What Writing Looks Like

by Stephen Houston (Brown University)

Beginning as puffs of air, channeled and shaped by the throat and mouth, words travel out from the body to reach other human ears. After cognitive processing, the puffs release their message, and communication ensues. [1] But words create their own problems. How is an assortment of meaningful exhalations, clicks, articulations, bellows, flutings, and affrications made more permanent and their recollection preserved beyond the memory of speakers and listeners? As many have observed, that is exactly what writing does. It takes ephemeral and invisible words and transforms them into fixed and visible graphs, to be seen as much as any picture. [2]

The ability to picture language creates its own kinds of play. Other graphic possibilities present themselves, other ways of linking with images. Other sorts of information become available. The claim that writing only concerns a phonic or linguistic message is a partial understanding at best, misleading at worst. Frolics with graphs, a luxuriation in their visible, material nature—these can be as important as any representation of sound. For Classicists, there is a relevant scene painted by Douris in Athens, at c. 490–485 BC (Fig. 1). In it, a schoolmaster holds a partly opened scroll, whose text reads: MOIΣAMOI AΦIΣKAMANΔPON EYPΩNAPXOMAI AEINΔEN. Translations of this sentence seem to vary by the translator, but it concerns a Homeric appeal to a muse and a reference to a good place for singing by the banks of the fast-flowing Scamander (Skamandros, the modern Karamenderes River in Turkey).



Figure 1. A schoolroom scene by the painter Douris, red-figure kylix, c. 490–485 BC, Athens, (Berlin, Antikensammlung, Staatliche Museen 2285).


One theory suggests that we are looking at a bemused schoolmaster and a botched text from an “F” student: a subtle joke about dullards (Sider 2010:548). A representation (a painting by Douris of a scroll and a schoolmaster) embeds a second representation (a record of sound and meaning in an addled text). But the eye darts between the two levels. It reads the text, yet it also depicts those phrases as something physical, an inking on papyrus that opens up within a picture. In other instances, such as a vignette in an illuminated manuscript from c. AD 1450–1475, there can be a mind-bending mix: a representation of a representation of a representation (Fig. 2, Houston 2018b). An image of a northern Italian apothecary’s shop shows jars rimmed with pseudo-Hebrew or pseudo-Kufic characters, the latter a simulation—a representation—of legible writing.

Figure 2. Ibn Sina/Avicenna, Canon Medicinae, Bibliotheca Universitaria, Bologna, Italy, MS 2197, fol. 492.


Some pictured texts come close to trompe l’oeil, that clever trick by which the viewer or reader is led to confuse and blur materials (Fig. 3). In this way, a two-dimensional image triggers the perception of a three-dimensional object (Houston 2014:61, 62, 147fn.40). Miriam Milman (2009:22–23) explains how to activate the ruse: make the object as close as possible in size to the original it replicates; blend it into surroundings; limit depth; avoid live subjects that move; and create edges that do not compromise the deception. As one case of many, a painter, perhaps Ludger tom Ring the younger (1522–1583), created an open missal (a book for saying mass) that offers a tantalizing glimpse of a gilded page (likely a Crucifixion), surrounded by columbine, insects, corn flowers, and musical notation. The pages flutter slightly, about to be consulted, and a leather strap marks the first passage that is about to be read (Loeb Open Missal). There must have been some market for these ingenious deceptions, for a nearly identical painting is in the Galleria degli Uffizi in Florence (N. Cat. 00124048, Inv. 1890, 6191). Other than a possible signature under the music (“Ludevi rinki”) no part of the text is readable. The work itself may have been an amusing surprise that lay on a sloping stand in a bookseller’s shop (Stirling 1952:33). Glossy and expensive, it hinted at knowledge that could never be accessed.



Figure 3. The Open Missal, attributed to Ludger tom Ring the younger, c. AD 1570, Frances Lehman Loeb Art Center, Vassar College, 1956.5).


A later painting, by Laurent de La Hyre (1606–1656), also displays a text as though in three-dimensional space (Fig. 4). A literate audience was the intended target, one that would recognize the figure as a liberal art (“Grammar”), watering a plant that is out-of-scene—thirsty growth stands for young minds. On its ticker tape, there reads in Latin: “a meaningful utterance which can be written down, pronounced in the proper way.” The whole was inspired by an illustrated book, Iconologia, 1603, by Cesare Ripa (Wine et al. 1993:23–25). It formed part of a larger set of seven half-length panels extolling each of the liberal arts. A bookish audience, smug its own accomplishments, would have appreciated the painting and wanted its message multiply among the young. The letters seem to move in real space. They distort, and some of the letters disappear in part. The back of the text occurs too, the letters washed out in a brown-tinged reversal. An artful ploy simulates what the eyes might actually see in a hand-held scroll.



Figure 4. Allegory of Grammar, Laurent de La Hyre, 1650 (National Gallery of Art, London, NG6329, photograph by Stephen Houston).


The Classic Maya showed writing in the same way: as representations of representations, on physical objects in pictorial space. [3] For example, most Maya books are shown, as first suggested by Robert Sonin and amply documented by Michael Coe, in the form of leporellos or screenfolds (Coe 1973:91; Coe 1977:332–33, figs. 4–7). A few are unopened or about to be read (Fig. 5).



Figure 5. Opossum scribe (K’IN-ni ya-sa u-chu) with Maya codex and vulture accountant (k’a?-na u-su) holding single sheet with numbers, perhaps a mythic Long Count date of (BAMW Photography). 


Others are folded up tidily, two pages viewable at a time (Fig. 6). A curious feature, not often noted, is that the books are being examined or painted in an impossible manner. The scribe sits perpendicular to the correct position for writing, for the folds are always vertical in a book, not horizontal as shown here. Doubtless this was for clarity of presentation. A scribe in front of a book would obscure it to the viewer.



Figure 6. Trickster rabbit-scribe, northern Guatemala/southern Campeche, c. AD 725 (K511, Princeton University Art Museum, y1975–17, photograph © Justin Kerr, used with permission).


Another feature is that, with one exception, such pictured books never disclose their contents. Viewers can readily identify a codex by its sumptuous jaguar-pelt covering and the thin, smoothed excellence of its page-edges (usually 4 to 10 visible, i.e., rather terse works by the standards of surviving examples). But they are not given any view of the glyphs within. The exception is late, a vessel from the final decades of the Classic period (a vase by the same artist may be found in the Museo Popol Vuh, Guatemala, #5335, Chinchilla Mazariegos 2005). It shows a mythic tableau of animals bringing offerings of food and drink that are presumably being tallied in an open book by two monkey scribes (Fig. 7). The deity receiving these treats may be a high god known to specialists as “God D,” but with unusual touches, for he is borne aloft by a coiled snake (on the combinatory complexity of this character, see Martin 2015:214–215, fig. 37). Unexpectedly, the book shows, at slight angle, in awkward display, some bars, cross-banded signs, and a few dots. These offer a casual hint of content, rapid flicks of ink to suggest writing, but not its detail. On present evidence, all such scenes are mythic, the participants gods or supernaturals. Not a one appears to be dynastic. Indeed, historical images are decidedly phobic about depicting books, despite the undoubted presence of many such tomes in Maya cities (a lone dynastic image may include a codex, but, oddly, it serves only as a support for the mirror of a preening lord, K6341).



Figure 7. Monkey scribes, scene of food tribute or serving, Peten, Guatemala, c. AD 800 (K3413, photograph © Justin Kerr, used with permission).


Far more evident are glyphs on depictions of ceramics (Figs. 8, 9, 10). They appear where they should, as rim bands, but largely as pseudo-script, ovoids with thickened outlines and interior features in more delicate, thinner lines (Houston 2018b). They offer a graphic primer of what Maya scribes thought the formal attributes of writing should be.



Figure 8. Vessel with (pseudo-)glyphs, c. AD 700 (K2800, photograph © Justin Kerr, used with permission).

The tributary scene mentioned before revels in such labels on ceramics (Fig. 9). Each animal—as a whole, they constitute a near-complete typology of Maya mammals and quadrupeds—offers up a drinking vessel with prominent glyphs on the side visible to the viewer. The lucid presentation seems not to involve legibility, however, for they appear to repeat pseudo-glyphs (a large sign with appended suffixes) that resemble the glyph for “sky,” ka’n. The scribe, a painter with a hand for inventive scenes roiling with energy, was probably someone with only a light grip on glyphic literacy. His two works demonstrate a familiarity with a few signs and their customary arrangement as suffixes and larger glyphs, but he had little understanding beyond graphic display. His writing was pure picture.



Figure 9. Animals serving food, Peten, Guatemala, c. AD 800 (K3413, photograph © Justin Kerr, used with permission).


The tenuous line between legibility and pseudo-writing is less a necessity than a strategy for other scribes, as in the fully literate Akan Suutz’, a painter of a vessel now in the Los Angeles County Museum of Art (Fig. 10, M.2010.115.12, see also K1599). The main text on the vessel is legible, even bold and confident. This is someone who understood, as do illustrators today, the impact of the la ligne claire (Clear Line; Ligne Claire). Small vessels throughout the scene have glyphs that appear to repeat, if with the usual alternation or juxtaposition of “affixes” and larger signs. Yet there is also an expert execution of a “12 Ajaw” on a jar for pulque. That may correspond to a date of, in the Maya Long Count system, (an ending for a 20-year span often commemorated with Ajaw signs written in this way, without months), or, in the Western calendar, a Julian Date of Jan. 21, AD 771.



Figure 10. Polychrome vessel from area of Tikal or even Aguateca or Dos Pilas, but likely made near Lake Peten Itza, Guatemala (Los Angeles County Museum of Art, M.2010.115.12, see also K1599). 


The lively scenes of marketing found in Structure Sub 1-4 at Calakmul, Mexico, offer both examples of glyphs painted on textiles (a possible u chu-?, u chuy, “sewing”?), but also, in another panel, a cup lifted to the lips of an atole drinker (Fig. 11, Carrasco Vargas and Cordeiro Baqueiro 2012:fig. 8, close-up fig. 33; Martin 2012:64–65, fig. 6). A different technique intruded here, “a minutely incised inscription” with yu-li (Martin 2012:64) that may refer to atole or maize-drink, ordinarily spelled ul, or to the act of carving or incision itself, yul-il (Houston 2016:424–425, fn9). Post-fire texts do not occur Late Classic pottery, especially in such a prominent position, but, with this enhancement, the legible text evoked the direct action and presence of a scribe.


incised CLK.jpg

Figure 11. Glyphs on blue-painted atole bowl, Structure Sub 1-4, Calakmul, Mexico (Carrasco Vargas and Cordeiro Baqueiro 2012:close-up fig. 33, photograph by Rogelio Valencia Rivera, Proyecto Arqueológico Calakmul). 


Glyphs on textiles afford an insight about gender. By common belief—the assertion is plausible yet hard to prove—most weavings were done by women (Halperin 2016:435). Yet there is also overwhelming evidence that the scribes and literate sculptors were men (Houston 2016). The occurrence of pseudo-writing on some textiles (Fig. 12, left), but legible texts on others (Fig, 12, right, Laporte and Fialko 1995:82, fig. 69), may have several explanations. If an actual textile is being shown, then this may reveal variable literacy among those painting textiles. Note that few appear to be woven into the fabric, i.e., they were added later. Or, if the painter of the pot is the relevant party, then it simply speaks to their representation of textiles.



Figure 12. Polychrome vessels with glyphs on textiles: (left) sash around waist, with pseudo-writing (Los Angeles County Museum of Art, M.2010.115.12, see also K1599); and (right) vase from Tikal, Burial PNT-007, with seemingly legible signs (K2697, photograph © Justin Kerr, used with permission).


The glyphs on a vessel from Tikal inspire confidence that literate productions appeared in some of these images, including a possible reference to a male youth (Fig. 13ch’o-ko? CHAK-la-ya ‘a?).



Figure 13. Close-up of glyphs, vase from Tikal, Burial PNT-007, with seemingly legible signs (K2697, photograph © Justin Kerr, used with permission).

A conundrum for any person looking at ancient art is that divide, at times close, at times yawning, between depiction and the depicted. These are no snapshots. They express a considered view of what to show and how to show it. But the occlusions, partly visible in several images (Fig. 12, leftFigs. 13, 14, 15), along with Laurent de La Hyre’s, Allegory of Grammar (Fig. 4), reinforce a view that an ocular effect is being entertained here, that painters and carvers are displaying not what they know to be there but what they can see (Houston 2016:fig. 13.5). Occasionally, glyphs are obscured by another piece of cloth or ornament (see also a partial sculptor’s name, in the Princeton University Art Museum, #2012–78, Houston 2016:fig. 12.5, in a lead from Bryan Just). The glyphs painted at the end of Classic period in the Bonampak murals refer explicitly to “cloth” in one case (u bu ku), but to secondary painting in another (u tz’i ba-li), to the medium of transmission, line-like paint applied after weaving, and to the intended display surface. These probably operated in a setting of tributary offering (hence the T’AB-yi, “raise up,” in Fig. 14, Room 1, Caption 5c; Houston 2018a:152). Texts specified that someone painted them, that they were offered, and that the textile belonged to someone, perhaps a maker, perhaps an owner.



Figure 14. Pictured texts on textiles, Bonampak Murals (images by Stephen Houston and Gene Ware, drawing by Stephen Houston, courtesy of Bonampak Documentation Project).


This pattern has also been attested in a carving now in the Dumbarton Oaks collection,  Washington, D.C. (Fig. 15Tokovinine 2012:69–71, fig. 32, 33). It refers to the painting on the cloth and to the ownership (or making) of that cloth, but by someone whose name disappears behind a (now-eroded) belt ornament. The statements are almost coy in providing the phrasing of possession but not any particulars about personal identity.



DO Chancala overlay.jpg

Figure 15. Chancala-area panel, Chiapas, Mexico, Dumbarton Oaks Research Library and Collection, PC.B.537 (drawing by Alexandre Tokovinine, with added highlighting in red of text on textile).


The art historian Meyer Schapiro paid close attention to pictured text in Western art. Some of his observations are parochial, as in his categorical insistence that writing consisted of “arbitrary marks” violating the “unified whole” of a pictorial work (Schapiro 1996:119). In the Maya case, sundering imagery and writing hardly makes sense for an iconically based script. But, to useful extent, Schapiro was concerned with the “material reality of the spoken and written word” (Schapiro 1996:120) and with the problem of viewpoint. Was inserted text to be “read” by a figure within a picture, a seated Evangelist examining a Gospel oriented to his “gaze”? Or was the pertinent observer “outside,” looking at that same Gospel but now laid out for clarity, not as any real book would be? Evolutionism creeps in: for Schapiro the latter was “an archaic object-oriented attitude,” to be contrasted with “the foreshortenings and overlappings that transform the constant shapes of objects,” crafting “an image coherent to the eye with a unifying perspective” (Schapiro 1996:121, 132, 141, 181).

“Archaic,” “ordered,” “whole,” “coherent,” and “unifying” are words of prejudicial intent. Schapiro’s voting record is clear. Yet pictured writing among the Classic Maya recalls similar patterns and a roughly parallel contrast of “attitude.” During a few decades in the Classic period, perhaps over a century, and in certain kingdoms or ateliers only, the need for presentational clarity gave way, in playful experiment, to what the eye could see, not what was known to be there. (Codices seemed strenuously off-limits.) This could be understood by the culturally laden term of “realism,” but it points more to a privileging of viewers, a means of summoning direct experience, and bringing observers into physical communion with acts on record. For the Maya, this was what writing looked like.


Note 1.  “Communication” is sometimes not quite the right label. Speaking to oneself can be seen as a disorder in Western psychiatry, which orders up lithium and other drugs to control such an impulse. To more recent thinking, chatter without an audience helps to organize the brain and to direct the tasks we perform (Kirkham et al. 2012). Moreover, in communicating with others, lip-reading offers a non-phonic option, provided that labial movement can be clearly seen (Auer 2010).

Note 2. Tactile scripts like braille and the “night writing” of Charles Barbier de la Serre present another story of sensory messaging. They are, as relatively recent innovations, far more restricted in use (Weygand 2009: 39, 299).

Note 3. Left to the side is an unusual occurrence: glyphs that appear as objects when they are most unlikely to have been seen in this way (e.g., K771, in which an “8 Ajaw” day sign “sits” on a surface, much like seated figures—all supernaturals—posed nearby). Year-bearers, numbered days marking the shift of years, also perform in this way (Stuart 2004:fig. 4).


Acknowledgments  Megan O’Neil kindly shared an image of the vessel at the Los Angeles County Museum of Art.



Auer, Edward T., Jr. 2010. Investigating Speechreading and Deafness. Journal of the American Academy of Audiology 21(3):163–168. Speechreading

Carrasco Vargas, Ramón, and María Cordeiro Baqueiro. 2012. The Murals of Chiik Nahb Structure Sub 1-4, Calakmul, Mexico. In Maya Archaeology 2, ed. Charles Golden, Stephen Houston, and Joel Skidmore, 8–59. San Francisco: Precolumbia Mesoweb Press.

Chinchilla Mazariegos, Oswaldo. 2005. Cosmos and Warfare on a Classic Maya Vase. RES: Anthropology and Aesthetics 47:107–134.

Coe, Michael D. 1973. The Maya Scribe and His World. New York: Grolier Club.

Coe, Michael D. 1977. Supernatural Patrons of Maya Scribes and artists. In Social Process in Maya Prehistory: Essays in Honour of Sir Eric Thompson, ed. Norman Hammond, 327–47. New York: Academic Press.

Halperin, Christina T. 2016. Textile Techné: Classic Maya Translucent Cloth and the Making of Value. In Making Value, Making Meaning: Techné in the Pre-Columbian World, edited by Cathy L. Costin, 433–467. Washington, D.C.: Dumbarton Oaks Research Library and Collection.

Houston, Stephen D. 2014. The Life Within: Classic Maya and the Matter of Permanence. New Haven: Yale University Press.

Houston, Stephen D. 2016. Crafting Credit: Authorship among Classic Maya Painters and Sculptors. In Making Value, Making Meaning: Techné in the Pre-Columbian World, edited by Cathy L. Costin, 391–431. Washington, D.C.: Dumbarton Oaks Research Library and Collection.

Houston, Stephen D. 2018a. The Gifted Passage: Young Men in Classic Maya Art and Text. New Haven: Yale University Press.

Houston, Stephen D. 2018b. Writing that Isn’t: Pseudo-Scripts in Comparative View. Unpublished ms., www.academia.edu.

Kirkham, Alexander J., Julian M. Breeze, and Paloma Marí-Beffa. 2012. The Impact of Verbal Instructions on Goal-Directed Behaviour. Acta Psychologica 39(1):212–219. Speaking Aloud

Laporte, Juan Pedro, and Vilma Fialko. 1995. Un reencuentro con Mundo Perdido, Tikal, Guatemala. Ancient Mesoamerica 6:41–94.

Martin, Simon. 2012. Hieroglyphs from the Painted Pyramid: The Epigraphy of Chiik Nahb Structure Subt 1-4, Calakmul, Mexico. In Maya Archaeology 2, ed. Charles Golden, Stephen Houston, and Joel Skidmore, 60–81. San Francisco: Precolumbia Mesoweb Press.

Martin, Simon. 2015. The Old Man of the Maya Universe: A Unitary Dimension of Ancient Maya Religion. Maya Archaeology 3, ed. Charles Golden, Stephen Houston, and Joel Skidmore, 186–227. San Francisco: Precolumbia Mesoweb Press

Milman, Miriam. 2009. Does “Real” Tromp l’oeil Exist. In Art and Illusions: Masterpieces of Trompe l’oeil from Antiquity to the Present Day, ed. Annamaria Giusti, 21–32. Florence: Mandragora.

Schapiro, Mayer. 1996. Words, Script, and Pictures: Semiotics of Visual Language. New York: George Braziller.

Sider, David. 2010. Greek Verse on a Vase by Douris. Hesperia 79(4):541–554. Schoolroom

Sterling, Charles. 1952. La nature morte de l’antiquité à nos jours. Paris : P. Tisné.

Stuart, David. 2004. New Year Records in Classic Maya Inscriptions. The PARI Journal 5(2):1-6. Pre-Columbian Art Research Institute, San Francisco. Stuart Yearbearers

Tokovinine, Alexandre. 2012. Carved Panel. In Ancient Maya Art at Dumbarton Oaks, eds. Joanne Pillsbury, Miriam Doutriaux, Reiko Ishihara-Brito, and Alexandre Tokovinine, 68–73. Pre-Columbian Art at Dumbarton Oaks, No. 4. Washington, D.C.: Dumbarton Oaks Research Library and Collection.

Weygand, Zina. 2009. The Blind in French Society from the Middle Ages to the Century of Louis Braille. Stanford: Stanford University Press.

Wine, Humphrey, Paul Ackroyd, and Aviva Burnstock. 1993. Laurent de La Hyre’s “Allegorical Figure of Grammar.” National Gallery Technical Bulletin 14:22–33. Allegorical Figure