Aztec Writing: How does it really work?

I have never written on this blog about colonial or precolonial Nahua iconography or glyphic writing. This is partly because I tend to find codices boring, prefering to work with spoken language, and partly because other people are much more knowledgeable about these things than I am. Recently, however, I have started reading up on this, and had some talks with some specialists in Nahua writing. So here comes my attempt to describe the current debates within the field of Nahua writing studies.

First of all perhaps we need to point out what we mean by Nahua glyphic writing - most Nahua texts are of course written in Latin letters - but here we mean texts written with non-European signs. In the codices this kind of writing abounds, mostly used to write personal names and the names of places.

The name of Acamapichtli "handful of reeds",
ruler of Tenochtitlan,  written with
glyphs  and Latin letters.
The glyphs shows a hand holding
a bunch of reeds.

The name of the town of Xochimilco "Flower field place" written with Nahua glyphs:
The squares on the bottom represent a "field" - representing the root MIL- in Nahuatl
and the flowers on top represent the root XOCHI "flower" in Nahuatl.

As can be seen from the examples, we have a pretty good idea about how to read Nahua glyphs (because almost all glyphic texts are accompanied by writing in Latin letters), and generally we know what the say. But there is still some discussion about the principles based on which the Nahua wrote. Scholars have for example discussed if the Nahua simply depicted concepts with pictures, or whether their signs actually represent words and sounds? Today we know that Nahuas seem to have definitely used their signs to represent spoken words and sounds, but specialists are still discussing the technicalities of how the Nahua writing system worked.

Aztec Writing: Logographic? Phonetic? Semasiographic? Rebus-writing? 

The decipherment of Aztec writing is not really the classic kind of situation where we have a lot of long monolingual texts that we are unable to read because we don't know the language or the principles they used to write with, and where a bilingual text (such as the Rosetta stone for Egyptian hieroglyhs, or Landa's syllabary for Maya hieroglyphs) can provide the clues to finally crack the code. On the contrary, we know that Aztec glyphic texts were written in Nahuatl, and they supply us both with the glyphic script, and with a transliteration in to the Latin alphabet, and they often even with a Spanish translation. So most of the time the question is not what the glyphs mean, the question is how they work to give that meaning. Furthermore, most texts written in Aztec glyphs are short, and consist mostly of names of places and people.

Nonetheless, the decipherment of Aztec writing did have to pass through some phases of confusion. Particularly, a phase in which scholars didn't consider it to be a script at all, but just drawings that could only be read non-linguistically. In some sources we can still find the claim that the Aztecs did not have "true writing". The concept of "true writing" in this sense is used to describe the fact that it does not seem the Nahuas wrote long narrative texts using their script. Rather they wrote stories using sequences of pictures, and used the script to name the persons and places who participated in the stories.

The first scholars to note that the Aztec script in fact has a lot of "phoneticism" (i.e. it represents the sounds of spoken language, and not just ideas), were those 19th century antiquarians who first studied the early colonial codices. One of them was Joseph Marius Alexis Aubin, a young French scholar who accompanied Maximilian of Habsburg who went to Mexico with a contingent of French forces because Napoleon III (no, not Bonaparte) had appointed him the new Emperor of Mexico. Aubin had studied Nahuatl in France, and probably also had help from Maximilian's main Nahuatl specialist the Nahua nobleman Faustino Chimalpopoca who had worked with many of the Nahuatl language codices. Reading Aztec maps and tribute lists Aubin (1849 [1885]) realized that all the Nahua rulers and all the cities in the codices had little name glyphs attached to them, and that these name glyphs consisted depicted concepts or words that were part of the name they represented. For example all the places that had the locative ending -pan in their name, had a little flag in their name glyph. "Flag" in Nahuatl is of course pantli or pamitl - so Aubin realized that the flag represented the ending. Aubin deciphered a number of glyphs in this way, by seeing that they depicted short words that could be used both to signify the word they depicted but also words that sounded similarly (e.g. a stone "tetl" signified the syllable TE, a pot "comitl" signified the syllable KO, etc.). And he noted that there was a good deal of systematicity in the way the word signs were used, going as far as describing the system as an écriture syllabique [a syllabic script]. Aubin was particularly inspired by finding a version of the "Pater Noster" written partly in Aztec script - it was apparently used as a medium with which the friars could better teach the Nahuas the catechism. In the Nahuatl script the title said:

Flag            Stone              Cactus               Stone
PAN(TLI)     TE(TL)        NOCH(TLI)       TE(TL)
pa                  te                   noch                  te  

Aubin realized the flag glyph was /pantli/ "flag" representing the root pan, the second symbol was the stylized symbol for /tetl/ "rock" representing the syllable te, the cactus symbol was /nochtli/ representing the root noch (the tli/tl suffix is not part of the stem). The phonetic reading was PA TE NOCH TE - a fairly close approximation of the Latin "Pater Noster" (remember that colonial Nahuatl has no /r/). This convinced Aubin that the Aztec script operated from a syllabic principle - and he considered that Nahuas were already literate in this script when the friars arrived, which why it made sense for the friars to use it.

However, his argument was not very well received. A number of scholars argued that the phonetic signs were introduced by the friars simply using a rebus principle, and that it didnt represent pre-contact usage. Since all of the known texts using the script were from the early colonial period, this was hard to refute, and for many decades the consensus came to be that the Nahua script was not true writing but a "pictographic" form of primitive picture writing. Some scholars argued that Nahua iconography, and central Mexican indigenous iconography in general, is "semasiographic"in character which means that it does not represent a spoken language directly, but rather offered clues to the "reader" through which they can improvise a narrative in whichever language they happen to speak. Nonetheless a number of works on the principles used in Nahua writing were published, by Barlow & McAfee (1954), Dibble (1971), Prem (1992) and Nicholson (1973) - Nicholson particularly argued that perhaps there was a greater degree of phoneticism in precontact Nahua writing than had previously been thought - but the mainstream view continued to be that pre-conquest Nahuas did not "write".

Lacadena's Syllabary:

Lacadena's 2008 Syllabary,
published in PARI journal 8:(4):p. 23

In 2008 a major event happened in Nahuatl writing studies. Spanish linguist and Epigrapher Alfonso Lacadena published an analysis of Nahuatl writing in which he argued that Aubin had been right to call the script "syllabic". Lacadena had previously participated in the work of deciphering the Maya script, and he argued that the Nahua system was essentially similar - combining a set of logograms representing entire words, with a syllabary that could be used to phonetically spell most of the possible syllables of the language - making the system capable of representing spoken language.

Lacadena had been able to arrange the known phonetic elements of the script into a syllabary similar to the Maya syllabary - showing that Nahuas in fact had syllabic signs for almost all the possible syllables in the language. He argued that several of the already recognized signs that had previously been read as logograms ("whole-word signs") in fact ought to be read as syllabograms ("syllable signs") consisting of only a consonant and a vowel - the flag glyph for example ought to be read pa and not PAN,  the pot glyph ko and not KON, etc. He shows that it is possible to demonstrate that some signs that were originally read as logograms, had to be read as syllabograms in certain words. For example the drum sign which had traditionally been read WEWE, he proposed should be read - at least sometimes - simply as we. And he showed this by demonstrating that it was used to write names such as tohuexiuh /towexiw/ where it only represented one syllable, not two.

A very interesting set of readings provided by Lacadena deal with the syllable /wa/ which he shows can be written either with two parallel lines, with two leaves or with a grasping hand. He argues that the two lines come from the verb wawana "to scratch", that the two leaves are amaranth leaves wawtli (Personally, I think they are more likely from ixwa "leaf" and in one of his examples the glyph actually co-occurs with an IX logogram (the eye), suggesting a dual representation of the ix-syllable - but clearly in many other cases it is only read as wa), and that the grasping hand is a logogram based on the possessor suffix -wa' (e.g. in michhua "fish-possessor" or mazahua "deer-owner"). The two former readings he considers syllabograms, while the third he considers a logogram (inspite of not representing a word, but a grammatical morpheme). These readings are very important and constitute an important move towards the syllabary.

The name of Spanish encomendero Luis Vaca written
in Nahua glyphic writing. The bottom sign is an olote (dry maize stalk) (OLO),
then an eyeball (IX) and on top two leaves (wa [or maybe (IXWA]), 
the long plate shaped glyph is probably a plate CAX or ca. Giving the full reading (o)LO-IX-WA-KA(X) 

Lacadena demonstrated that as in the Maya (and Egyptian) scripts often the syllabograms were used as phonetic complements, added to logogram to support a specific reading - or even at times to force one of two phonetically distinct readings of the same logogram.

In Lacadena's interpretation the Nahuas had a fully developed logo-syllabic writing system already in the pre-contact period, although he also recognizes that the use of phoneticism may have become more prominent as the Nahua logosyllabic script competed with the Latin alphabet. By setting up the Nahuatl phonetic signs into the syllabary format, Lacadena moved the discussion forward - it was now no longer relevant to discuss the phoneticity itself, from now on the discussion would have to be about the degree of systematicity with which the Nahuas used phonetic symbols.

Whittaker's Challenge:

In 2009, anthropologist and epigrapher Gordon Whittaker challenged Lacadena's vision of how the Aztec script worked. Following earlier work by Hanns Prem, Whittaker had studied the phonetic principles of Aztec writing for awhile and did not agree with Lacadena's strictly syllabic interpretation. In his paper "Principles of Nahuatl Writing" he proposed that rather than being a strict logo-syllabic script, Nahuatl writing allowed a much wider range of possibilities. For example he argued, most of Lacadena's syllable signs could be read both as syllable signs with a consonant and a vowel, but also as logograms including syllable final consonants; and sometimes polysyllabic logograms could be read as syllable signs, selecting just one of their syllables as their phonetic value. In this way the Nahuatl script included both logograms, syllabograms and bisyllabograms. Whittaker therefore suggests that the way that Nahuas read their texts were not as straightforward as Lacadena suggests - they needed access to a wider array of interpretive processes than simply reading the syllabograms. In the example below, the place name Chipiltepec is written with the syllabogram chi (depicting the edible seed "chian"), the logogram HUIPIL"blouse", the logogram TEPE and the syllabogram te used as a phonetic complement.

chi     (hui)pil     TEPE    te
"Chipiltepec"

The interesting part here is that the logogram HUIPIL is not read as a logogram but as a syllabogram, ignoring the first syllable and reading only the second syllable pil. This challenges the strict division Lacadena poses between logograms and consonant-vowel syllable signs.

Whittaker also notes that several of the signs that Lacadena considers syllabograms, can be read both as syllabic Consonant-Vowel (CV) reading, as well as as a Consonant+Vowel+Consonant (CVC) group. For example the "flag" can be read both pa and pan, and the "tooth" can be read both tla and tlan. He seems to consider that the reading with final -n is not a real logogram, because it is not usually used to represent the words "flag" and "tooth", but rather to write the locative endings on place names - locative endings that happen to be homophonous with the words depicted by the logograms, but which have completely different meanings. This is another reason Whittaker considers some of the Nahua signs to be phonetic, but to represent neither CV syllables nor words - but longer sequences such as CVC or even CVCV.

Furthermore, Whittaker implies, it does not seem that Nahuas thought of their phonetic writing as a self-contained closed system in the way that a syllabary is generally thought of. Rather the way he describes the system it is an array of conventions, many of them fairly loose, which the Nahua used in a creative and intuitive way. Whittaker notes for example (in a talk he gave at Yale in May 2016) that the Nahuas used ways of expressing concepts that are unique to the Nahua script. For example they used color of glyphs to convey color words (i.e. the word chichiltepec "red mountain" written as a red mountain where the color of the mountain is used as a logogram).

Whittaker also produces some interesting new readings of specific signs, for example he shows that a worm-shaped glyph can be read COCHIN, based on the word ihcochin "earthworm".  For example the name Tlacochin can be written with a combination of four signs: TLACOCH "javelin", (ih)COCHIN, tla/TLAN "tooth", and co/CON (here I allow for both the possibility of reading these as logograms and as syllabograms, hence the /). Interesting here each syllable (except -in) is spelled phonetically twice - providing no possibility for misreadings. Here, following Lacadena's method, co and tla can be read as syllable signs, and TLACOCH as a logogram. But the worm, spelling syllables COCHIN can neither be a logogram (because the name is likely derived from tlacochtli "spear" and not from ihcochin "earthworm") or a simple syllabogram.

The name Tlacochin written both with Latin letters and with
four Nahua glyphs. It is the wormlike creature on top of the teeth
that Whittaker reads as (ih)COCHIN.

Finally, Whittaker argues, that if the Nahuas indeed thought of their system as a complete syllabary, then it was a very deficient syllabary since there are no known syllabograms for several of the most frequently occurring syllables in the language such as ti, ni, ki, tli, kwa, ya, etc. This would make it impossible to actually represent spoken language using only syllabograms. And this in turn suggests that the Nahuas did not conceive of their phonetic signs as a systematic complete syllabary, but that rather than an organized system the syllabograms formed a repertoire of signs that had grown through accumulation in the course of the Aztec writing tradition. An incipient syllabary perhaps, but not quite there yet.

So, Whittaker and Lacadena agree on considering the Nahua script to be highly phonetic - also in pre-contact times - but they disagree on what kind of system it was, and how Nahua scribes interpreted the phonetic signs. Nonetheless, it doesn't seem that either of the two epigraphers are willing to concede many points to the other. To my knowledge, Lacadena has yet to respond publicly to Whittaker's argument, while Whittaker's papers seem to me to exaggerate some of the differences between the two models (for example the difference between logograms and bisyllabograms isn't very big in practice - it is mostly a question of labeling I think). So where does Nahuatl glyph studies go from here?

My 5 Cents Worth:

While I am not an expert on Nahuatl writing, to conclude I will share my own evaluation of Lacadena's and Whittaker's competing models.

Lacadenas demonstration of pervasive phoneticity of the Nahua script is extremely important, and did inaugurate a new way of looking at the Nahua script. But I am not convinced by the argument that the script can be understood as a strictly logo-syllabic script. Whittaker demonstrates that the script is considerably more flexible than that, including types of signs not known in any other scripts (e.g. the use of color to signify color words), and making more readings available to the reader. I simply dont think that there is a convincing reason to think that Aztec scribes and readers thought of their signs as representing mainly CV syllables rather than for example representing entire word-stems that could e read for their phonetic value as well as as the word-concept they represent. On the other hand I am not sure if Lacadena is actually argueing that the Aztec scribes thought of their script as a syllabary, and that the syllabic interpretation is "psychologically real" - this I think would be a problematic claim. But it is possible, I think, that he simply meant to demonstrate the pervasiveness of syllabic signs in the system to show that it is not as different from the Maya system as it has been claimed. If this is the case I think the point was well-made and Whittaker's argument can simply be considered a demonstration of additional features of the script, that Lacadena did not spend time exploring because his aim was another.

On the other hand, I think Whittaker exaggerates the differences between the two interpretations. Particularly, it seems that Lacadena does allow for many of the readings that Whittaker considers disyllabograms and CVC syllables - he just considers those logograms used according to the rebus principle (Whittaker apparently doesnt like using the word "rebus" for phonetic readings of logograms). This seems to me as a mostly theoretical discussion that doesn't really impinge on our understanding of the script itself. I assume that Lacadena acknowledges that when a logogram can serve the function of expressing another word than the one that is actually depcited, this is indeed a phonetic rather than a logographic reading of the sign.

Some of Lacadena's readings also beg the question about whether each sign really has only one possible reading. For example many of his syllabograms clearly must have originated as logograms, for example the tla- and pa syllabograms which are represented "logographically" with signs that must have been read originally as TLAN and PAN - and which sometimes still are clearly used to write out those exact phonetic sequences with final -n (for example in toponyms). Hence it seems to me that we must concede that these signs are either used sometimes as syllabograms and sometimes as logograms, or that we can simply conclude that the final consonants can sometimes be omitted to produce syllabic readings from logograms. Probably, over time some logograms were conventionalized as being mostly used syllabically, while others kept their double readings. Whittaker produces further evidence for this type of process when he shows that the HUIPIL logogram can be read simply as PIL. And another example would be if I am right that the double leaf symbol was originally and occasionally a logogram IXWA which then became conventionalized as a syllabogram wa. Although the possible alternative  reading of "Luis Vaca" as OLO IX IXWA (where IX is a phonetic complement noting that the leaf sign should be read IXWA and not simply wa), instead of Lacadena's reading OLO IX wa, suggests that maybe two readings continued to coexist).

Where, I think Whittaker's interpretation is strongest is in that it paints a picture of a very different process of writing than Lacadena's  - a process that is highly creative, combining intuitive conventions and a repertoire of established signs with local innovation. I think it is clear that Nahua writing was in many ways unique when compared to other writing systems, for example in being seemingly open-ended, and without many fixed conventions.

In any case, I think any future engagement with Nahua glyphic script needs to take both Lacadena's and Whittakers arguments into account, in order to pick out the important insights of both of them. A question that needs further exploration is how systematic and or flexible the system really was  - what were the limits? Which kinds of readings do we not get that we might? Also additional support for the syllabary could be provided for example if we were to find syllabograms used to spell final consonants in the Maya style (where for example BALAM can be spelled ba-la-ma).

In the end it seems to me that we cannot escape the fact that even though the Nahua script was highly phonetic, and clearly shared principles with the Maya logo-syllabic script, their approach to writing was very different from that of the Maya. Nahuas do not generally seem to have considered it a function of their script to represent spoken language much beyond personal and place names used to label narratives that were represented with pictures. No clear examples of full sentences have been deciphered, and indeed it doesnt seem that the script was even capable of writing such sentences given the lack of syllabograms for such important grammatical morphemes as ni, ti, and ki  which would be necessary to represent almost any slightly complex Nahuatl sentence. [Edit: After publishing this, I was reminded of the long sequences of glyphs found in the Codex Xolotl which may well represent something like full sentences - however I do not believe these have been deciphered yet].

Bibliography:

• Aubin, Joseph Marius Alexis. 1849. Mémoire sur la peinture didactique et l’écriture figurative des anciens mexicains. Paris: Paul Dupont

• Barlow, Robert H. and Byron McAfee. 1949. Diccionario de elementos fonéticos en escritura jeroglífica (Códice Mendocino). Mexico, D.F.: Universidad Nacional Autónoma de México.

• Dibble, Charles E. 1971. Writing in Central Mexico. in Gordon F. Ekholm and Ignacio Bernal (eds.), Handbook of Middle American Indians: Archaeology of Northern Mesoamerica, Part 1, Vol. 10, pp. 322-332. Austin: University of Texas Press.

• Lacadena, Alfonso. 2008a. Regional scribal traditions: Methodological implications for the decipherment of Nahuatl writing. The PARI Journal 8 (4): 1-22.

• Lacadena, Alfonso. 2008b. The wa1 and wa2 phonetic signs and the logogram for WA in Nahuatl writing. The PARI Journal 8 (4): 38-45.

• Lacadena, Alfonso.  2008c. A Nahuatl syllabary. The PARI Journal 8 (4): 23.

• Nicholson, Henry B. 1973. Phoneticism in the late Pre-Hispanic Central Mexican writing system. in Elizabeth P. Benson (ed.), Mesoamerican Writing Systems, pp. 1-46. Washington: Dumbarton Oaks

• Prem, Hanns J. 1992. Aztec writing. in Victoria Reifler Bricker (ed.), Handbook of Middle American Indians, Supplement, Vol. 5: Epigraphy, pp. 53-69. Austin: University of Texas Press.

• Whittaker, Gordon. "The Principles of Nahuatl Writing." Göttinger Beiträge zur Sprachwissenschaft 16 (2009): 47-81.