact of tone on phonemes, accent on morphemes, or intonation on syntax. These are not secondary features but another treatment of language that no longer operates according to the preceding categories.

(In A Thousand Plateaus: Capitalism and Schizophrenia, tr. Brian Massumi (MN: The University of Minnesota Press, 1987), p. 103-4. Italics added.)

24.  Apart from the concept of noise (vs. signal) in information theory, this prosodic continuum could also be compared with Jacques Lacan’s notion of the Not-all or lalangue, the agrammatical domain which exceeds the grammatical, thereby making the latter the formal superstructure of the former. The distance between the grammatical and the agrammatical is that gap occupied by the anthropic. See Ann Banfield’s “Introduction: What do Linguists Want?” in Jean-Claude Milner, For the Love of Language, tr. Ann Banfield (NY: The MacMillan Press, 1990).

   Many remarks cited above already point out the abstraction which carves discrete units (the “sign” etymologically) out of the stream of speech sound in general, and this is just one physical channel which registers the abduction of salience from experience.

     Robert Ladd (2014, p. 21):

As noted by Kornai (2008: 234), even in a theory like AP [autosegmental phonology] that posits distinct speech gestures as primitive elements, ‘the gap between the discrete and the continuous is left unbridged....Whatever we may do, we still need to recover the discrete articulatory configurations from a continuum of signals’.

     In note # 3 at the end of her article, Alice Faber (1992, p. 128) writes:

The segments identifiable in an acoustic waveform, to the extent that their boundaries are well-defined, are not isomorphic with the segments of phonological analysis or of alphabetic writing.

(Alice Faber, op. cit., pp. 112, 114. Italics added.)

   On the notion of source vs. filter, see Fox (op. cit., pp. 4-6, some passages below), concerning the attempt to locate prosody physically in terms of articulatory phonetics. What we see, at the least, is the more fundamental role of prosodic articulatory faculties in the production of segmented sounds, even if the duality prosodic/nonprosodic is once more redeployed and given a physicalist grounding as if the recognizability of such segments and non-segments did not require a schema of perception and recall already in place.

We thus have a phonetic basis for distinguishing prosodic from non-prosodic features of speech in the component of the speech process where the features can be localized. The one exception is the larynx, which appears to have both a segmental and a prosodic role....

There is a sense, therefore, in which the laryngeal and subglottal components are more basic than the supralaryngeal, and the subglottal component is more basic than the laryngeal. Since the prosodic features are associated with the latter two, and segmental features primarily with the supralaryngeal component, we could consider prosodic features to be similarly more fundamental, in the sense that segmental features involve the modification of an air-stream which is already specified for prosodic features.

25. Slavoj Žižek (cited with emphases), in his For They Know Not What They Do: Enjoyment as a Political Factor (2008), writes:

This gap is not simply external to language, it is not a relationship between language and a subject external to it; rather, it is inscribed into the very heart of language in the guise of its irreducible (self-)reflexivity. When Lacan repeats that "there is no meta-language", this claim does not imply the impossibility of a reflexive distance towards some first-level language; on the contrary, "there is no meta-language" means, in fact, that there is no language - no seamless language whose enunciated is not broken by the reflexive inscription of the position of enunciation. Here, once again, we encounter the paradox of the non-All: there is no (meta-language) exception to language, it is not possible to talk about it from an external position, precisely because language is "not all", because its limit is inscribed into it in the guise of ruptures in which the process of enunciation intervenes in the enunciated.

    Thus, a metalanguage is not a superior point above language: it is the reflexive motion by which languageness is manufactured as a referentiable object. In the impossible status of having a meta-language, and where language is in effect its own metalanguage by which it recognizes itself as language, the motion of intralingual or interlingual translation becomes the modus operandi by which language effectuates that infrathin difference or gap which edges it out of non-readability or obscurity and toward the logic of doubles and reflection:

The questions we have asked so far about language - Can a language be its own metalanguage? Does any language require, on the contrary, a distinct and possibly richer metalanguage? What about a hierarchy between language and metalanguage? Is it not better to assume that there is a kind of symmetry between them and to consider the question in terms of "translation" rather than "metalanguage"? - also apply, mutatis mutandis, to any discourse that has the ambition to speak of another discourse, to any discipline that claims to analyze another discipline. Epistemology is such a discourse since it strives to make a critical analysis of scientific knowledge. Perhaps even the whole of philosophy, but also semiology, have to ask themselves these questions, inasmuch as they readily adopt a "meta" position with regard to other disciplines.

(In Laurence Bouquiaux, François Dubuisson et Bruno Leclercq, “Modèles épistémologiques pour le métalangage”, Signata [en ligne], 4 | 2013, mis en ligne le 30 septembre 2016, consulté le 20 avril 2019. URL : http://journals.openedition.org/signata/544 ; DOI : 10.4000/signata.544, pp. 15-52.)

   Since interlinguistic or intralinguistic translations require the self-reflexive distance which functions as a metalingual perch by which language recognizes itself, this “interior disjunction” becomes the basis of the ineradicable difference which forces the work of mapping equivalences to begin to operate: translation, metaphor, analogy, correspondence, figures, grammar, and so on. In this doubled state where language is its own mirror, one part gets assigned the role of substance, another that of form; one becomes the essence, the other becomes the accident; one belongs to nature, the other to art; one takes up the body, while the other the shadowy image; one absorbs the nominal, the other complements with the adjectival; one performs the description, the other receives the description; one is assigned the surface, the other depth; one part is referred to as the signifier, another its signified. These are, however, porous border designations, and the roles are reversible, unless a third term intervenes to “resolve” or “dissolve” the dialectical antinomy, if only perhaps, once again, temporarily, via another doubling. This is the mirror stage of language, the arrival of metaphor as such, not as a singular event, but as a recursively reflexive process.

   The contradictory nature of metalinguistic spaces has been interrogated in many places, chief of which is Richard Rorty’s The Linguistic Turn: Essays in Philosophical Method (1967, 1992). Ever since Ludwig Wittgenstein, the notion of a metalanguage from where language could be surveilled as if from an absolute exterior has received critical refutation. For example:

As Gilbert Hottois explains, there is no meta-game or meta-calculus, or again: "There is no place outside of language (of “language games”) from where one could survey them. "There is no “Game” in general, no rules that govern all the particular cases. This phantasm of a common denominator is vain: language games are irreducible to one another. This hypothetical “metalanguage” would necessarily be “inside” language (since it cannot, as we have said, be outside); or, the part cannot grasp the whole.

(Laurence Bouquiaux, ibid.)

26. Faber, p. 127. Emphasis added. Faber writes:

My goal is, in particular, to trace the development of alphabetic writing, since it is the existence of alphabetic writing that is used (implicitly or explicitly) as evidence for the universality of segments as a building block of language. My discussion is informed by the view expressed by O'Connor... that the structure of an orthography for a particular language reflects, albeit not always systematically, native speaker analysis of that language.

   It will be necessary, of course, to delineate more accurately such Sapir-Whorf type of linguistic relativism from cognitive or epistemological or even discursive determinations or factors. The least that can perhaps be said is that a writing system as an abstract solution space for notational signification, in its role as the fundamental information storage and retrieval system for human beings, is not free from encoding biases. This doesn’t mean they cannot be balanced or complemented by other means or media.

27. Anthony Fox, op. cit., pp. 2-3. Italics added.

28. As already noted, most twentieth century approaches to phonological analysis rely on the idealization that the speech stream is divided into discrete segments, each representing an acoustic or articulatory steady state. Nevertheless, as is well known, there is no physical basis for this segmentation. As a consequence, there is a long and often under-appreciated tradition of alternative approaches to phenomena not easily amenable to segmental analysis, most notably of phenomena not easily localized within a word....

To summarize, there is little unambiguous structural support for positing segments as linguistic units. In particular, there is no need to appeal to segments in modeling ordinary language structure and use. Neither is there evidence that incorporation of segments into linguistic models leads to more satisfying or parsimonious analyses. In contrast, structural linguistic evidence suggests that models based on syllables and syllable components (either onset-rhyme or onset-nucleus-coda) might be more appropriate. However, there are spheres, notably reading, in which language users clearly do display an ability to recognize and manipulate segments.

(Alice Faber, pp. 128, 112, 114. Italics added.)

29. Neef, op. cit.

30. Gerhard Äugst, “Descriptively and Explanatorily Adequate Models of Orthography,” in Gerhard Augst, ed., New Trends in Graphemics and Orthography (Berlin: Walter de Gruyter, 1986), pp. 25-42. Emphasis added.

   See his diagram modeling of the rapport between the phonemic and the graphemic schema from page 37 below. The question that comes to mind, however, is how much his modeling depends on the opposition between types from the schema side and the tokens on the graph and phone side. We need to examine how the binary of constant vs. variation subtends the simple division between the schema and the realization parts, especially when the empirical location of the schema (operating from an ideal plane) is hard to determine beyond the language modeling done within language modeling itself.



Gerhard Äugst’s orthographic modeling with phoneme-grapheme correspondence rules (PGC)

   The fact that there are only empirical variants upon which to base any paradigmatic type seems to have been overlooked in such a modeling of the solution space which simply posits the existence of ideal targets in the schema domain (the phonemic and graphemic paradigmatic figura) distinguishable a priori from their realization in the performance domain (as phone and graph instants). Won’t the ontology of formal entities or units be resolved more locally in a radically empirical or temporary way than in terms of universals, invariants, or constants?

It is now well established that segmental realization is intricately conditioned by the prosodic structure of a given utterance, so that, for example, the actual phonetic form of a phoneme is determined by the prosodic position in which it occurs (e.g., Cho &Keating,2009; Cho, 2011; Cho, Lee,& Kim, 2014; Fletcher,2010; Fougeron & Keating,1997).

(In Holger Mitterer, Taehong Cho, Sahyang Kim, “How Does Prosody Influence Speech Categorization?” Journal of Phonetics 54:68-79. DOI: 10.1016/j.wocn.2015.09.002.)

31. Richard Cureton, op. cit., p. 245.

32. Des Ryan, op. cit., p. 64.

33. For a study of avant-garde experimental typography and its role in the semiotics of the signifier as a visual artefact, see Johanna Drucker’s The Visible Word: Experimental Typography and Modern Art, 1909-1923. However, contrary to her conclusion that experiments in types have declined after WW2, it would seem that, with the arrival of New Media, it just got more technologically diverse and heterodox, from prosodic and kinetic fonts to asemic and abstract glyphs globally running parallel to the mainstream type design industry. We need to ask, furthermore, if the previous semiotic analytical frames founded on the various versions of the dualistic coding of the sign still apply.

34. The modern analytical distinction between the “medium” as separate from the “message” is part of a larger set of parallel oppositions such as material/logical, pragmatic/poetic, and linguistic/paralinguistic where the aesthetic object (the perceptual target or message) is assigned to be expressed by the supporting non-aesthetic substance (the medium). The foundation of such dichotomies where the doctrine of the autonomy of the poetic from the pragmatic is derived is the historical result of the modernist dissociation of techné from aísthēsis:

Modern descriptions of art constantly mix two great conceptual legacies. The legacy of the Ancients is interested in the process of making any object or work; the aesthetics of the Moderns is interested in the sensations that the object produces for the beholder. The two perspectives do not precisely coincide. The “art”” of the Ancients includes every kind of making, and thus what we would call “technique” or “technology.” The Moderns’ aesthetics include every kind of admirable beauty, and thus the beauty of natural phenomena (the sublimity of volcanoes). When studying the art-technique of the Ancients, we have to abandon as false oppositions antinomies that are legitimate from the point of view of Moderns. Art did not have the beautiful as its exclusive domain, and technique was not limited to the useful. Art was not the realm of mysterious things and “artistic” vagueness, as opposed to technique as the realm of serious things, rigorous procedures, and guaranteed results. Clarifying the vocabulary was as important as relativizing, as a dictatorship or caricature, any scientific view of rationality modelled on industry and, later on, techno-science.

(Francis Goyet, “Art of the Ancients, Art of the Moderns: The Rules of Art,” in Barbara Cassin, ed., and tr. by Steven Rendall et al, Dictionary of Untranslatables: A Philosophical Lexicon (NJ: Princeton University Press, 2014), p. 43. Emphasis added.)

35. The term “geometric-iconic” combines the appellation mentioned in Richard Cureton’s essay (for the pictorial aspects) with the general notion of iconicity in semiotics.

As many critics have felt it necessary to stress, these poems seem to be rare in the Cummings' corpus. Cummings produced no volume comparable to Hollander's Types of Shape or May Swenson's lconographs, and he wrote few poems whose global visual form is as overtly iconic as Apollinaire's Calligrammes.

(Cureton, op. cit., pp. 249, 266.)

36. Richard Cureton prefaces his essay (p. 245) with an enumeration of questions generally focused on the aesthetic effects and role of visual prosody vis-à-vis other types of prosody:

In spite of these developments, however, we still know very little about the possibilities (and limitations) of visual prosodies. Discussions of visual form in poetry usually provide more polemic than analysis, more apology than illumination. Many of the major questions we might ask about visual form remain unasked, and many of the questions that we have asked remain unresolved: do visual prosodies create some of the same kinds of effects as other prosodies? Are there some types of effects that can only be achieved by visual form? Are the range and intensity of effects achievable with a visual prosody as great as with other prosodies? Is there a limit to the productivity of combining visual and other prosodies in the same text? Does the appreciation of a visual prosody demand that one entertain a certain type of aesthetic- an aesthetic, say, that is different from the usual aesthetic entertained by readers of non-visual prosodies? Are visual prosodies peculiarly appropriate to our contemporary world? These are very large questions, nothing to be answered in a day (certainly nothing to be answered here!). Nonetheless, if we are to become intelligent readers ('viewers') of visual poems, these are questions we all must ultimately confront.

   Cureton does not give any detail about the origin of the term “visual prosody.” He seems to assume some degree of currency for the term to skip the need to give us a basic origin history. In the same year Cureton’s essay on Cummings was published (1986), the renowned poetry scholar Marjorie Perloff was already using the term to characterize an important aspect of modernist and avant-garde poetry, from Futurism to Imagism:

Given this matrix, the Futurists’ foregrounding of the isolated word or even sound on the one hand (as in Khlebnikov’s “Zaklyatie smekhom”) and the insertion of “ordinary” prose, in the form of letters, charts, archives, and so forth, on the other, can be seen as a way of calling attention to the materiality of the signifier so as to reduce the transparency of language. The same is true for the incorporation of visual devices—the abandonment of the left margin, the typographical play, the use of ideograms—into the text. To visualize the poem is to insist that language does not simply point outside itself to some metaphysical reality but that it oscillates between representational reference and compositional game. Again, visual prosody calls into question the centrality of the foot, the line, the stanza, even the whole poem, and substitutes for the framed poetic text the basic unit of the poetic page.

(In Marjorie Perloff, The Futurist Moment (Chicago: The University of Chicago Press, 1986), p. 191. Italics mine.)

   We would want to keep that last sentence in mind for later discussions.

37. Cureton, pp. 245-246.

38. Christoph Küper, ed., Current Trends in Metrical Analysis (Frankfurt am Main: Peter Lang, 2011), pp. 3-4. Bold letters are in the original.

39. Kristin Hanson,Generative Metrics: The State of the Art,” in Christoph Küper, op. cit., pp. 50-1. Italics added.

40. Marina Tarlinskaja, Shakespeare and the Versification of English Drama, 1561–1642 (VT: Ashgate Publishing Company, 2014), p. 4. Highlights are in the original.

   That meter and rhythm are opposed as abstract is to concrete or as form is to realization can be seen from the following:

The opposition between meter and rhythm in contemporary theory of verse derives from Russian authors, first from the Symbolists, then from the Formalists. Meter is understood as an ideal structure which is realized in rhythm. In the structuralist interpretation, the opposition of meter and rhythm was treated as a special case in the Saussurean contrast of language (langue) and speech (parole) or in Jakobsonian terms of code and message. What constitutes a problem is how this distinction is related to the actual creation or perception of poetry. In the spirit of Saussure, the dominating understanding is as follows: meter (language) is primary and general, both the author and the audience must know it whereas rhythm (speech) is secondary and individual. Moreover, according to Victor Zhirmunsky’s statement, verse rhythm, that is the actual alternation of stresses or quantities, is always a secondary formation as compared to meter. Rhythm is the realization of meter.

(Mihhail Lotman, “Verse Structure and its Cognitive Model: Hexameter and Septenary,” in Christoph Küper, pp 307-8.)

41. Derek Attridge:  It will be necessary to make one initial simplification: I shall ignore the fact that there exist rhythmic differences among the varieties of English spoken both now and in the past.

   Also, in various places, he writes:

We need a way of talking about poetic rhythm which will be useful for all varieties of English verse, which will reflect their interconnections and their dependence on the rhythmic characteristics of the language itself, and which will make sharp distinctions only where these are genuine perceptions experienced by the reader. (Italics added)

(In Derek Attridge,The Rhythms of English Poetry (London & NY: Taylor & Francis, 1982). See pages 11, 14, 16, 21-22, 24-27, and 60.)

   Gilles Deleuze & Felix Guattari provide us with a critique of the notion of an ideal hearer/speaker in which ideal objects of a monolithic language reside:

A semiotic chain is like a tuber agglomerating very diverse acts, not only linguistic, but also perceptive, mimetic, gestural, and cognitive: there is no language in itself, nor are there any linguistic universals, only a throng of dialects, patois, slangs, and specialized languages. There is no ideal speaker-listener, any more than there is a homogeneous linguistic community. Language is, in Weinreich's words, "an essentially heterogeneous reality."

(In A Thousand Plateaus: Capitalism and Schizophrenia, tr. Brian Massumi (MN: The University of Minnesota Press, 1987), p. 7. Emphasis mine.)

   The absence of an ideal speaker-listener not only subtracts the court of appeal on an idealized Ear (which no one possesses) but also on an idealized Eye as the corroborating authority of the truth of Reading. These disembodied semiotic verifiers are versions of the generic Lacanian big Other governing the negotiations of the Symbolic order, like the empty signifiers which Slavoj Žižek spoke about governing politics (the People, the State, the System, Structure, and so on).

42. Of the terms used to refer to this distinction, and the expressive possibilities it offers the poet, the least helpful are those which imply that there are two levels of structure simultaneously perceived by the reader; Hopkins’s influential borrowing of the term ‘counterpoint’ from music, for instance, gives the erroneous impression that the double structure is the equivalent of two voices in a polyphonic composition, each clearly perceptible, and each with a distinct character of its own. But what we are aware of in reading a metrical line is an onward movement which at times approaches a marked regularity and at times departs from it, constantly arousing and thwarting rhythmic expectations. It is in this sense that we can apply the term ‘tension’ to poetic rhythm, without implying foot-scansion and substitution, or the perception of two discrete patterns at different levels and a relationship between them....Tension arises out of the twin tendencies of language, towards variety and towards regularity: the voice, or rather the speech faculty of the human brain, enjoys its freedom to range over a finely gradated scale of intensities, timbres, pitches, and durations, but also feels the pull towards simple patterns and repetitions. This is a feature of all speech, perhaps of all human activity; but metre marks off the language of poetry from the language of daily existence by formalising and controlling this natural tension, and the classical approach to prosody has always shown an awareness of this central fact.

(Attridge, pp. 17-18, italics added.)

   Richard Cureton explains:

In terms of the general morphology of his theory... is to claim an additional construct on the rhythmic side of the rhythm-language dichotomy, a construct he calls underlying rhythm. In Attridge’s theory underlying rhythms generate a small number of highly abstract patterns, patterns that are then particularized by the conventional metrical patterns of specific verse traditions.

(Richard Cureton, Rhythmic Phrasing in English Verse (London & NY: Longman Group, 1992), p. 43. Italics in the source text.)

43. The central criterion of any metrological frame will, hence, find it difficult to skirt to question of “markedness” in the determination of the metrical and nonmetrical:

As Kager (1999) notes, at the outset of generative linguistics the concept of markedness, though acknowledged, was in general not intrinsic to the structure of rules. Over the forty years since, it has come to be understood as the very substance of phonology within universal grammar, with particular phonologies arising only from how the constraints which express markedness are ranked with respect to each other and to the faithfulness constraints which keep lexical representations recognizable (Prince and Smolensky 1993). In consequence, a theory of meter which takes meter to be a linguistic form cannot help but grant markedness an intrinsic role. In particular, one that takes it to be a stylization of the phonology of rhythm entails the inalienable involvement in the experience of meter of the entire gamut of universal intuitions about markedness defining the phonology of rhythm, and hence their availability for the creation of aesthetic effects.

(Kristin Hanson,Generative Metrics: The State of the Art,” in Christoph Küper, op. cit., p. 50-1. Italics added.)

44. Richard Cureton, Rhythmic Phrasing in English Verse, p. xii-xiii:

Rhythm is one of our basic cognitive abilities; therefore, we might expect a prosodic theory to take the same basic form as theories of other cognitive abilities – for example vision... We use our cognitive abilities to construct useful representations of the external world. A rhythmic structure is one of these useful representations. (Italics added)

   To carry any useful information, such a rhythmic design, however, still presumes an “experienced reader.” What has been experienced: natural or formal, remains the unanswered question.

45. See Cureton , Rhythmic Phrasing in English Verse, Chapter I, pages 1-70. Here is a brief summary of most approaches discussed. From “foot substitution” prosody (but aren’t all metrical frames substitution prosodies in one way or another?): strong and weak stresses represent lexical tonal or suprasegmental dimensions of the (English) language. Among the “phrasalists” are lineal “cadences” which are either seen as sonic or as semantic units. “Prose rhythmists” relate cola units to the oratorical beat because of their connection to the “laws of breathing.” Slavic theorists like Jakobson propose a “glottic” hierarchy in syntax. Derek Attridge views rhythm as pulses of energy combining verbal and nonverbal elements. The “intonationalists,” “generative metrists,” and “metrical phonologists” (where musical rhythms inform spoken language) all obviously allude to the phonological or phonetic dimension as the mimetic or counter-mimetic ground of metered and unmetered linguistic or poetic rhythm. David Wesling’s “grammetrics” postulates a primary connection between prosody and an abstract “voice,” a connection illustrated in the interaction of prosodic and syntactic unities, including speech segmentation. In free verse prosody, the awareness of metrical scheme has foregrounded “linguistic volatility” more than regularity. By having no “strict normative organization,” the free verse line is less the medium of “ordinary” speech or “thought groups” than the reorientation toward the visual rhythmic divagations of the line.

46. Miranda Cleary and David B. Pisoni, for example, note how “speech perception was equated for many years with what we would term “phoneme perception” (“Speech Perception and Spoken Word Recognition: Research and Theory,” in E. Bruce Goldstein, ed., Blackwell Handbook of Sensation and Perception (MA: Blackwell Publishing Ltd., 2005), pp. 499-534).

   Alice Faber, within the context of her essay on phonemic segmentation as a function of alphabetical thinking, refers to the dominance of the “phonemic conspiracy” in the analysis of language (Faber, p. 127).

   The volume edited by Christoph Küper, Current Trends in Metrical Analysis (Frankfurt am Main: Peter Lang, 2011), highlights the inescapable role of the spoken dimension in the both the writing and analysis of metrical language. For example, Peter L. Groves, in his essay on Shakespearean meter in the said volume, observes that:

Metrical structure in English is produced through the placement of beats in the spoken line, and this is enabled and limited by three prosodic phenomena: by the disposition of lexical and syntactic stress, by the location of syntactic junctures, and by the speaker’s (contextually motivated) placing of pragmatic accent (used for pointing contrast or highlighting information) within the utterance. (p. 145)

   The American linguist Richard T. Oehrle summarizes the major aspects of a phonovocalic metrological frame:

[W]e have tried to provide a set of parameters, based on psychophysical and phonological principles, that yields the observed cases of lines of English verse consisting of periodically iterating, temporally rigid, rhythmic figures. In doing so, we have relied on a number of hypotheses:

There is a natural temporal interval (called "the Measure"), which defines the domain of relative prominence in speech

• In periodic, iterating instances, the Measure can be divided into two, three, or four periodic subintervals, with the points of subdivision called "beats"

• Every beat either dominates an accompanying beat (or set of beats) or is dominated by an accompanying beat (or set of beats)

• If beat A dominates beat B, then beat B cannot be more prominent than beat A along the relevant parameter of prominence (namely, pitch, intensity, or duration)

(R.T. Oehrle,Temporal Structures in Verse Design,” in Paul Kiparsky & Gilbert Youmans, Rhythm and Meter: Phonetics and Phonology (CA: Academic Press, 1989) p. 106. Emphasis mine.)

   The American linguist Bruce Hayes, to give another example, views meter as phonological, that is, glottographic:

I would like to suggest that metrical rules NEVER refer to syntactic bracketing, only to prosodic bracketing. In other words, syntax has effects in metrics only insofar as it determines the phrasings of the Prosodic Hierarchy. This claim is the metrical counterpart of Selkirk's (1981) contention that syntactic effects in phonology are limited to the determination of phrasing. Intuitively, the hypothesis states that meter is essentially a phonological phenomenon; thus we might call it the Hypothesis of Phonological Metrics.

(Bruce Hayes, “The Prosodic Hierarchy in Meter,” in Paul Kiparsky & Gilbert Youmans, eds., Rhythm and Meter. Phonetics and Phonology, Volume 1 (CA: Academic Press, 1989), p. 224. Caps in the source.)

   That is to say, for him, prosodic meter is glottographic. However, suprasegmentals like “stress” are by no means fully represented in written language, and verse does not record them in a transparently explicit way. Secondly, as Jennifer Gross et al observe, the rather complex nature of English rhythm comes from the current realization in contemporary linguistics that “stress is a dynamic rather than... an immutable property of syllables.”

In sum, spoken and written English have a canonical rhythm created by an alternating pattern of stressed and unstressed syllables. The rhythm of English is not simply the by-product of the concatenation of words with given stress points. Rather, the stress-alternating rhythm of English is dynamic and responsive to contextual conditions. Stress and rhythm are not marked in English’s orthography. Skilled users of English must infer stress, and the stress-alternating preferences of its language, when writing and reading aloud.

   We would certainly want to think that scanning a text is a straightforward activity, requiring only the counting of syllables, accents, and stresses to determine the so-called “ideal, stress-alternating rhythm [of the] English” language. Yet, even if “spoken and written English have a canonical rhythm created by an alternating pattern of stressed and unstressed syllables,” these are not indicated by the orthography of the language.

(Jennifer Gross, Bo Winegard, & Andrea R. Plotkowski.”Marking Stress ExPLICitly in Written English Fosters Rhythm in the Reader’s Inner Voice,” Reading Research Quarterly, 53(3), 2017, pp. 305–321 | doi:10.1002/rrq.198, pp. 305-08. Emphasis added.)

   Thirdly, it has been recognized that the binary W-S alternating pattern for stress is an idealism. Since Chomsky & Halle, the number of stresses is theoretically three to four, raising the complexity further. Thus, the notion of an “unstressed” segment seems bizarre given the fact that all signs are aesthetic objects.  Finally, “stress is not a very well-defined property but rather a broad cover term for a set of properties that tend to cluster together.”

(Harry van der Hulst, “The Study of Word Accent and Stress: Past, Present, and Future,” in Harry van der Hulst, ed., Word Stress: Theoretical and Typological Issues (NY: Cambridge University Press, 2014), p. 8.)

47. From “A Grammar of the English Tongue,”A Dictionary of the English Language. A Digital Edition of the 1755 Classic by Samuel Johnson, https://johnsonsdictionaryonline.com/a-grammar-of-the-english-tongue/. Italics added.

   It would be the British poetics scholar Richard Bradford who, in his study of the development of “silent poetics” in 18th century England, would formulate the foundations of free verse in the dialectic of the “double pattern” of Eye vs. Ear in which “speech” becomes a shadow:

The shadow of speech is discernible when the balance between formalization and spontaneity is shifted so far towards the explicit and self-conscious manipulation of graphic materiality that speech becomes a memory, a shadow of its realization in sound...

(Richard Bradford, “Cummings and the Brotherhood of Visual Poetics” in Words Into Pictures: E. E. Cummings Art Across Borders, edited by Jirí Flajšar and Zénó Vernyik (Newcastle: Cambridge Scholars Publishing, 2007. p 13. Italics added.)

See his argument that a shadow of speech still lurks behind Cummings’ visual prosody, pp. 131 ff.

48. See William Wordsworth, “Preface to Lyrical Ballads” (1801). Italics mine.

   The glottographic orientation can still be felt dominant right up to T. S. Eliot:

Yet, in a paradoxical gesture, Eliot suggests that the movement away from elaborate forms of this kind and back towards the relative formlessness of colloquial language is also fundamentally ‘musical’. Unlike his fin- de-siècle precursors, Eliot does not see the sound of poetry as a separable element; for him, ‘the music of poetry is not something that exists apart from the meaning’. The patterns and rhythms of everyday speech are the basis for those of poetry; and, since this speech is always changing and renewing itself, poetry must periodically abandon its fixed and elaborated forms and reinvigorate itself by renewed contact with the colloquial. The result of this movement is a new and more subtle music, one which depends on the poet’s intimate knowledge of and sensitivity to the rhythmical structures of everyday language.

(Joseph Phelan, The Music of Verse: Metrical Experiment in Nineteenth-Century Poetry, (London: Palgrave and Macmillan, 2012), p 179.)

49. The American linguists Wallace Chafe and Deborah Tannen have noted “the inextricability of speaking and writing even in those modes of discourse that seem most exclusively a matter of writing and reading, and the inherently social nature of all discourse.”

(Wallace Chafe and Deborah Tannen, “The Relation between Written and Spoken Language,” Annual Review of Anthropology, Vol. 16 (1987), pp. 383-407, http://www.jstor.org/stable/2155877, pp. 396, 398.)

   As Geoff Hall, quoting Chafe and Tannen, has argued in his promotion of the “discourse”” as a broader category subsuming the domains of speech and writing, “there is no single feature or dimension that distinguishes all of speech from all of writing,” and “Distinctions between orality and literacy on the one hand, and spoken versus written language on the other, do not suffice to characterize real discourse . . . the relationship of literary to conversational language [is] . . . closer, and distinctions between them foggier, than had previously been thought.”

(In Geoff Hall, Literature in Language Education (NY: Palgrave MacMillan, 2005), p. 66.)

50. Deborah Tannen, in addition, refers to conversation as “daily, unrehearsed found art,” foregrounding the presence of many poetical procedures in discourse. This is what she calls a “poetics of discourse”:

My claim is that conversation is inherently poetic because of its structure, its use of figures of speech and ellipsis (or indirectness), imagery and detail, and its rhythmic or musical quality, all of which serve to move hearers (or readers)—that is, in Friedrich’s terms, to affect our imaginations, the pot in which knowledge, conviction, and emotion are brewed in aesthetic constraints.

(“Ordinary Conversation and Literary Discourse: Coherence and the Poetics of Repetition,” in Edward Bendix, ed., The Uses of Linguistics. Annals of the New York Academy of Sciences, vol. 583, 1990, p. 18.)

 51. Tannen (ibid.) writes:

Scattered findings from disparate fields... yield evidence of features common to literary language and ordinary conversation.... I group these in the categories of 1) rhythm 2) surface linguistic form and 3) contextualization or audience participation in sensemaking.

Surface linguistic form includes patterns of sound (alliteration, assonance, rhyme), morphology, lexical items, syntactic constructions, and discourse structures such as line and verse, and organization of information associated with conventional discourse genres such as narrative. Included here... are what Levin (1982) calls “style” figures of speech.

Audience participation in sensemaking includes indirectness, ellipsis (called implicature in conversation), imagery, detail, dialogue, and “thought” figures of speech (Levin 1982), or tropes.

52. Tannen, p. 19. Italics added.

53. Deleuze & Guattari, in A Thousand Plateaus: Capitalism and Schizophrenia, on “regime of signs” (p.140, italics in the original):

What is a semiotic, in other words, a regime of signs or a formalization of expression? They are simultaneously more and less than language. Language as a whole is defined by "superlinearity," its condition of possibility; individual languages are defined by constants, elements, and relations of a phonological, syntactical, and semantic nature. Doubtless, every regime of signs effectuates the condition of possibility of language and utilizes language elements, but that is all. No regime can be identical to that condition of possibility, and no regime has the property of constants. As Foucault clearly shows, regimes of signs are only functions of existence of language that sometimes span a number of languages and are sometimes distributed within a single language; they coincide neither with a structure nor with units of a given order, but rather intersect them and cause them to appear in space and time. This is the sense in which regimes of signs are assemblages of enunciation, which cannot be adequately accounted for by any linguistic category: what makes a proposition or even a single word a "statement" pertains to implicit presuppositions that cannot be made explicit, that mobilize pragmatic variables proper to enunciation (incorporeal transformations). This precludes explaining an assemblage in terms of the signifier or the subject, because both pertain to variables of enunciation within the assemblage.

   A regime of signs is, hence, a formalization of superlinear or pragmatic factors which exceed the formalized confines of the regime.

54. Tannen, pp. 19-20.

55. In the context of new media and textual criticism, “script act theory” as conceived by the textual criticism scholar Peter L. Shillingsburg, is the term referring to techno-social processes in the constitution of the parameters that define textuality:

By script acts I do not mean just those acts involved in writing or creating scripts; I mean every sort of act conducted in relation to written and printed texts, including every act of reproduction and every act of reading....

‘‘Script acts’’ identifies my overarching subject: how constructions of texts and constructions of understandings from texts in individual acts of writing and reading ‘‘happen’’ (or don’t).

(Peter L. Shillingsburg, From Gutenberg to Google: Electronic Representations of Literary Texts (NY: Cambridge University Press, 2006), pp. 40, 41.)

   In the context of this paper, such a global sense of textual reproduction is indeed useful in the way it goes beyond the naked or unmediated view of speech acts. Strictly linguistic texts cannot, of course, include nonverbal cues like those in face-to-face conversation, but it is less a question of absence or presence, or the poverty of the writing vis-à-vis actual or real-time experience, than a difference in the channel of reiteration of memories in various media. Objects, for example, are iconic reiterations of themselves and are, therefore, signs in nonverbal formats existing in redundancy (and not in faithful “representation”) with other formats like < table >.

56. Deleuze & Guattari:

There is no individual enunciation. There is not even a subject of enunciation. Yet relatively few linguists have analyzed the necessarily social character of enunciation.... The social character of enunciation is intrinsically founded only if one succeeds in demonstrating how enunciation in itself implies collective assemblages. It then becomes clear that the statement is individuated, and enunciation subjectified, only to the extent that an impersonal collective assemblage requires it and determines it to be so. It is for this reason that indirect discourse, especially "free" indirect discourse, is of exemplary value.... Indirect discourse is not explained by the distinction between subjects; rather, it is the assemblage, as it freely appears in this discourse, that explains all the voices present within a single voice....

That is why every statement of a collective assemblage of enunciation belongs to indirect discourse. Indirect discourse is the presence of a reported statement within the reporting statement, the presence of an order-word within the word. Language in its entirety is indirect discourse. Indirect discourse in no way supposes direct discourse; rather, the latter is extracted from the former, to the extent that the operations of signifiance and proceedings of subjectification in an assemblage are distributed, attributed, and assigned, or that the variables of the assemblage enter into constant relations, however temporarily. Direct discourse is a detached fragment of a mass and is born of the dismemberment of the collective assemblage; but the collective assemblage is always like the murmur from which I take my proper name, the constellation of voices, concordant or not, from which I draw my voice.

(pp. 79, 84, italics added.)

57. Caroline Féry, p. 16. Emphasis mine.

58. Wallace Chafe, “Punctuation and the Prosody of Written Language,” in Written Communication, Vol. 5 No. 4 (October 1998), pp. 396-426, doi 10.1177/0741088388005004001.

59. As the linguists Jennifer Gross, Bo Winegard, & Andrea R. Plotkowski explain:

In contemporary linguistics, stress is viewed as dynamic rather than as an immutable property of syllables.... In sum, spoken and written English have a canonical rhythm created by an alternating pattern of stressed and unstressed syllables. The rhythm of English is not simply the by-product of the concatenation of words with given stress points. Rather, the stress-alternating rhythm of English is dynamic and responsive to contextual conditions. Stress and rhythm are not marked in English’s orthography. Skilled users of English must infer stress, and the stress-alternating preferences of its language, when writing and reading aloud.

(Jennifer Gross, Bo Winegard, & Andrea R. Plotkowski.”Marking Stress ExPLICitly in Written English Fosters Rhythm in the Reader’s Inner Voice,” Reading Research Quarterly, 53(3), 2017, pp. 305–321 | doi:10.1002/rrq.198.)

60. Katy Carlson, “How Prosody Influences Sentence Comprehension,” in Language and Linguistics Compass 3/5 (2009): 1188–1200, 10.1111/j.1749-818x.2009.00150.x

61. The pre-punctuated “neutrality” of scriptio continua could be viewed as the ground state of the prosaic stream. As the late English palaeography scholar Malcolm B. Parkes has remarked:

The merit of scriptio continua was that it presented the reader with a neutral text. To introduce graded pauses while reading involved an interpretation of the text, an activity requiring literary judgement and therefore one properly reserved to the reader.

(Malcolm B. Parkes, Pause and Effect:  An Introduction to the History of Punctuation in the West (NY: Routledge, 2016), p. 11.)

   The redefinition of what constitutes punctuation has been the subject of the English literary scholar John Lennard’s “Mark, Space, Axis, Function: Towards a (New) Theory of Punctuation on Historical Principles.” The editors explain the import of Lennard’s redefinition in the following introduction to the essays in the volume they edited:

John Lennard challenges this distinction, proposing a new axis of punctuation on eight levels and examining the historical development of each. This axis ranges from level 1, the letter-forms which punctuate the blank page, to level 8, the book itself as a complete object punctuating space. Lennard's extended view of punctuation intersects with McGann's retheorisation of the text as comprising 'lexical' and 'bibliographical' codes. According to Lennard, levels 1-4 (including punctuation marks and details of the mise-en-page ) may reasonably be labelled as lexical, while levels 5-8 (including layout and spacing) seem more bibliographical. The fact that punctuation thus blurs the codes would not, Lennard notes, pose a problem for McGann, for whom the lexical and bibliographical are not exclusive and opposed, but rather constantly intertwined.

(In Joe Bray, Miriam Handley, Anne C. Henry, eds., Ma(r)king the Text: The Presentation of Meaning on the Literary Page (London: Routledge, 2018), https://doi.org/10.4324/9780429432309, p. 16ff. For Lennard’s essay, see Chapter 1, pp. 24 ff.)

62. See Richard Bradford, Graphic Poetics: Poetry as Visual Art (London: Continuum, 2011), pp. 138-139. Italics added.

   Bradford also calls it “silent poetics,” a part of what he calls the “double pattern” which opposed the visual against the verbal in the history of poetry. See note # 80 below.

63. See entry on “Line” in Roland Greene et al., eds., The Princeton Encyclopedia of Poetry and Poetics, 4th ed. (NJ: Princeton University Press, 2012) pp. 801-803.

64. Chafe, op. cit., p. 406.

65. Roland Greene et al., op. cit., pp. 1118-9, emphases mine. In this connection, see also Reuven Tsur’s discussion of the pentameter and the caesura being determined most often after the fourth syllable (a 4/6 pattern iconic of the iambic beat) in his Poetic Conventions as Cognitive Fossils (2017). He argues that this is a cognitive constraint stemming from the chunking threshold of the brain’s information processing rate.

66. Throughout the history of punctuation attitudes to the use of the symbols have been bound up with developments in traditional attitudes to discourse. These are reflected principally in discussions of punctuation based on different modes of analysis: grammatical analysis and rhetorical analysis, which are most commonly found in the works of schoolmasters in all periods. Grammatical analysis has been concerned with the application of punctuation to identify the boundaries of sententiae (later, 'sentences') and the units of sensus or grammatical constituents within them. Rhetorical analysis has been concerned with the ways in which punctuation reflects the periodic structure of a discourse, and indicates the periodus and its parts (commata or incisa, cola or membra). With its emphasis on pauses for breath this mode of analysis has been preoccupied with bringing out correspondences between the written medium and the spoken word. A rhetoricians' periodus should not be confused with a grammarians' sententia (the length of a periodus was a matter of opinion), and cola and commata should not be confused with such grammatical units as clauses (some of the examples cited by Cicero, in Orator,§ 225, do not contain a verb). However, there is usually some agreement between grammarians and rhetoricians as to what constitutes incomplete and completed sense.

(Malcolm B. Parkes, op. cit., pp. 3-4.)

67. Elizabeth R. Schotter, Bernhard Angele, & Keith Rayner, “Parafoveal Processing in Reading,” Atten Percept Psychophys (2012) 74:5–35, DOI 10.3758/s13414-011-0219-2, p. 6.

   The large body of empirical studies on the physiology of visual reading and perception is beyond the scope of survey and evaluation for this paper. Just to cite another research along these lines:

We directly investigated in a new experiment the association between speech spectrum and eye-movement sampling frequency at a person-specific level and found a significant correlation. Based on this evidence, we argue that during reading, the rate of our eye movements is tuned to supply information to language comprehension processes at a preferred rate, coincident with the typical rate of speech.

(Benjamin Gagl et al., “Reading at the Speed of Speech: Alignment of Eye-Movement Sampling in Reading with the Speech Production Rate,” bioRxiv preprint first posted online, Aug. 14, 2018; doi: http://dx.doi.org/10.1101/391896.)

68. Peter Gallman, “The Graphic Elements of German Written Language,” in Gerhard Augst, ed., New Trends in Graphemics and Orthography (Berlin & New York: Walter de Gruyter, 1986).

   Gallman imagines abstract supragraphemes as a category in analogy with phonological suprasegmentals. His table of graphic means from p. 49 is reproduced below.

                           


69. Gallman, op. cit., p. 54, italics mine.

70. The oppositions between concrete and abstract, basic and marked graphemes would be difficult to sustain, given the premise that all graphemes are abstract entities to begin with, resulting from the combination of more fundamental graphetic elements on the featural level. (See Beatrice Primus, “A featural analysis of the Modern Roman Alphabet,” Written Language & Literacy 7:2 (2004), pp. 235–274. https://doi.org/10.1075/wll.7.2.06pri). That there is no basic unmarked grapheme should have been already apparent from Gallman’s listing of the polyvalence of the dash <  -  > (Augst, p. 74). The use of “context” as an explanation of this polyvalence becomes a way of maintaining the notion of an integral or essential form, curtailing the radical idea that any form is already an overlay of various abstract supragrapheme operations.

   Even if we concede that a pure concrete mark or grapheme can exist in the state of a value-free “ceneme” (Louis Hjelmslev), its codification with the value of the “absence of value” places it in the outer limits of the operation of abstract supragraphemes, a placement possible only from the point of view of the success of that operation. (See the remarks on Gomringer’s “Silencio” concerning the paradoxical nature of having “silent” glyphs or writing without any semiotic value.) To bring home the point more sharply, the full implication of Gallman’s definition leads to saying that the whole writing system itself is a grand abstract supragrapheme. It is the writing system's phonographic or logographic principles which abduct a cenemic mark to overlay it with a semiotic charge, transforming it into a “plereme” (Hjelmslev) with marked signatures like “speech” or “writing.” That is, it is only within the operations of a writing system that a glyph obtains a letteral or non-letteral, punctuational or non-punctuational, grammatical or non-grammatical valence. (As we can already foresee, it is for the suspension of such essentialist grammatological distinctions that E. E. Cummings accrued the fame and notoriety of his poetics.)

   In grammatology, letters are not simple basic graphemes but are the bearers of several properties. The chart below from Gabriel Altmann illustrates this clearly.



(From Gabriel Altmann, “Towards a Theory of Script,” in Gabriel Altmann & Fan Fengxiang, eds., Analyses of Script Properties of Characters and Writing Systems (Berin & NY: Mouton de Gruyter, 2008), p. 150.)

71. The poet Denise Levertov characterizes lineation as a form of punctuation: “The line-break is a form of punctuation additional to the punctuation that forms part of the logic of completed thoughts. Line-breaks together with intelligent use of indentation and other devices of scoring represent a peculiarly poetic, alogical, parallel (not competitive) punctuation.”

(From Jan Mieszkowski, Crises of the Sentence (Chicago : The University of Chicago Press, 2019), p. 85 footnote.)

   John Lennard (2018), alluding perhaps to a passage Malcom Parkes’ Pause and Effect, describes the punctuational role as “the pragmatics of the written text.”

72. Just to mention a few titles under or about this deviationist emphasis are:

1. Silvia Chirila, “Ambiguity and Idiosyncratic Syntax in the Poems of E. E. Cummings,” AMBIGUITIES, 2, no. 2 (2009): 61-72. Italics added. Online version: https://blogs.kent.ac.uk/skepsi/files/2010/06/vol-2.2-5-Chirila.pdf.

2. Sami B Al Hasnawi, “Morphological Deviation as a Stylistic Marker in E.E. Cumming's Poetry,” Journal of Al-qadisiya in Arts Education and Science, 7 (January 2008).

3. Xin Li & Mengchen Shi, “A Stylistic Study on the Linguistic Deviation in E. E. Cummings’ Poetry,” Journal of Pan-Pacific Association of Applied Linguistics, 19, no. 2 (2015): 23-54. Online version: https://files.eric.ed.gov/fulltext/EJ1092438.pdf.

73. The well-known Swiss linguist Martin Hilpert, in his study of English language construction, summarizes the issue thus:

In synchrony, constructions are studied because of the insight that grammatical idiosyncrasies are ubiquitous: Not all transitive clauses can be passivized, not all verbs of communication can occur in a ditransitive construction, and not all monosyllabic adjectives exclusively form the morphological comparative. Exceptions, it turns out, are the rule, so that any attempt to reduce grammar to a single underlying set of rules, variable or not, is bound to fail. The slogan “Grammars contract as texts expand” (Hopper 2010) captures the idea that with every new text genre and every additional speaker, more variation enters the picture, and less of an invariant core system remains.

(Martin Hilpert, Constructional Change in English: Developments in Allomorphy, Word Formation, and Syntax (Cambridge: Cambridge University Press: 2013), p. 2. Italics added.)

74. Many of the deviations observed in Cummings are seen to be surface phenomena that are amenable to normalization, implying that there is a canonical syntax and grammaticality under such surface stylization. Many of his morphological deviances turn out to be in line with standard morphological processes which are simply carried over locations not usually subjected to such processes.

[T]here is in fact much less deviation than it might seem. Though the extensive use of conversion leads to superficial ambiguity, one can obtain a well-formed and relevant syntactic structure, and thus a semantic interpretation, through a good knowledge of Cummings’ system of values and using phonetic, morphological and syntactic parallelism. Grammatical analysis can thus explain certain clear intuitions and help reach an interpretation for some less intuitively obvious passages.

(Philip Miller, La Déviation Grammaticale chez E. E. Cummings. Une Étude de what
if a much of a which of a wind, Études anglaises 2004/2 (Tome 57): 187-201, https://www.cairn.info/revue-etudes-anglaises-2004-2-page-187.htm.)

75. Deleuze & Guattari:

Words are not tools, but we give children language, pens, and notebooks as we give workers shovels and pickaxes. A rule of grammar is a power marker before it is a syntactical marker.

(Deleuze & Guattari, 76).

   For the politics of literacy and its role in the determination of correct forms, see Mark Sebba, Spelling and Society: The Culture and Politics of Orthography around the World (Cambridge: Cambridge University Press, 2009).

76. The application of theories of complex systems would go beyond the traditional six or seven subsystems of phonetics, phonology, morphology, lexicology, syntax, semantics, and pragmatics in linguistics study. Titles such as Complexity Applications in Language and Communication Sciences (2019) or Complexity Perspectives on Language, Communication and Society (2013), “Evolutionary dynamics of language systems” (2017) or the discussion of “Complexity Theory” or “Dynamic Systems Theory” in the Encyclopedia of Applied Linguistics by Carol Chapelle, (Oxford: John Wiley and Sons, Inc., 2013)) all testify to how far linguistic research has gone from the rudiments of grammatical or synchronic modeling. I can only allude to such domains in this paper focusing on the intersection of the graphic and phonic subsystems. Just to cite a passage in this connection from Carol Chapelle’s Encyclopedia entry on “Dynamic Systems Theory Approaches”:

In sum, from a DST point of view, the language system can be assumed to consist of embedded subsystems for all levels of language production and perception, such as conceptualization, semantics, syntax, lexicon, phonology, and phonetics.... Contrary to the modular approach, dynamic subsystems must be assumed to be open, interacting, and emergent systems.

   What is, therefore, superficial and that demands explanation is the regularity or symmetries that are posited as universal across a wide range of events in the sea of noise and variation. Manuel DeLanda explains how in biology, population thinking is opposed to typology in the way types are seen not as archetypes but individual operation on larger time scales. The type is then only an average or a mean. In a note, he writes:

While for Aristotelians homogeneity is the natural state and variation is what needs special explanation, for population thinkers it is variation which is natural, while homogeneity, when it exists, is what needs to be explained.

(Manuel DeLanda, Intensive Science and Virtual Philosophy (NY: Continuum, 2002), pp. 47-8.)

77. The American literary scholar Gillian Huang-Tiller, in characterizing the meta-formal critique that underlines Cummings’s prosodic modification of the Sonnet genre, focuses attention on the manner by which the Sonnet in Cummings’s hands was “destabilized” from its fixed and “festishized” iconic status as the structure of “aesthetic and cultural purity.”

Cummings not only published experimental typographical verse but also gave a prominent place to his experimental sonnets, which form the crucial part of his structural design in each publication and visually turn the sonnet into the iconic meta-form. He observes this formal task by calling attention to the process through which the sonnet achieved iconic status and by exposing this status as artifice. The self-referentiality and deliberate violation of conventions in Cummings’ sonnets explode the myth of the sonnet’s purity and estrangement from daily life, commenting on the genre and culture both aesthetically and critically....

For one, Cummings’ use of the visual form of the sonnet as a structural device directs our attention to the iconicity of the genre itself, along with the cultural prestige and stylized emotions and feelings its long pedigree supports. We have to wonder what effect this visual performance of the iconic “meta-sonnets” has, besides giving the “additional shape to the rise and fall of narrative tension” within the individual sonnet that Cureton finds.... It seems that the mimetic structural play of the sonnet affords Cummings an effective means of destabilizing the fetishized form, thereby freeing the genre from its convention. In so doing, Cummings’ “countersonnets” in No Thanks are indeed meta-sonnets—a self-reflexive comment on the iconic fixity of the traditional genre.

(Gillian Huang-Tiller, “The Iconic Meta-Sonnet, Manhood and Cultural Crisis in No Thanks,” Jirí Flajšar and Zénó Vernyik, eds., Words into Pictures: E. E. Cummings’ Art across Borders (Newcastle: Cambridge Scholars, 2007), pp. 27-57.)

78. No Thanks is generally recognized as Cummings' most experimental volume, and therefore it should represent the extremes of his visual practice (for instance, it contains his most often cited visual experiments, 'r-p-o-p-h-e-s-s-a-g-r' and 'bright'). At the same time, however, the volume also contains a large number of poems in more conventional forms (every fourth poem is a sonnet), and it provides a balanced sample of poems across his characteristic thematic 'genres' (satires, comedies, impressions, transcendental poems, and poems of the city).

(R. Cureton, “Visual form in E.E. Cummings' No Thanks,” p. 246.)

79. The particular care with which Cummings handled (or overmarked) the notion of the line as a periodic device is described by Huag-Tiller as Cummings’ “lineation code”:

The variations of the lines for each sonnet change arbitrarily from stanza to stanza. Furthermore, in most of the iconic quatrains or tercets, Cummings embeds split lines, rendering line structure into something more than four or three lines. With the use of the split lines or descending triadic lines, what William Carlos Williams might term as “variable foot” in his discussion of “A New Measure,” darting across the sonnet plane, Cummings’ sonnets are both sonnets and meta-sonnets engaged in mimicry of the form. The composition of lines in various patterns further casts light on broader interpretive possibilities, based on his play with form and his use of form, syllable counts, and line numbers as metaphors.

The variation of line numbers reinforces Cummings’ demonstrable concern for numbers and sequencing or rhetorical patterns that the orthodox sonnet represents and contains, at the cost of spontaneity and freedom. Cummings’ own notes on lineation in the unpublished papers housed in the Houghton Library show how he frees the idea behind his numerical patterning. Cummings deems 1,2 as simple lines; 121, 123, 1212, 122 as compound lines: 11 as “repetition”; 12 as “change”; 123 as “direction,” “body goes somewhere,” or movement in “time,” or “Growth” (life to death); 1212 meaning “ad infin., circular, a returningness, parallel repetition”; 122, meaning “incomplete, up in the air, unfinished like a poem ending w, a comma”; 13-2 as “jump.” Cummings further writes that “certain motions wh. pay no attention to (are absolutely independent of) the 123; those are based on 2 things: change 12 (movement//iteration 11 (standing still”).  Borrowing Annie Finch’s term “metrical code,” (1993, 3) I consider Cummings’ notes on lines as his “lineation code.” Using this code, Cummings seems to give new meaning and life to the line pattern of each of his sonnets that the convention of the sonnet form has stultified.

(In Jirí Flajšar and Zénó Vernyik, p. 34-5.)

80. R. Bradford has framed this literary question in terms of the historical conflict between two notions of the double pattern. The traditional one involved seeing all of poetry in terms of the tension between meter and language as phonolinguistic phenomenon, while the second one places a third dimension beside this two to form a new and more complex double pattern of “silent poetics” or visual against verbal prosody.

I emphasize the mutual dependency between the ideal of the double pattern and the methodological belief in oral counterpoint in order to prepare the ground for my examination of the concept of silent poetics. Counterpoint, to be consistent with its origins in music, denotes the simultaneous production of two contrasting effects. But there is already evidence that some readers, such as Llewellyn Jones, find that such interactions can depend as much upon the existence of ‘a convention in the mind of the reader’ as they do upon what Attridge calls ‘the inherent tension of the line’. The possibility that counterpoint could in certain types of poem register as a distinction between what we see and what we hear – silent poetics – becomes evident in an unwitting debate that took place between the eighteenth-century critic Thomas Sheridan and John Hollander, arguably the most incisive modern commentator on the visual/oral dimensions of poetry. Often focusing upon the same lines from Paradise Lost, these critics reached two separate conclusions on how the visual structure of verse affects the transfer of meaning through the communicative circuit between text and reader. Hollander’s apparently innovative theories of visualism remain anchored to the orthodox interpretative prominence of speech over writing, whereas Sheridan developed a method of reading in which the contrapuntal relation between verse design and verse instance is divided, respectively, between the eye and the ear.

(Bradford, Graphic Poetics, pp. 19-20.)

81. In a note to his survey of approaches to rhythm, Cureton observes that “the major find of the free verse poets in this century [is] visual form.”

(Cureton, Rhythmic Phrasing in English Verse, p. 72.)

   Bradford, on his part, traces the beginnings of visual prosody to as far back as John Milton and the debates in the 18th century British literary landscape. See his article, “The Visual Poem in the Eighteenth Century” (in Visible language XXII, no. I) or his discussions of the same debates in his book Graphic Poetics.

82. The three-body problem exclusively involves the prosodic dimensions in theory only, for the reason that syntax and grammar are viewed as non-prosodic, that is, as organizational aspects of langue, and not rhythmic components of parole. Considered as primarily primitive formal structures first, they obey only the dictates of their self-internal rules, and don't need to follow or be patterned after the prosodic aspects of speech to function, i.e., they don’t need any specific speaker for their operational existence. For Bruce Hayes (note # 46), meter also does not refer to syntactic but to phonological or prosodic segmentation, an assumption which secures the naturalistic grounding of prosody. Even when subjected to prosodic redistribution, syntax and grammar retain the role of a stable abstract frame (even if grammar admits many exceptions or that syntax itself has no natural foundation, that is, until the innatism of generative linguistics) along or against which any prosody is arranged.

   However, the question remains regarding what role “beatless” or “toneless” syntax plays, if it has any, in a predominantly phonolinguistic prosody, apart from the traditional bearer of complete or incomplete “thought” (the Sentence). Because it does not have the primary command in prosodic territory, it is relegated to being a pure systemic artefact of language like the writing system: necessary but impersonal and mere tools to an end: glottography. Hence, the eternal question persists: is syntax dependent on prosody for its partitions and hierarchies, or is it the other way around? Is there a forgetting that syntax might have had some prosodic origin, an origin in the creation of grammatical categories, cognitive salience, or markedness where the verbal is distinguished from the nominal or adjectival, the grammatical from the lexical, the syndetic from the asyndetic, the tonal from the atonal, or the head from the argument, but then grammaticalized or synchronically “promoted”? Would the fixed locution “parts of speech” traditionally referring to the grammatical categories of syntax be any indication of such an origin? Or, wouldn’t this simply indicate the long equation of language with speech and thought since ancient times? The “primitive” status of syntactic categories may no longer be axiomatic today, but we still wonder where and how—in both Constituency and Dependency grammars—they get their primitive status out of a mass of input. In the area of Construction Grammar, for example, construction templates which have cognitive and pragmatic import are the fundamentals of syntax, and not grammatical categories.

83. Citations of Cummings’ work will be taken from the edition by George J. Firmage, E. E. Cummings: Complete Poems, 1904-1962. Revised, Corrected, and Expanded Edition Containing all the Published Poetry (NY: Liveright Publishing, 1991). “On souls robbed of their birth-right’s better part” is the Petrarchan sonnet in p. 1070 of that edition.

84. So far, Richard Cureton is perhaps the only poetics scholar who has thoroughly investigated the iconic aspects of E. E. Cummings’s oeuvre. Aside from some parts of his essay on “Visual Form in Cummings’ No Thanks,” his major research work of poetico-linguistic iconicity is found in his “E. E. Cummings: A Study of Iconic Syntax” (1981) based on his PhD Dissertation work. The exploration of iconicity as a compositional principle in E. E. Cummings will be limited to the notion of prosodic alignments in this paper, apart from a few forward looking notes on its rapport with the structural principle of semiotic or poetic assembly.

   He differentiates iconicity from parallelism in the passages below. I, on the other hand, have a broader notion of iconicity as any form of diagrammatic analogy.

Iconicity involves "resemblances" between a form of expression on one linguistic level and a theme or form of expression on another level. Thus, iconic effects are interlevel (as opposed to intralevel, parallel) correspondences. The clearest case of iconism on another linguistic level is sound symbolism, but other linguistic icons are common as well. For instance, Cummings often uses the visual presentation of his poems for iconic effects (e.g., the poem will fall down the page like a falling leaf it describes, or, on a smaller scale, Cummings will frequently write moon as mOOn, so that it contains orthographic moons).....

While iconicity is interlevel "coupling" where the degree of similarity of the sign and the object it refers to creates the major effect, parallelism is intralevel "coupling" where the differences in the meanings of the "coupled" forms create the major effect. In iconicity, the form of the iconic sign is_ the effect, and, therefore, this effect is usually perceptual. In parallelism, the form of the sign is a means to another effect—a conceptual comparison between the differences or similarities in the meanings of the items coupled. As many commentators have pointed out (e.g., Jakobson (1960), Kiparsky (1973), Levin (1962), Leech (1969) ), parallelism occurs at all levels of linguistic form. On the phonetic level, phonetic schemas such as rhyme and alliteration "couple" items. On the morphological level, items can be coupled with shared bases or derivational and inflectional affixes, and, on the orthographic level, one again could use capitalization or simply the visual placement of items to create a parallel. Of course, on the semantic/narrative level, parallelism can be achieved in a variety of ways—for example, by placing different characters in similar situations in order to "couple" their various reactions to that situation.

(Richard Cureton, The Aesthetic Use of Syntax: Studies on the Syntax of the Poetry of E. E. Cummings. Ph.D. Dissertation, University of Illinois, 1980. pp. 250, 253. Online: https://www.researchgate.net/publication/35613457_The_Aesthetic_Use_of_Syntax_Studies_on_the_Syntax_of_the_Poetry_of_EE_Cummings.)

85. J. Alison Rosenblitt, E.E. Cummings’ Modernism and the Classics: Each Imperishable Stanza (Oxford: Oxford University Press, 2016), pp. 33-34

86. Rosenblitt, ibid.

87. R. Bradford cites the American critic John Livingstone Lowes (1919) to explain one of the key revolutions of free verse: “What free verse would strike out . . . is the recurrent rhythm of the line.” One step further from this is dealing a blow to the very notion of the line itself:

With the arrival of free verse, accentual pattern and syllabism, the remaining concessions to regularity, were discarded. The line became something that evaded abstract definition; it was neither a syntactic unit nor a measure of metrical regularity. Its use as an axis between what we hear and what we see is in my opinion the most fruitful and innovative consequence of the free-verse revolution – but it would be a decade after the birth of the new form before William Carlos Williams and e. e. cummings shook themselves free of the restrictive, phonocentric conventions of the first generation....

(R. Bradford, pp. 11, 14).

88. R. Cureton, 1986. As Cureton has noted, the use of tmesis in Cummings is an iconic way of indicating an intimate association between the forms and meanings of different terms, with negative or positive outcomes depending on context.

89. R. Cureton (1986) explains how Cummings’ visual and vocal range straddles the macro-geometric down to the microscopic level of the linguistic hierarchy; that is, from the level of the page down to the features of glyphs.

90. In the framework of Cognitive linguistics, Cummings would be using the schema Good is Up and Down is Bad. See the well known work by George Lakoff & Mark Johnson, Metaphors We Live By (2007).