Visual Poetics: Semic, Asemic, Nonsemic . . . and AI
by Michael Betancourt
Abstract:
The wide spread emergence of machine learning has brought the question of an ‘AI poetics’ into focus as these systems demonstrate the centrality of human perception and interpretation to the creation of meaning. Rather than rendering human creativity obsolete, AI emphasizes the importance of reflective, critical judgment supported by connoisseurship. These developments are not unprecedented: the recognition of language in visual poetry and the collage techniques invented by the historical avant-garde anticipate the nonsemic works of contemporary AI by attenuating the semiotics of intentional encoding. This capacity to impose coherence on disparate elements arises from distinctions between reading and seeing that are not inherent to written language, but emerge from the entanglement of perception and enculturation. They reveal how semantic cues and past experience guides the recognition and completion of fragmentary or ambiguous forms to play a central role in poeisis.
Article:
All written language, both handwriting and typography, is some form of mark on a surface, typically a page or screen. This commonplace fact requires overlaps between writing and the other visual arts, but their semiosis differs depending on the initial assessment of ‘what is seen.’ Subsequent apperceptions either remain within diagnostic recognition (depiction) or become symbolic interpretation (denotation). The rearticulations of poeisis engages the “visual or material properties of language”[i] by exploiting the human capacity for shifting between modes and developing the multiplicity of unbounded semiosis.[ii] How the human audience creates meaning is circumstantially sensitive, an aspect of aesthetic consideration that becomes increasingly obvious when confronting the ‘nonsemic category of expression’ enabled by AI, whose capacity to produce works that invite the same shifting and polyvalent engagements as poetry[iii] emerges from the ambivalence and instability of perception, however the multiple decisions that reframe and reinterpret apperception follow predictable paths that semiotician Umberto Eco termed “recognition”:
Recognition occurs when a given object or event, produced by nature or human action (intentionally or unintentionally), and existing in a world of facts as a fact among facts, comes to be viewed by an addressee as the expression of a given content, either through a pre-existing and coded correlation or through the positing of a possible correlation by its addressee.[iv]
Eco’s concern is with the communication of a specific meaning, not how perception becomes apperception, yet when confronting visual poetry as its legibility decreases towards asemic form, and its identity as-language becomes ambiguous, communication becomes unstable, complicating his idea of ‘recognition’ as an easy diagnostic. The nonsemic products of AI systems expose the human dimensions of this process as inextricable from encounters with ambivalence.[v] The same semantic cues that allow the identification of a/semic writing are modeled by the nonsemic expressions generated by AI [Figure 1]. This ‘page of illegible handwriting accompanied by doodles’ demonstrates how the emergence of an ‘AI poetics’ through the nonsemic produces a weird return to the problematic distinctions between art and non-art.[vi] Machine learning thus reveals how those essential diagnostic recognitions that begin semiosis depend on past knowledge and experience to define legibility,[vii] which then tautologically directs/determines when it is appropriate to understand a selection of marks are writing or letterforms.[viii]
AI automates the “creative industry,”[ix] creating things that resemble human poetics precisely because the database employed by AI is purely human-made material—it contains the essential cues recognized by human audiences.[x] The problem of communication posed by these outputs is explicit: communication assumes the encoder comprehends the meaning being expressed, has something to communicate, a prerequisite for it a statement being encoded.[xi] AI does not understand, invalidating this assumption. They create an illusion of communication that should be neither surprising nor unexpected given the operations of statistical generation; AI produces nonsemic outputs because their ‘expressiveness’ happens without encoding only because these machines are explicitly designed and built to produce the response that their outputs “signify.” It belongs in the same ambivalent category of pseudo-signification as parrots imitating human speech (although the parrot’s speech may still occasionally be a meaningful demand for a “cracker”). Rather than being an expression, the nonsemic either contains familiar semantic cues whose encoding is implicit (a product of the material being collaged/reproduced already being encoded), or has an illusory encoding fully invented by the audience—both potentials render any decoding (interpretation) of uncertain status.[xii]
What complicates the question of encoding is that AI operations often resemble the a/semic creations of human poets,[xiii] and the expressive work of artists who exploit ambivalence and ambiguity.[xiv] How twentieth century avant-garde works by Dada and Surrealist poets employed collage to create nonsemic poetry parallels its role in AI generation. These types of writing and their expressive and communicative dynamics can be fully simulated by AI systems without any requirement for comprehension by the machine, because the avant-garde used them as exploit-demonstrations to challenge the protocols of expression used by human audiences. The Surrealist poem “Tu seras parmi les victims” (You will be among the victims) produced by André Breton in ~1924 [Figure 2] is a randomly assembled sequence of texts cut from magazines and newspapers that produce a series of in/coherent statements. The avant-garde produced nonsemic work as a challenge to coherence and interpretation, attempting to eliminate familiar recognitions, as Raul Hausmann explains about his Dada ‘phonetic poetry’ in 1962:
The phonetic poem may be compared with a drum that can play half and quarter tones, as for example in Arabic or Chinese music, where these notes lend their timbres at different pitches to the general melody, changing intonation and thus altering the sense of words pronounced. A similar process occurs in phonetic poetry through architectural arrangement of contradictory groups of vowels and consonants as well as other non-linguistic noises; snoring, humming, croaking and so on. […]
Take any sentence and print it backwards!
sdrawkcab ti tnirp dna ecnetnes yna ekat !
Now redistribute the sentence on different lines
sdrawkcab
ti tnirp dna
ecnetnes yna
ekat !
and that’s how to write a phonetic poem![xv]
“Music” is Hausmann’s model, posed in opposition to the communication of language, a challenge to coherence since it includes what would typically be considered ‘noise.’[xvi] His Dada protocol for ‘phonetic poetry’ resembles an algorithm, adaptable to any input, but producing a predetermined output that transforms its source material in predictable ways. The spontaneous emergence of recognizable words within the ‘nonsense’ of Hausmann’s example converges on the allusive significance of Breton’s Surrealist poem “Tu seras parmi les victims.” Each piece in itself is fragmentary, but their combination is allusive: Breton made it from a random sequence of texts pulled from a hat, while each “choice” by an AI is guided by an algorithm (statistical likelihood models) where the next piece emerges from a mathematical operation independent of its significance. In both cases, selection and placement is autonomous; lexical collage is and can only be nonsemic because the materials being arranged, even by large language models such as ChatGPT, are employed without concern for the meaning of the terms being arranged in the output. These techniques produce a chimera for intentional action and communication: nonsemic ‘expressions’ are parasitic on human intentions and encoding in the materials reconfigured in/by the creation of these expressions, and which simultaneously exploit/depend upon their audience’s capacities to integrate disparate terms and materials to form a coherent poeisis—the differences, illegibilities, and ruptures between parts becoming expressive bridge the disparactions to produce a new synthesis. The lineage of the avant-garde’s work with nonsemic poetry justifies addressing the products of AI generation as a potentially expressive despite its clearly unencoded foundation. They are a magic mirror onto human aesthetics and interpretations, an engagement with a socially coded, but unconscious structure of meaning and form that centers the audience’s apprehension.
Ambivalence problematizes all expressions (semic, asemic, nonsemic) because human audiences respond to semantic cues without concern for their ontology or origin.[xvii] The decoder’s interpretation is independent of any significance that might be assigned to the utterance,[xviii] allowing semiosis to proceed even if the “coding” is nebulous, a reflection of how ambivalence must be resolved for the symbolic identification linking signs to signifieds and producing significance to begin. Semiosis is independent from, but entangled with, the diagnostic recognition that provides the category assignment as-language. This split between the identification of ‘what things appear to be’ and ‘what they communicate’ is instrumentalized by AI, which only addresses forms and ignores their meanings. These results of technological ingenuity produce a structural monotony precisely because the machine is constrained by the data used in its operations—the banality and cliché generated by AI’s nonsemic expressions is a demonstration of their nonencoding. Because these machines proceed without consideration of significance, their outputs tend towards the statistically most likely, creating variations on the familiar—the necessary and sufficient qualia for their structural monotony. The ‘novelty’ of AI is not a matter of invention, but of permutation. Machine learning does not challenge conventionality precisely because it operates without concern for meaning, a rote process instrumentalizing convention as production.[xix]
The role of human symbolic interpretation in shaping poeisis (and even the poetics based in the products of AI) becomes self-evident when addressing ambivalent works such as the visual poem Two Degrees of Face [Figure 3]. Resolving uncertain perceptions into coherence requires a continual shifting between distinct (but related) modes of interpretation that organize the empirical encounter. Category assignments are foundational and immediate[xx]—they pass entirely unconsciously, yet ambivalence returns their formative role to consciousness. To change from just seeing abstract patterns to looking at writing begins with the identification as-language that shapes how the next diagnostic recognition of particular signs, letters, or words proceeds.[xxi] Past experience and established knowledge thus modify the human audience’s experiential encounter via apperception—establishing enculturation as an inherent constraint for semiosis because it is always limited by the reader’s proximate awareness of what matters in perception intersecting with their subjective interpretive skill.[xxii] Choosing which paradigm (reading or seeing) to interpret the ambivalent marks in Figures 1 and 3 depends on the audience’s desires that shape how they understand what they encounter: the semantic cues in Figure 1 never resolve into coherent signs, while the ones in Figure 3 readily becomes legible. Semiosis begins with this foundational refraction differentiating reading from seeing via implicit, immediate, and often unacknowledged decisions about category assignment as-image or as-language. This prima causa determines the protocols for higher level signification via the “gignomenological law of identity,” an intellectual finger trap which constrains what apperceptions are appropriate to consider, while rejecting all others as irrelevant noise.[xxiii] The paradigm shift between recognizing Figure 3 as a visual depiction of ‘face in profile’ and/or as a text stating ‘two degrees’ demonstrates the power of this foundational desire. The choice to shift between reading the marks lexically as letterforms, or seeing them graphically as a face, becomes a poetic act familiar from considerations of Asemic poetry. When viewed as-language, Figure 3 allows the unambivalent identification of the marks as ‘typography’: the number /2/ and the symbol /°/ for degree, accompanied by an underscore character. How this text states ‘two degrees’ is so fully legible and familiar that some viewers may only see typography; however, it can also be understood as an image. The ‘face in profile’ recognition returns it to an graphic paradigm of depiction, and responds to a subjective, internal choice by the audience to shift the foundations of their engagement. That change reveals the ambivalence of the marks themselves.
Nonsemic poetics exploit what semiotician Hartmut Stöckl noted about the distinction between the recognitions of text and the expressive significance that typography has. This ambivalence anticipates how the AImage separates lexicality (denotation) from the parallel diagnostics of depiction, suggesting direct engagements with the act of perception itself:
Typefaces may point to the nature of the document, carry emotional values or indicate the writer’s intended audience, and aspects of the layout may serve to reinforce the thematic structure of a given text and facilitate access to its information. Finally, on yet another level of typographic meaning making, the graphic signs of writing can assume pictorial qualities. Thus, letters may form visual shapes which stand for objects from reality, signal states-of-affairs or actions, and illustrate emotions. Materials and techniques of graphic sign making, too, may be made salient in text design and can thus convey something about situation, genre and stylistic intent of a communicative occurrence—this is also a pictorial kind of communication. It is this threefold semiotic nature of typography that provides its communicative flexibility.[xxiv]
The scope of nonlexical significance Stöckl identifies draws attention to how design parallels the more common and familiar interpretive activity of reading text. But these expressive dimensions of typography are equally absent from consideration by semiotics and art history, an omission that has not received systematic consideration,[xxv] even though the psychological processes of letterform identification and legibility have been extensively studied.[xxvi] The “pictorial qualities” he describes—the depiction of typography and design—are readily apparent in the various traditions of calligraphy as well as in the wide range of visual poetry that includes calligrams, concrete poems, and asemic writing. All these varieties of poeisis exploit the graphic aspects of writing.[xxvii]
These interpretive dynamics are not questions of text and lexicality, but emerge from ambivalent visual perception. The initial decision that identifies language orders perception and elides attention to its depiction, thus rendering the ‘face in profile’ in Two Degrees of Face as a pure excess, irrelevant to lexical semiosis. This capacity of depiction to inflect, deflect, and modify meaning is precisely poetic. It reveals how the implicit identification of writing can reveal autonomous and unconscious interactions between perception, enculturation, and apperception. Any diagnostic recognition that ignores the designation as-typography transforms these marks into an image-object distant from the coded signification chain that defines judgments of lexical form: [A] the legibility of arrangement and marks (semantic cues), [B] the reader’s anticipation of what comes next (subjective encounter) that parses the sequence, and which enables [C] their lexical fluency (past experience with the ‘intentional function’) to organize both signs and signifieds into coherence. This whole process responds to their desire for comprehension. Each of [A, B, and C] teleologically support the others in a self-referential tautology; they are not a hierarchy, but a network or web of mutually validating relations that enables the existence of a series of mutually exclusive (and intermediate) interpretations authorized by their place within the range of interpretations[xxviii]—a state of information that AI instrumentalizes[xxix] by exploiting how [A and B] are empirically present features that audiences rely upon to make interpretive “bets” [C] appealing to what in perception is ready-to-signify.[xxx]
Lexical recognition derives from the human audience resolving the ambivalence[xxxi] exploited by metamorphic illusions: the AImages of a landscape which also states ‘READ THE TEXT’ in Figure 4 reveal the instabilities of [A and B] arise in a mental paradox linked to the complexity of perception.[xxxii] This ‘text’ creates the same shifting relationship between reading text and seeing image evident in Figure 3. These ‘subliminal texts’ were produced using AI which neither recognizes nor understands the words: understanding the graphic outlines and shapes as-language [C] is culturally defined, not just by the capacity to read, but in the more basic identification of the patterns as-being language at all.[xxxiii] Producing these AImages requires human mediation via a ‘control net’ image that already contains the text which the machine simply uses as a compositional guide to generate a visual output. This technical role is obvious on examination: although the words ‘READ THE TEXT’ appear as a demonstrable “fact,” they are not actually present. To become legible, the audience must mentally fill-in the missing parts of words and letterforms [A], linking the fragmentary elements to form something familiar [C]. Lexicality is justified by how the audience makes the interpretive bet[xxxiv] that facilitates it via a tautology masking the imaginative ‘filling-in’ of missing details [B] which reinforces lexical identification, but which also reveals it is illusory. This capacity to ‘complete’ letterforms when confronted with fragmentary elements is what causes the lettering to both appear and then disappear on inspection in the same ways that the AImage it itself will decompose into abstract marks on close examination. Depending on the weight assigned to the ‘control net,’ the facility of this recognition can be modulated to be more or less legible as the differences between the top and bottom variants in Figure 4 show.
And although both reading and seeing equally originate in perception, for language and its recognitions, enculturation dominates. A foundational entanglement between fluency and perception defines the refractive process confronting language. Both AImages in Figure 4 demonstrate how category assignments are constructs, rather than being an innate, ontological qualia of the objects thus engaged. Contingency dominates via implicit refractions that recognize, complete, and interpret perceptions as-text derive from encultured knowledge gleaned from past experiences with myriad other texts and images whose ambivalence is resolved or minimized, and for which the appropriateness of semic interpretation has been established. The ‘intentional function’ identifies the mediation of this expertise [C]. Changes in initial assumptions about component parts [A] cause modal shifts that restructure ambivalence into a different interpretation.[xxxv] The recognition of familiar language in Figure 4 authenticates how past experience completes partial letterforms into a legible rendition of each word,[xxxvi] even when parts of the text are obscured or missing.[xxxvii] This discovery of ‘meaning’ is an example of what gestalt psychologist Erich von Hornbostel described in his essay “The Unity of the Senses.”[xxxviii] Interpreting ambivalence is autonomous, suggesting a bridge between the conceptual differences separating semiotic analysis from what psychology experimentally examines, anticipating their relevance for the nonsemic:
The perceptible is not less perceptible because it is more than merely perceptible. Appearance are not only a means by which we get knowledge of something—not otherwise communicable—which stands behind, beside, and beyond. It is not hidden behind the appearance, but is beheld directly therein. We do not hear sounds which someone once put together in such and such a matter in order to express this or that—we hear Mozart.[xxxix]
Immanent–emergent affinities parse ‘form and enunciation’ from experience, the predictability granted by familiar forms and structures being foundational to the recognition of semantic cues and their deployment in nonsemic works.[xl] The mutually exclusive decoding paradigms (reading versus seeing) demonstrate the centrality of past experience to the realm of interpretive activity. Embracing the ambivalence at the perceptual foundation of semiosis expands the scope of poeisis dramatically beyond only those unquestionably encoded expressions to include the nonsemic products of AI generated via a complex process of autonomous collage.
While identifying letterforms appears to be a perceptual act comparable to seeing color, an immanent encounter independent of cultural knowledge—but this fallacy ignores how the ‘recognition’ of language [C] depends on the viewer. These problematics of interpreting writing and language were commonly masked in the twentieth century because the Modernist demand[xli] for maximal legibility required a refusal of the graphic aspects of writing and presentation in favor of a metaphoric “transparency” where typography seems to vanish, leaving only the meaning communicated,[xlii] and transcends the nature of the encounter: a photograph of typography rarely falls into the same category as other varieties of photography such as the holiday snapshot, the journalistic news image, or even the selfie. Its contents are identified as-typography, and the vehicle of its presentation—the photograph—is irrelevant to that designation. The same problem confronting Jasper John’s paintings of numbers or flags: the difference between depiction and denotation collapses.[xliii] These seemingly simple identifications mask a structural complexity of hidden decisions which are central to the stability of the knowledge written language communicates.[xliv] While it is possible to make a photograph of a book where you can see the words, as that image coincides with the shape and extent of the page photographed, a curious transformation happens: it stops being a photograph of a page in a book and becomes that page. Photolithography and photographic typesetting both depend on this coincidence of image-as-page. Although it is always possible to read the words in a photograph of a book page, (and thus to treat the image as the page of text), this potential regard only serves to make the difference of lexical and pictorial more obvious. Text retains its notational identity no matter how it appears, while images do not. The shift in recognition makes the distinctions between pictorial images and language evident. “Encoded meaning” begins with an examination for empirically present cues (i.e. the identification of ‘writing’) whose identities as a residue of intentional action must be learned. This diagnostic for the potentially communicative is mediated by the ‘intentional function’ which justifies interpreting the semantic cues as-if encoded, thus beginning the sign formation process and its decoding of expressive articulation.[xlv]
Human apperception will invent the details and connections required for semiosis from incomplete and fragmentary elements, a necessary adaptation to ambivalent experiences. These completions reinforce how signification is not an inherent property of anything encountered, but must be imposed through a process of un/conscious ordering that follows the predictable and precisely conventional protocols that are exploited in metamorphic images [Figure 4] and other optical illusions such as the Kanizsa triangle[xlvi] [Figure 5, Left]. The plastic recognition of triangles in Figure 5 parallels the lexical recognitions in Figure 3 that requires a completion of the letterforms to become visible. Although the processes of modal [Figure 5, Center] and amodal [Figure 5, Right] completion[xlvii] are different than the completion of the metamorphic letterforms in Figure 4, it demonstrates that not all the elements of semiosis necessarily exist in the empirical perception of a world of facts, as a fact among facts. The ‘triangles’ in Figure 4 are simultaneously empirically absent, yet paradoxically present,[xlviii] a product of past experiences creating phantom edges[xlix]—mentally ‘filled in’ forms—that produce a familiar image.
These recognition and completion processes shape the Asemic poem in Figure 6, which presents the human audience with fragmentary typography that is easily identified as-language, despite how reading that lettering is impossible, as in Figure 1 which presents the semantic cues of handwriting without containing handwriting. Fragmentary marks provide enable the letterforms’ identification, rendering them as-language, but withholding their legibility.[l] The refraction that changes pictorial form into written language sets in motion a series of decisions that become visible in/as these “letterforms,”[li] even though the letters are imaginary: individual suggestions of familiar typography disappear upon direct examination.
Ambivalence is central to these nonsemic poetics, a point of continuity between earlier avant-garde[lii] uses and contemporary AI: the nonsemic text in both Figures 1 and 4 (made by AI), and the collage poem in Figure 2 or the asemic poem in Figure 6 (human made), rely on the viewers’ capacity for ‘word recognition’ to read fragmentary words, but this capacity also describes the emergence of individual letterforms. Figure 6 thus splinters in precisely the same ways that the depictions in Figure 4 decompose into abstract blocks and smears on close inspection, or the triangles in Figure 5 vanish, only to return as phantoms: apperception integrates the role of memory and past experience with perception. This identification of elements moves from opposite ends of the same range to understanding semantic cues. Apperception sets in motion a series of recognitions that exceed the scope of what is empirically present. The recognition of familiar lexicality creates a coherent meaning even when aspects of the letterforms themselves are ambivalent [Figures 1 and 4], their arrangement nonsensical [Figure 2], or missing [Figure 6]. Poeisis exploits how the semantic cues empirically present in the thing perceived can become a source of ambivalence—demonstrating the constitutive, atomic particles of expression by creating, containing, or manipulating language without producing definite or legible statements.
This insistence on image-as-language is the secret of perception revealed by ligatures in general—that all three instances shown in Figure 7 present a complete letterform that shares one or more elements with the adjacent characters, linking them together to produce a coherent figure. Parsing these connections between and around fragmentary letters resolves their ambivalence into familiar coherence where all elements of each letterform are present, but ambivalent, shared across pseudo-independent parts that combine several letters into a compound character in both handwriting [Figure 7, left] and the established fonts of typography [Figure 7, center]. The combination of both potential approaches—standardized ligatures with the more dynamic linkages common of handwriting—demonstrates the potentials of design [Figure 7, right] to exploit the graphic fusion of letterforms. It produces a range of relationships that increasingly segments the components while linking them together. Irregular ligatures [Figure 7, left] such as the ‘handwritten’ letters forming the emergent word ‘why’ share elements; traditional typographic ligatures [Figure 7, center] such as ‘ae’ and ‘oe’ where paired vowels are conjoined but remain distinct; the abstract block [Figure 7, right] that creates a singular unit that fuses all the letterforms into a graphic symbol which nevertheless maintains its visual identity as the discrete word “book.”
Acknowledging these entangled roles of perception, ambivalence, and the instability of semiosis when confronting the products of AI generative systems challenges the commonplace assumption that these machines will render human creative labor obsolete. Instead, they emphasize the critical act of semiosis by undermining assumptions about the necessity for human virtuosity and differentials in execution of finished works. This change in address leaves the aesthetic problem of poeisis intact as a product of how the human audience’s fluency, expertise, and knowledge are guided by their desires for comprehension. Instead of an expressive untergang, the nonsemic poetics of AI elevates the subtle and infinitesimal differences to greater prominence and aesthetic importance.[liii] The capacity of generative systems to automate a permutation and variation in the arrangement of semantic cues without concern for their role in signification is not a new aesthetic potential, but is well established in the historical avant-garde’s uses of collage to challenge coherence and authorial intent. Ironically, this aesthetic lineage, instead of displacing critical analyses, makes reflective judgments[liv] and the connoisseurship they require into the central concern for poeisis.
Figure 1: AImage of nonsemic poem where the “marks” suggest language, but are not. Copyright © 2022 Michael Betancourt / courtesy Artists Rights Society (ARS).
Figure 2: Nonsemic Surrealist collage poem, “Tu seras parmi les victims” (“You will be among the victims”) [left] French text including original typeface choices [right] English translation Surrealist collage poem by André Breton, ~1924.
Figure 3:
Two Degrees of Face is a visual poem that can be understood as either a text stating ‘two degrees’ or as an image showing a ‘face in profile.’
Copyright © 2020 Michael Betancourt / courtesy Artists Rights Society (ARS).
Figure 4: Two examples of a metamorphic AImage showing a landscape where the ‘trees’ are also ‘subliminal’ words stating READ THE TEXT. Copyright © 2024 Michael Betancourt / courtesy Artists Rights Society (ARS).
Figure 5: [left] the full Kanizsa triangle [center] example of modal completion [right] example of amodal completion The recognition and relationship between empirically present elements and the perception of the ‘triangles’ these forms enable depends on human perception.
Figure 6:
Asemic typoem 2024-008 presenting ambivalent suggestions of letterforms.
Copyright © 2023 Michael Betancourt / courtesy Artists Rights Society (ARS).
Figure 7:
Three examples of ligatures:
[left] in handwriting
[center] in traditional typography
[right] in abstract design
Copyright © 2015 Michael Betancourt / courtesy Artists Rights Society (ARS).
[i] Drucker, J. “What is a Word’s Body?” What is? Nine Epistemological Essays (Victoria: Cuneiform Press, 2013) p. 38.
[ii] Borges, J. “From Allegory to Novel” Other Inquisitions (Austin: University of Texas Press, 1975) p. 155.
[iii] Vasillakis, N. These Gem Effluvia (Redfox Press, 2024).
[iv] Eco, U. A Theory of Semiotics (Bloomington: Indiana University Press, 1979) p. 221.
[v] Garrison, M. “The Poetics of Ambivalence” Archetypal Psychiatry (Spring 1982) pp. 213-232.
[vi] Milić, N. “Resistance to Nihilism: Benjamin, Baudrillard, and Duchamp” AM Journal no. 22 (2020) pp. 15−22 DOI: http://dx.doi.org/10.25038/am.v0i22.377
[vii] Betancourt, M. Typography and Motion Graphics: The ‘Reading-Image’ (Routledge Studies in Media Theory and Practice) (Routledge, 2018) pp. 11-18.
[viii] Betancourt, M. “Asemic Reading, Visual Writing: an introduction to ‘asemic poetry’ as metalanguage” Psiax Pintor: Estudos e Reflexões sobre Desenho e Imagem, no. 6 (“Writing on, for and with Drawing”), published by i2ADS, (Faculties of Architecture and Fine Arts of the University of Porto and the School of Architecture, Art and Design of the University of Minho), Porto, Portugal (December 2022) pp. 33-38.
[ix] Mould, O. Against Creativity (New York: Verso, 2018).
[x] Schiller, D. Digital Capitalism: Networking the Global Market System (New Haven: MIT Press, 2000).
[xi] Betancourt, M. Glitch Theory: Art and Semiotics (I’m Press’d, 2023) pp. 13-15.
[xii]Knapp, S. and Michaels, W.B. “Against Theory” Critical Inquiry vol. 8, no. 4. (Summer, 1982) pp. 717-728.
[xiii]Jakobson, R. Selected Writings III: Poetry of Grammar and Grammar of Poetry (New York: Mouton Publishers, 1981) p. 25.
[xiv] Betancourt, M. “A Note on AI and the Ideology of Creativity” Academia Letters, Article 328 (March, 2021) DOI: https://doi.org/10.20935/AL328
[xv] Hausmann, R. “Optophonemes (1962)” Ubuweb DOI: https://ubu.com/sound/hausmann.html retrieved July 26, 2024
[xvi] Attali, J. Noise: The Political Economy of Music (Minneapolis: University of Minnesota Press, 1985).
[xvii] Eco, U. Interpretation and Overinterpretation ed. Stefan Collini (Cambridge: Cambridge University Press, 1992).
[xviii] Eco, U. “The Poetics of the Open Work” Participation: Documents of Contemporary Art ed. Claire Bishop (Cambridge: MIT Press, 2006) pp. 20-40.
[xix] Munn, L. Automation is a Myth (Stanford: Stanford University Press, 2022) pp. 29-45.
[xx] Schyns, P. G. “Categories and Percepts: A Bi-Directional Framework for Categorization” Trends in Cognitive Sciences vol. 1, no. 5 (August, 1997) pp. 183-189.
[xxi] Jastrow, J. Fact and Fable in Psychology (London: Houghton, Mifflin and Co., 1901) pp. 280-281.
[xxii] Harris, R. Introduction to Integrational Linguistics (Oxford: Pergamon 1998) p. 105.
[xxiii] Ziehen, T. Lehrbuch der Logik auf positivistischer Grundlage mit Berücksichtigung der Geschichte der Logik (Bonn: A. Marcus & E. Webere Verlag, 1920).
[xxiv] Stöckl, H. “Typography: body and dress of a text” Visual Communication 4.2 (June, 2005) p. 78.
[xxv] Drucker, J. “What is a Word’s Body?” What is? Nine Epistemological Essays (Victoria: Cuneform Press, 2013) pp. 33-46.
[xxvi] Haslerud, George M., and Clark, R. Ernest. “On the Redintegrative Perception of Words,” The American Journal of Psychology, vol. 70, no. 1 (1957) DOI: 10.2307/1419236
[xxvii] Betancourt, M. Typography and Motion Graphics: The ‘Reading-Image’ (New York: Routledge, 2018) pp. 68-101.
[xxviii] Eco, Op. cit. p. 97-98.
[xxix] Betancourt, M. The Critique of Digital Capitalism (Brooklyn: Punctum Books, 2016) pp. 153-190.
[xxx] Harris, R. “The Integrational Conception of the Sign” Integrationist Notes and Papers 2006–2008 (Gamlingay: Bright Pen, 2009).
[xxxi] Bleuler, Eugen. “Theory of Schizophrenic Negativism,” translated by William A. White, Journal of Nervous and Mental Disease, 1950, p. 266.
[xxxii] Caws, Mary-Ann. The Surrealist Painters and Poets: An Anthology (Cambridge: The MIT Press, 2001), pp. 179.
[xxxiii] Solomon, R. L., and Postman, L. “Frequency of Usage as a Determinant of Recognition Threshold for Words,” Journal of Experimental Psychology vol. 43, no. 3, (1952) pp. 195-201 DOI: 10.1037/h0054636
[xxxiv] Harris, R. Integrating Reality (London: New Generation Publishing, 2012).
[xxxv] Bleuler, Op. cit., p. 477.
[xxxvi] Bouma, H. “Interaction Effects in Parafoveal Letter Recognition,” Nature no. 226, (1970) pp. 177–178. DOI: 10.1038/226177a0
[xxxvii] Haslerud, George M., and Clark, R. Ernest. “On the Redintegrative Perception of Words,” The American Journal of Psychology, vol. 70, no. 1 (1957) pp. 100–101 DOI: 10.2307/1419236
[xxxviii] von Hornbostel, E. “The Unity of the Senses,” A Source Book of Gestalt Psychology ed. Willis D. Ellis (Gouldsboro: Gestalt Journal Press, 1997) p. 214.
[xxxix] von Hornbostel, E. “The Unity of the Senses,” A Source Book of Gestalt Psychology ed. Willis D. Ellis (Gouldsboro: Gestalt Journal Press, 1997) p. 214.
[xl] Harris, R. Integrating Reality (London: New Generation Publishing, 2012).
[xli] McMurtrie, D. “The Philosophy of Modernism in Typography” Texts on Type: Critical Writing on Typography ed. Steven Heller and Philip B. Meggs (New York: Allworth Press, 2001) pp. 146-148.
[xlii] Warde B. The Crystal Goblet: Sixteen Essays on Typography (London: Sylvan Press, 1955).
[xliii] Danto, A. The Transfiguration of the Commonplace (Cambridge: Harvard University Press, 1981) pp 83-86..
[xliv] Foucault, M. The Archaeology of Knowledge (New York: Pantheon, 1972).
[xlv] Betancourt, M. “The ‘Intentional Function’ in Still and Moving Photographic Images” Semiotica: Journal of the International Association for Semiotic Studies, vol. 246 (2022) DOI: https://doi.org/10.1515/sem-2020-0065
[xlvi] Kanizsa, G. (1955). “Marginiquasi-percettivi in campicon stimulazione omogenea,” Rivistadi Psicologia, vol. 49, no. 1, pp. 7-30.
[xlvii] Monk, A. F. and Hulme, C. “Errors in Proofreading: Evidence for use of word shape in word recognition” Memory & Cognition vol. 11, no. 1 (1983) pp. 16-17.
[xlviii] Kanizsa, G. (1974). “Contours without gradients or cognitive contours?” Giornale Italiano di Psicologia, vol. 1, no. 1, pp. 93–113.
[xlix] Gregory, R. L. The Intelligent Eye (New York: McGraw-Hill, 1970) p. 20.
[l] Bouma, H. “Visual Recognition of Isolated Lower Case Letters,” Vision Research vol. 11, no. 5 (May, 1971) pp. 459-474. DOI: 10.1016/0042-6989(71)90087-3
[li] Solomon, R. L., and Postman, L. “Frequency of Usage as a Determinant of Recognition Threshold for Words,” Journal of Experimental Psychology vol. 43, no. 3, (1952) pp. 195-201 DOI: 10.1037/h0054636
[lii] Poggioli, Renato. The Theory of the Avant-garde, (Cambridge: Harvard University Press, 1968).
[liii] Betancourt, M. Art, AI and Culture (Savannah: I’m Pressed, 2022) pp. 69-96.
[liv] Kant, I. The Critique of Pure Reason trans. Werner Pluhar (Indianapolis: Hackett Publishing Company, 1996) p. 675.
