torsdag den 6. juli 2017

The relation between Nahuatl, Cora and Huichol: initial thoughts

Lately I have been studying the two languages Cora (Náayeri) and Huichol (Wixarika) spoken in the Mexican state of Nayarit which like Nahuatl belongs to the Uto-Aztecan language family. 

Everyone agrees that Cora and Huichol are closely related to eachother, and the languages are grouped together under the name Coracholan, but it is an open question whether Nahuatl is more closely related to Coracholan than to the other Uto-Aztecan languages of Mexico. 

There are some significant similarities between the three languages, but some scholars such as Karen Dakin (2017) and Terrence Kaufman (2001) argue that these similarities are most likely due to borrowing between the languages. In 1978,  Lyle Campbell and Ronald Langacker proposed that Nahuan and Coracholan had undergone a shared sound change in which they all changed the previous proto-Uto-Aztecan (PUA) *u to /ɨ/. This would suggest that the languages all began as dialects of a single ancestor language (proto-Corachol-Nahuan) which was itself a daughter language of proto-Southern Uto-Aztecan (PSUA). Karen Dakin in turn has argued that Nahuatl never underwent this change but that Nahua instead changed PUA *u directly into /i/ with some instances changing to /e/ instead.

It is an interesting question whether Coracholan and Nahuan form a single group; a question, which, if we could answer it in the positive would bring us much closer to understanding the origins of Nahuatl, both in terms of geography (it would then have originated in Western Mexico, and not as Kaufman (2001) suggests in North-Eastern Mexico), but also in terms of the linguistic development the language has passed through.

As I have studied Huichol and Cora from the sparse materials about the languages (and from a short field stay in Nayarit, which is still on-going as I write), I have compiled a list of cognates (related words) between the three languages - and I believe that they are indeed closely related and that a single ancestor language can be reconstructed. I have not fully analyzed the patterns of similarity, but there are some that look particularly promising. Below I describe five particularly interesting etymologies:

Three women figures standing in front of hanging hearts
  (detail from a Huichol yarn painting, found at wikicommons).

1. "To live/heart" 

In Nahuatl the root yo:l-  has to do with living, and being alive, and with the heart. From this root we get the intransitive verb yo:li "to live", yo:llo:tl "heart" (</yo:lyo:tl/) and yo:lkatl "animal/living thing", and the spiritual term te:yolia "that which causes people to live" or "life force".  The noun yo:llo:tl or yo:llohtli "heart" is composed of the root yol:- and the suffix -yotl/-yoh that derives an abstract noun or an inalienably possessed noun suggesting a meaning of something like "life-essence" or "aliveness" as the original meaning of the Nahuatl word for heart.

In Huichol the word for "heart" or "soul" is 'iyáari, and as in Nahuatl it is central to the way that Huichol people conceive of life and life force. When one dreams the 'iyáari travels to other worlds, and when one dies the 'iyari of a person begans a journey through the otherworld towards the land of the dead. Similar beings have similar 'iyáari because the 'iyáari travels through kinship bonds, which also means that the Huichol have a different 'iyáari from other beings.

In Cora, life is expressed with the verb rúhuri or rúuri"he lives" and the noun life with the word rúh. And the word for "soul" or "living thing" is rúuri-kame (=yol-katl).

At first glance the three words look quite different, but I consider them cognate. The correspondence between Nahuatl /l/ and Corachol /r/ is easy to explain - it is a very common change in the worlds languages, and in fact Huichol /r/ is often pronounced as /l/. Here we may propose that Nahuan changed Corachol-Nahuan *r to /l/ .  And Huichol and Nahuatl both have /y/ as the first consonant. This leaves us to explain why Cora has begins with /r/ and why the three languages each have completely different vowels Nahuatl /o:/, Huichol /aa/ and Cora /u(h)u/.

Cora has the reflex /r/ corresponding to Nahuan /y/ in several words: Cor. šaihruh "fly": Nah. sayo:lin "fly",  Cor. rú-viveh "tremble": Nah. wewe-yo-ka "tremble" (notice that Cora constructs the compound with the roots reversed relative to Nahuatl), Cor. ɨra’a "wife", Hui. 'iya "wife", Nah. ye "mother", (found in the Isthmus varieties),  and in the locative suffix describing a place full of x, e.g. Nah. okoyoh , Hui. hukuyá and Cor. úkúuré all meaning "place full of pinetrees". So Cora /r/ is in fact a regular reflex of Nahuatl and Huichol /y/ and it is particularly common before Cora /u/ corresponding to Nahuatl /o/. We may posit as a sound law that Cora changes a proto-Coracholan *y to /r/ before back vowels (and maybe some other vowels).

As for the differential vowels, there are also patterns that explain this: the first pattern is that Nahuan long /o:/ often corresponds to diphthongs /ew/ or /eu/ and /aw/ or /au/ in Coracholan for example:  Cor. sau'uh, Hui. xiauri, Nah. so:lin "quail"; Hui. teukíya, Cor. tye'ukwa, Nah. to:ka "bury", Cor. taya'u ("god"), Hui. tayew "name of the sungod", Nah. teo:tl; "deity/sacred"; Cor. tyauhsa Nah. to:san "mole" and Hui. tau "sun" and Nah. to:na "heat (of the sun)". Indeed one Cora dialect, that of Santa Teresa also has the change to a long /o:/ in its reflexes of those particular words- and the change is well known also from European languages, it is for example the same change that turned Latin aurum into French or and Spanish oro. Hence we can propose one more sound change: Nahuatl turns proto-Corachol-Nahuan *au into /o:/. But in this word Huichol and Cora does not have /au/ but instead Huichol has /aa/ and Cora has /u(h)u/. What went wrong? There is also a correspondence between Nahuatl o and Corachol u - and in those case we can be fairly sure that /o/ was the original (e.g. the words for fly, tremble, pine, and the "full of x" suffix). So the Cora form *uu could from a previous *oo - in that case Nahuatl and Cora would have both changed the *au to /o:/ and Cora would then have changed /o:/ to /u:/. This would be a sound change shared by Cora and Huichol, which would considerably strengthen the claim that the languages are closely related. It could however also be the case that the word was borrowed into Cora from Nahuan, since Nahuan /o/ is always borrowed as Cora /u/. But the fact that there is a Huichol cognate suggests that it is not borrowed, and the fact that Huichol has the vowel /aa/ suggests that perhaps this word simply saw a previous corachol-Nahuan *au change to /o:/ in Nahuan, to /aa/ Huichol and to /uu/ in Cora- with Cora and Huichol each leveling the previous diphthong, but one language choosing to level to the a and the other to the u.

This would suggest that all three words derive through fairly regular changes (at least changes of which there are other examples between the languages) from a proto-form close to *yauri or *yahuri .

This seems like a likely explanation which just requires us explain why Cora and Huichol leveled the diphthong in this word but not in the many other words with a sequence of /au/. I will keep looking for an explanation of this exception.

2. "Breath/to breathe"

The Nahuatl word for breath is ihiyo:tl, and it is also a culturally salient word because like yo:li and te:yo:lia it refers to one of the lifeforces that characterize human beings (Nahuas recognized at least three life forces, the te:yo:lia (tied to the heart), the ihiyo:tl (tied to the breath) and the to:nalli (tied to the sun and the day)).

The Cora word for breath is yéh, and the form that means "to breath heavily" is í'iyeh with the prefixed í'i-.

The Huichol word for "lungs" or "breath" is 'iiyáte. 

While the words seem quite similar to the word for to live (I would say they are probably derived from the same root at an earlier point in the languages' development), here we see two major differences: Cora does not have /r/ but /y/, and Cora does not have /u/ but /e/. The reflexes for Nahuatl and Huichol are the same as for the previous etymology y:y and o:/a. So why is Cora different? First we may posit that Cora retains y from *y when occuring infront of the vowel e (i.e. front vowels). This makes good sense since /y/ is a palatal consonant and palatals like front vowels - so we may propose a modification of the previous rule (Cora *y>r) to say that Cora depalatalizes *y to *r except before front vowels.

But then we need to understand why the front vowel is even there to begin with! Here we may propose the same development as before - we have a rising diphthong beginning with /a/  and while Huicho levels to the first vowel Cora levels the diphthong to the second vowel. So that would suggest a diphthong of /ae/. But why would Nahuan turn /ae/ into /o:/?  I don't think it would, but it might turn the diphthong *aɨ into /o/ - at least that seems to happen in the words for "waste/trash" which is Cora saɨri and Nahuatl sol-. And in Corachol PUA *ɨ regularly becomes /e/. So that lets us reconstruct PCN *aɨ which becomes PC *ae and Huichol /a/ and Cora /e/ (I don't know why there doesn't seem to be a long vowel in any of these except for Nahuatl, this could be a secondary shortening). And the previous rule allows us to show that the depalatalization of *y to /r/ in Cora happened after the leveling of the diphthong.

In this way, using basically the same sound changes proposed for the first etymology, leads to a final reconstruction of a proto-Corachol-Nahuan root *iyaɨ "breath/to breathe". Note also how the Huichol nominal ending -te seems to match the Nahuatl absolutive suffix -tli, suggesting that this suffix comes from a previous *(Karen Dakin and Alexis Manaster-Ramer have already suggested this). 

3. "To degrain corn/corn kernels"

In Nahuatl there are several words for maize: in general maize is sentli or sintli, though in some varieties this refers specificlly to the dried corncob as opposed to the fresh corn cob which is called yelo:tl ot elo:tl. But another word is for the dried corn kernels after they have been taken off the cob, these are the ones that are boiled with lime to make the nixtamal porridge that is in turn ground into dough. These dried kernels are called tlao:lli (or in some varieties tleo:lli, tlao:yalli, or tau:lli and other variants) this is a patientive noun derived from the verb oya "to degrain corn". The root of this verb is o: and -ya is one of the different verbal stem formant suffixes.

I would usually posit that the old common-Nahuan term was /tlao:yalli/ which ponemically is /ƛa-o:ya-ƛi/, where the /y/ is assimilated to the /ƛ/ to become a double ll. But the most common form of the word in Nahuan is /tlao:lli/ where the thematic vowel -a of the transitive verb seems to be deleted (as sometimes happens to penultimate vowels in Nahuan).  But it is also possible that the common-Nahuan form was tlao:lli with the phonemic form /ƛao:l-ƛi/. Based on the Nahuan form alone we would reconstruct a pre-Nahuan (that is, the stage before proto-Nahuatl where the language had still not undergone the changes that would turn it into Nahuan) /*ta- o:-ri-tɨ/ or /*ta- oya-ri-tɨ/ where the -ri- is the passive morpheme that becomes -l- in Nahuan.

Now, in Huichol the corresponding forms I have found in a dictionary is 'urika "to degrain corn" and reuyusáata "corn grain", ("corn" in general is 'ikú). the root here is 'u-.   In Cora the forms are yuu "degrain corn", yuuri “corn grain (and in general)” and yuhuri  the verbal form for “it has been degrained”, the root clearly being yuu-. 

Where Nahuan has the stem formant -ya, Huichol has the transitive verbal stem formants  -rika, but in the noun we see the form uyu suggesting that the long o in Nahuatl is a simplification of an original oyo sequence (making the full Huicho root uyu). We could then hypothesize that the Huichol re- prefix is cognate with the Nahuan tla- prefix, and that the -sáa suffix is a patientive like the Nahuan -l- and the -ta suffix is cognate with the Nahuan absolutive suffix.

In Cora the -ri suffix corresponds to the patientive suffix in Nahuatl, and the  root yuu seems to correspond with the Nahua root o: and the Huichol root 'u(yu). We know from the Cora vocabulario published by Father de Ortega in 1732 that Cora in the 18th century still had traces of an original set of absolutive prefixes, one of which was -t (Vázquez Soto 2000), and if we look up the word maiz (yes, the 1732 cora vocabulary is available online!) we find the form yuurit showing the final absolutive suffix. Now the three languages match entirely phoneme for phoneme and morpheme for morpheme.

This gives us the following correspondence between the forms in the three languages:

Proto Nahuatl:
/*ta- o:   -ri  -tɨ
/re-  uyu -sáa -ta/
/Ø- yuu  -ri -t/

 OBJ- degrain -PASS -ABS

This comparison gives us some valuable information about the grammatical development of the languages as well: Nahuan includes the verbal formant when it derives verbs (unless we argue that the form tlao:lli, is in fact the one found in common Nahuatl and that it didn't drop the -ya verbalized but simply never included it, which I think is likely). It also shows us that if this comparison of morphemes holds, then Huichol and Nahuatl are similar in both having the prefix tla- which would then be expected to be an object marker referring to an inanimate object "something", but that Cora does not need such an object marker on a patientive noun derived from a transitive verb.  This suggests that Huichol and Cora have both developed a grammatical requirement of transitive nouns to mark their objects arguments explicitly, while Cora has not. On the other hand Nahuatl groups with Cora in having the patientive suffix -ri,  suggesting that we need to reconstruct it for proto-Corachol-Nahuan and consider the Huichol -sáa an innovation. But finally, it is interesting that while the word in each language is composed of cognate morphemes and following the same grammatical model, they result in different words, suggesting that while the proto-Corachol-Nahuan language had a verb for "degraining corn", each daughter language derived the noun from this verb in a separate process.

So we can reconstruct *oyo or *oo as the proto-Corachol-Nahuan verb for  "to degrain corn"; and while we could reconstruct the patientive noun as something like *ooritɨ or *oyoritɨ, we should probably not do that, since it seems each language derived the noun on its own.

4. "To bury/plant" (and "spiders" and "names")

Apart from grinding corn kernels into dough another thing you can do with them is to plant them, that is use them as seeds. In Nahuatl the word for planting is /to:ka/, and it also means "to bury" (persons or things) and the word for seed is /ač-ƛi/.  Interestingly the word for name and naming is a homophone /t:oka/ "to name" and /to:ka-yi-(ƛ)/ "a name", and there is another set of near homophones namely /toka/ "follow" and /tokaƛ/ "spider", with a short /o/ instead of a long one.  Such homophones and near homophones make for interesting cases in reconstructing a relatoinship because they test our understanding of the phonology of the proto-language.

In Huichol "to plant" is tuárika, but "to bury" is a separate verb tewkíya. "Seed" is hasí,  "name" is tewa and spider is tuuká. A word that seems related to "name" is tewkari which refers to that grandfather of a child who gives it its name in a naming ceremony. The word for "follow" is the seemingly unrelated veiyarika, which seems to be a transitive form of the verb for "to go".

In Cora "to plant" is the seemingly unrelated wíité, but "to bury" is  tye’ukwa or in Cora of Sta. Teresa tyo’kwa. "Seed" is haȼí, "name" is tyáwa and spider is tu'uká (1732 tùcati). The word for to go is given as tabahara in the dictionary of the McMahons. 

This gives us a set of comparisons like this:


"to name" 
to:ka-yi-(ƛ) “name”
"to name someone"
tewá “name”
tewkari “name-giver”
"to name"
"to name someone"
tyawarit (1732)
tùca-ti (1732)

The two that are straight forward are "seed" and "spider": 
The differing vowel in tu'uka/toka points towards an original *o, and we can see that apparently the double vowel in Huichol and the glottalized vowel in Cora do not correspond with a long vowel in Nahuan. We might suggest that the first syllable ended in a glottal stop, *to'  and that in in Corachol it became glottalization with rearticulation, and in Huichol the stop disappeared between two identical vowels which then coalesced into a long vowel, whereas in Nahuatl it simply disappeared. Cora and Nahuatl show reflexes of the absolutive suffix -tɨ, whereas Huichol does not. This gives us a proto-Corachol-Nahuan reconstruction of *to’ka-tɨ.

In the word for "seed" Nahuatl lost the initial *h (probably from an earlier PUA **p), palatalized *ȼ to č before the front vowel i and added the absolutive suffix (the word is not recorded in the 1732 Cora vocabulary, but it probably did have the suffix in Cora). Huichol systematically changes Corachol *ȼ to /s/. This gives us a proto-Corachol-Nahuan reconstruction of *haȼí-tɨ.

For "bury", "plant", and "name" it looks like we may have two or three different etyma which have become homophonous in Nahuatl. In the etymology for "life/heart" we saw that PCN *au/aw diphthongs became Nahuan /o:/. Here it looks like the same happened for the difthong eu/ew. All the forms in all thre languages can be derived from a root of the shape /*tewa/: In Nahuatl the root becomes /to:/ which then combines with the verbalizing stem -ka, but in Corachol the development seem to depend on the accent, which in Huichol seems to cause an apocope of unstressed syllables.

In Huichol we see four different developments (acute accent shows primary accent with high pitch, and grave accent a secondary rhythmic accent with low pitch, unmarked syllables are unstressed and low pitch):




We can see that when the root /tewa/ combines with a bisyllabic suffix which has the accent on the first syllable then the unstressed syllable in the root is lost. The accent doesn't seem to be lexical though because the "name" root has stress on the second syllable when occurring alone and on the first syllable when occurring as part of the word for "name giver". Probably rather accent shifts were used for grammatical effects, as they still are in Huichol where plural forms of some word are made by shifting the accent from one syllable to the other, or as in Cora where some noun/verb pairs have alternating accent patterns. Incidentally, it is interesting to consider whether the Huichol word for "bury" can be analyzed as "dust-enter/house-Verbalizer". In Nahuatl "dust" is tew-, enter is -aki and -ya is an inchoative verbalizing formant. In Huichol is "house", "enter" is haayá, but  I have not found a root corresponding to Nahuan tew-. 

In Cora we see two different developments in the word for "bury" and "name":  In both words the *t is palatalized to /ty/, which again tells us that the first /a/ in the Cora word for name is a secondary development since this palatalization usually happens only before front vowels. This suggests a process of regressive vowel harmony having affected the word and turning the original /e/ into an /a/ in anticipation of the vowel in the second syllable - this process is found both internally in Cora and in Nahuan.

In the 1732 vocabulary the word has two additional suffixes not found today: tyawa-ri-t. The first suffix -ri could be one we have seen before as the patientive suffix, and the -t is the remnant of the absolutive suffix *-tɨ. But in fact rather than being a patientive suffix I think the -ri- suffix in Cora is cognate to the Nahuan suffix -yi- which signals abstractivation or inalienable possession. This would make the Cora morpheme sequence -ri-t cognate to Nahuatl the Nahuatl sequence -yo:-tl.   

In the word for "bury" the Cora forms show a glottalization between the /e'u/ in the form from Jesús Maria and after the /o'/ in the form from Sta. Teresa. The Sta. Teresa form is a secondary development of the eu sequence, and in fact we see Sta Teresa changing both the sequences /eu/ and /au/ to /o/ - just as in Nahuatl! The final change to be explained is why Cora has /kwa/ instead of /ka/ - here there are two possible explanations: One is that it is simply a different verbalizing stem that is used, the other is that Cora has labialized the /k/ which is not unlikely since Cora frequentlky labializes both /t/, /k/ and /m/ before /a/. 

The word for plant in Cora is wíité which seems unrelated to the *tewa forms - unless the root for planting is -and the -wa is and added stem formant. In which case we might be able to analyze plant as a trimorphemic root *te-wa-ka, "bury" as bimorphemic *tew-aka (dust-enter) and name as bimorphemic *tewa-ka. 

Finally the Nahuatl and Cora forms for "follow" look like they might be cognate, but from a stem tawa- instead of tewa-, but I am not going to explore that possibility now.  

Additionally we should mention that Brian Stubbs reconstructs the PUA word for name with the vowel **ɨ and with a nasal as **tɨnwa. PUA ɨ regularly becomes e in Coracholan and Nahuan, and the Cora form for "to name someone" seems to show a reflex of the nasal.

Notice how interesting and suprisingly beautiful it is that the word should become a homophones that describe both the first and the last ritual in the human life cycle, and the act of planting a lifeless seed into the ground to see it return as a life-sustaining plant.

5. Sun, heat, deities and sacred energy

Having covered now terms of religious and spiritual significance to the life and the soul and to nourishment and the life cycle it enables we turn now to the last set of words, which are no less significant: The words for the sun, its heat, its status as primary deity, and the sacred energy that is associated with it. 

A depiction of a ruler called "tlacateotl" in the Codex Xolotl,
his name written simply with the glyph for teo:tl, a sun.
The Nahuatl word for "sun", "day" and the solar energy that provides humans with their individual fates is called /to:nalli/, and is another patientive noun derived from the verb /to:na/ "to be warm" (about the sun and weather), the adjective "hot" in Nahuatl is /to-to:n-ki/ a reduplicated form of the verb with an perfect suffix (so basically "heated"). The word for "deity" (today mostly the Christian god) and for "sacred" is teo:tl. There is no obvious relation between those two words - except that in the Aztec glyphic script the word teo:tl was written logographically with a symbol depicting a sun. Nevertheless, in the following I shall show a relation between the two, though it is not a direct etymological relation.

In Huichol the word for sun is tau, plain and simple. But the word tau can also be used to describe the solar deity who also has the personal name /tayéu/. There is also a prefix tu- used to describe sacred things tukí "traditional temple building" (tu+house), tutú means "the ancient tradition (i.e. Huichol religion)", tutúma "deity" (any deity in the large Huichol pantheon), and túukah means "midday". 

A Huichol  sikɨri  sun-symbol placed in the landscape.
In Cora the word for "sun" and "day" and "heat" is sɨkah, which seems unrelated to the Nahua and Huichol words for sun, but the word for "God" (the christian, and other deities) is taya'u. The word for sun seems to be related to the Huichol word for a sacred sun related symbol called "the eye of god" in Spanish, but sikɨri in Huichol. If we look under "sol" in the 1732 Cora vocabulary we find out why the Cora shifted to calling the sun sɨkah because it says "Sol - xeucat. Decirle tayoappa es idolatrar porque le decián que era su padre" [Sun - sɨkah. Saying tayoappa is committing idolatry because they said that was their father ]. It seems that the Cora also originally had a name for the sun and solar deity which was probably something like /tayau-pa/ where the -pa is a likely locative suffix, but that missionaries in the 18th century forbade them to use that word.  

The picture is complicated however by the fact that both Cora and Huichol has borrowed a lot of words from Nahuatl, many in the ecclesiastical realm as Nahuatl was the langage used as a lingua franca by missionaries in the 16th and 17th centuries. For example the word for "church" is teyeupáni in Huichol, and teyuhpwa in Cora both borrowed from Nahuatl teo:pan (Knab 1976). Luckily it seems that we can count on the borrowed form of teo:tl to be identified by having /e/ as the first vowel. 

This gives us the following breakdown of probable relations: 


to:- (na)

 At this point we can see that Nahuatl /to:/ is the regular reflex of proto-Corachol-Nahuan *tau, the -na is a verbalizing stem formant deriving the verb "to sunheat" from the noun root.  

The word for deity/sacred energy/name of the Solar deity/God has two syllables: we can assume that *e was the original and that it either god blended with the word for sun, or that the /e/ assimilated to the following vowel if that was /a/. The long Nahua /o:/ could be from either /ew/ or /aw/, so that doesn't provide a definitive clue. We should probably reconstruct the proto-Nahuan form with a /y/ glide between the two syllables as *teyo:-, this glide simple disappears phonologically when found in between a front and a back vowel (the same reason the glide in the Nahuatl verb /piya/ "to have" can only be distinguished when the final /a/ drops in the past tense and the /y/ devoices to [š]).  

This leaves the Huichol shortform tu to be explained: Here I would again invoke the accent phenomenon that causes apocope of unstressed vowels: tú < *taú, in the same way that tuá- "to plant" comes from tewá.    

Stubbs reconstructs a PUA root **tɨyo "deity" which he only finds in the Aztecan branch, I have hereby demonstrated that the root is shared with Coracholan and should be reconstructed as *tɨyaw or *tɨyew. 

6. Conclusions: 

In this post I have basically attempted to solve the relations between life and death, sustenance, birth and the sacred in Nahuatl and Corachol - at least the etymological relations. And I believe that I have shown that the three languages are indeed very closely related, sharing not only a set of linguistic forms derived from the same root, but also a system of cultural conceptualizations of the relatoins between life, death, food and religion. I cannot say that I have demonstrated that Corachol and Nahuatl are more closely related than either group is to other Southern Uto-Aztecan languages (to do that I would need to include those groups in the comparison), but I think that it is fairly improbable that the same degree of phonologica, semantic, lexical and morphological similarity can be found with any of the other UA languages of Mexico. 


  • Campbell, L., & Langacker, R. W. (1978). Proto-Aztecan vowels: part I, II, III. International Journal of American Linguistics44(2), 85-102; no. 3: 197-210;  no. 4: 262-279
  • Dakin, Karen. "Western and Central Nahua dialects." Language Contact and Change in Mesoamerica and Beyond 185 (2017): 263.
  • Dakin, K. (1983). Proto-Aztecan vowels and Pochutec: An alternative analysis. International Journal of American Linguistics49(2), 196-203.
  • Kaufman, T. (2001). The history of the Nawa language group from the earliest times to the sixteenth century: Some initial results. Paper posted online at http://www. albany. edu/anthro/maldp/Nawa. pdf. University of Pittsburgh.
  • Knab, T. (1976). Huichol-Nahuatl borrowings and their implications in the ethnohistory of the region. International Journal of American Linguistics42(3), 261-264.
  • Manaster-Ramer, A. (1996). On Whorf's law and related questions of Aztecan phonology and etymology. International journal of American linguistics62(2), 176-187.
  • McMahon, A., & de McMahon, M. A. (1959). Vocabulario cora. Instituto Linguistico de Verano, Vocabularios Indigenas 2.
  • Stubbs, B. D. (2011). Uto-Aztecan: A comparative vocabulary. Shumway Family History Services.
  • Vázquez Soto, Veronica (2000). Morphology and Syllable Weight in Cora: The Case of the Absolutive Suffix-ti. Uto-Aztecan: Structural, Temporal, and Geographic Perspectives: Papers in Memory of Wick R. Miller by the Friends of Uto-Aztecan, 105.

onsdag den 26. april 2017

The Aztecs did not need a leap year: Introducing the Nuttall-Ochoa model for the Aztec Calendar

This blogpost is inspired by a recent blogpost by Itztli Ehecatl at Calmecac Anahuac, in which he lays out the different calendar correlations between the Aztec/Mexica calendar and the gregorian one, and in which he argues in favor of the correlation recently made by Rubén Ochoa. I agree that Ochoa’s correlation is the best available, and in this blogpost, I describe why that is – recycling some of the arguments from Itztli Ehecatl’s blogpost and adding others.

Depiction of the Celebration of the feast of Tlacaxipehualiztli
during the Spring Equinox in the Codex Borgia (folio 34)

 A recurring question in the study of Aztec calendrics has been whether the Aztec calendar did or did not correct for leap years. Generally the question has been answered in the negative (see e.g. 
Sprajc 2000). Archaeoastronomer and expert in Mesoamerican calendars Anthony Aveni (2016:110), in his recent chapter on calendars in the  Oxford Handbook of the Aztecs, concludes that “there is little evidence that the Aztecs employed a leap-year correction”, but that the question is open.
Rather, I think the question is wrongly posed - who says the Aztec calendar even needed a leap year correction the first place?

The basic problem is that there is little explicit evidence that the Aztecs used an intercalary leap year in their xiuhpohualli solar year count, but also it seems that the monthly feasts, many of which were related to an agricultural cycle and themes were usually celebrated at the appropriate time of year for a given agricultural event. This poses a basic question of how to reconcile the idea of calendric agricultural rituals with a calendar count that cannot be fixed in relation to the tropical year.

But as I see it the question can be resolved quite simply by recognizing, as several scholars have, that the beginning of the Aztec year was fixed by the vernal equinox and that the nemontemi are simply the days between the end of the 360-day cycle and the beginning of the next. This would make the question of leap year moot for the Aztec calendar since the equinox/new year would automatically fix the calendar to the tropical year every year, and the need to have an explicit principle for intercalating days or months due to the increasing disjuncture between the astronomical and calendric year would never arise.

Currently, none of the most widely accepted correlations between the Aztec calendar and the Gregorian adopt this solution to the leap year problem. But the solution emerges naturally from the argument that the vernal equinox marked the beginning of the Aztec year.

Today, the most scholars use one of two correlations – Caso’s or Tena’s.  Alfonso Caso’s correlation does not correct for the bissextile year and so does not begin on a specific day relative to the Gregorian calendar, it would result in the Aztec year starting on a different day each year with the monthly ceremonies gradually changing their position relative to the tropical year.  Rafael Tena’s model, in turn, does correct for leap year, but he argues the year begins on February 26th with the veintena of Atlcahualo. 

Oddly, as noted by Itztli Ehecatl, neither Caso nor Tena cites Nuttall’s 1894 paper or her 1904 article in American Anthropologist in response to Eduard Seler. The argument against the intercalation of an extra day to correct for leap year is that some scholars find that it disrupts the ongoing count of the other calendrical cycles – the 260 day tonalpohualli and the lunar and venus counts. The argument for an intercalary day has mostly been the seemingly correct correspondence with sowing and fertility rituals in the spring months and the harvest rituals in the fall. Some scholars (e.g. Graulich 2002) have argued that the agricultural associations of the monthly feasts were so vague that they didn't really require being performed at a specific time of year (which is debatable), and that farmers don't need a calendar to tell them when to sow and harvest (which is right). 

Nuttall was the first to propose the equinox as marking the end and beginning of the Aztec solar year. Her proposal however has been echoed also by Lopez Luján (2005), who argues, based on the astronomical alignment of the Templo Mayor, and in accordance with Nuttall, that Tlacaxipehualiztli was the first 20-day month of the Aztec year and that its beginning was fixed by the equinox. 

Recently, Rubén Ochoa has verified that Nuttall’s correlation is consistent with the evidence from prehispanic calendric codices, and that the length of the nemontemi must have varied so as to gradually adjust to the tropical year. He notes that the addition of a day is evident in the relation between the two most securely dated events in Aztec history:  the arrival of Cortés in Tenochtitlan on November 8th 1519 (2 quecholli/8 ehecatl) and the fall of Tenochtitlan on August 13th 1521 (15 miccailhuitontli/1 coatl). He notes that counting from the first date to the second requires at least one intercalary day, which he supposes would have extended the nemontemi of one of the two interceding years with one day from 5 to 6 days.  The explanation of Ochoa’s model and the evidence from the codices can be found here and the evidence showing how the model works for the known dates around 1519-1521 is found here.

Nuttall however, followed the colonial chronicler De La Serna in believing that there was a 13-day intercalary trecena every 52 years, when the New Fire ceremony was celebrated. But as noted by Hassig, this would have had the effect of throwing the correlation between the agricultural year and the calendric year off so much that it would be impractical as an agricultural calendar. Hassig proposes that rather than resetting directly to the astronomical year, the New Fire Ceremony reset the year to 7 days earlier than the exact match, which would mean that the total variance between the agricultural year and the calendar would vary over the 52-year period, but without becoming greater than about 7 days which is agriculturally tolerable. The main problem with Hassig’s explanation is that it is entirely conjectural.

What sets apart Ochoa’s solution from those of de la Serna, Nuttall and Hassig is that when following the principle of beginning the calendar year on the day after the observable equionox, no explicit intercalation is necessary, nor are any calculations or record-keeping of a deep day count.  The intercalation arises naturally as the number of days between the end of the 360- xiuhpohualli and the beginning of the next after the equinox naturally varies between five and six days. For the Aztec commoner, the principle is simply that when the last day of the year is over, the waiting period for the beginning of the new year begins, and the waiting period ends when the priests calls that the sun is in the right position. Calendar priests will be aware that every four years there are six days before the equinox rather than the usual five, but since the nemontemi are already out of count and simply fill up the rest of the year, they do not need to make explicit mention of any “intercalation”, they simply keep counting the tonalpohualli day count and then restart the year count on the equinox. In this way, the Aztecs had a smoothly functioning calendar which could function without significant deviation from the tropical year, and which could be observed in all communities that shared the custom of beginning the new year on the vernal equinox, regardless of what month they used as the first in the year.

The Nuttall-Ochoa model has a number of major advantages over all previous calendar models:
  •  It solves the leap year problem without the necessity of explaining how the intercalation of days was accomplished without further confounding any existing mismatch between local calendar variants.
  •  It requires minimal astronomical knowledge to use since it can be maintained as a self-correcting system through the simple cultural practice of beginning the year count the day after the observable equinox, without the necessity of making calculations into the past or future or keep records of bissextile years.
  •  It is congruent with the clear cultural relations between the calendar rituals and the agricultural cycle and the close cultural association between the year and the growth period.
  •  It is congruent with the archeologically attested importance of the equinox, and with the use of westward-oriented double temples that allow for ease of observation of the equinox sun between the two temples, and which symbolically divides the year into two deities who then represent the estival and hibernal solstices.
  • It is linguistically plausible based on the literal and cultural meanings of ‘nemontemi’ (“they vainly fill up”), and ‘xihuitl’ (year/herb/green(-stone)/comet).
  •  It is congruent with what we know about the practices surrounding the Maya solar calendar, including the architectural use of E groups as observation points from where the vernal equinox can be observed relative to the main temple structures.
  • It is congruent with the known correlation dates for the arrival of Cortés in Tenochtitlan in 1519 and the fall of Tenochtitlan in 1521.
      To me however the best argument comes from the Templo Mayor and has been used by Leonardo Lopez Luján to argue that the temple was tied to the equinox and to the calendar: If we assume that the Mexica new year started right after observing the spring equinox (usually march 21st in the Gregorian calendar), then the sun would rise exactly in between the two shrines on top of the Templo Mayor. AND: The feast for Huizilopochtli in Panquetzaliztli the sun would fall around the Winter Solstice when the sun would be above the Huitzilopochtli temple, and the feast for Tlaloc in Etzalcualiztli would fall on the summer solstice (during the rainy season in central Mexico).  With Ochoa's model, but not with any of the others, this would be the same year after year, without any need to discuss how to put the calendar back in sync with the sun. 

This image based on the official model of the Templo Mayor at the Museo Templo Mayor, shows how the sun would be observed from the Quetzalcoatl temple on the Equinox, and on the winter solstices - if Ochoa's model is correct.
Can this be a coincidence?

Finally, as a treat, here is a link to  Calmecac Anahuacs "Aztec Date app" which calculates any given date according to the Nuttall-Ochoa correlation. You can use it to find out what date you were born, married in the Aztec calendar.

Update (28/04/2017): Additional Arguments:

I have been encouraged to also mention the main arguments against the Nuttall-Ochoa correlation, so I will do this here. Traditionally it has been believed that the Aztec year was named after the day-sign (in the tonalpohualli ritual calendar) of the first day in the year. If we follow the Nuttall-Ochoa model, the first day of the year however is not the one that defines the year bearer - rather years seem to be named after the thirteenth day of the new year - this does seem a somewhat arbitrary principle, although thirteen is of course an important number.

Also a final argument in favor of the correlation , although a somewhat subjective one, is that in the Nuttall-Ochoa correlation and none of the others, my birthday falls on the Tonalpohualli day 13 Ozomahtli, about which the Florentine codex writes (in Anderson and Dibble's translation): "It was said that anyone who was born upon this day sign became a highly favored person; he succeeded and endured on Earth.He was well respected and recognized; he was famed and honored; hence he was one who prospered, enjoyed glory, was compassionate, and served others. As a chieftan, he was strong, daring in battle, esteemed, intrepid, able, sharp-witted, quick-acting, prudent, sage, learned and discreet; an able talker, and attentive. To everyone he brought happiness, as much as to comfort the afflicted and provide succor. And if such did not befall one, it was said that he himself, by his own doing, neglected and destroyed his day sign through vice, because perchance he took not good heed, perhaps did not perform the penances well.The closing day sign 13 Ozomatli was named Tonacatecutli. And likewise they said that the one then born would become aged; they said that he would finish his work, endure on Earth, and be admired" (Florentine Codex, vol. 4-5:53-54).


tirsdag den 26. juli 2016

How to spell Nahuatl? Nawatl? Nauatl?

My last blog post was about how the Nahua people wrote before the arrival of Europeans with their alphabetic writing system. But almost all Nahuatl texts from the colonial period onwards are of course written in alphabetic writing. In this blog post, I describe the many different conventions for writing Nahuatl using the Latin script.

For the past 80 years, Nahuatl scholars have argued about how to standardize Nahuatl orthography and what conventions to use. Different groups of scholars and activists have recommended  and supported different systems. Sometimes scholars and Nahauatl activists seem to be spending more time arguing about how to write Nahuatl than they do on actually writing it. There are even cases where a single community has two different lanugage revitalization projects that refuse to cooperate because they use different spelling systems!

In this post, I try to describe the different types of writing conventions that are in use for Nahuatl, and to show their relation to different schools of thought within Nahuatl scholarship.

Roughly we can classify Nahuatl orthographies into two main types, each of which has a bunch of variatoins. One group we can call "Classical orthographies", because they base their orthographic choices on the conventions used by the Spanish speaking friars who wrote the first alphabetic texts in the early 16th century. The other group we can call "Modern orthographies" because they were introduced by academic linguists working to find the most linguistically efficient ways to represent the Nahuatl language in writing in the early 20th century. Both types of orthographies can be used to represent colonial Nahuatl as well as contemporary varieties.

For the current purpose we can define the two types of orthographies in this way:

  • Classical Orthographies: are those that value continuity with the colonial tradition of nahuatl writing - and which adopt colonial conventions because of the value they have as connectors with that tradition.
  • Modern Orthographies: are those that value linguistic efficiency and which aim to represent the Nahuatl language in ways that are either easier to learn or which facilitate a higher analytical precision by representing linguistic elements (phonemes, morphemes) in ways that are minimally variable and maximally efficient.
In practice most orthographies include elements of both "classical" and "modern" principles. 

Classical Orthographies: 

Sometimes people talk about "classical orthography" as if it is a single well-established standard. Really it is not, and it never was. In the 16th century when Nahuatl was first written alphabetically, the idea of a standardized orthography didn't even exist - and there was no established orthography for any of the spoken main languages such as English or Spanish (as anyone trying to read Shakespeare or Cortés' letters will realize). Authors writing in any of these languages simply used the writing conventions they learned from their teachers and put them to the best possible use to get their points across in the easiest way. They tended to write these languages as they were spoken, representing the sounds more or less as they pronounced them. And when they began writing Nahuatl they did the same, tried to use the conventions they knew from writing Spanish to represent the sounds of Nahuatl. This is why the only thing that is really shared by all "classical orthographies" is the fact that they represent the sounds that exist both in Nahuatl and in Spanish using the letters that were most commonly used in Spanish to represent these sounds. For example, Spanish had adopted the Latin convention of writing the sound [k] with the letter <c> before the vowels [a] and [o] but with the letters <qu> before the vowels [i] and [e]. Luckily, actually most of the sounds in Nahuatl are also found in Spanish, which meant that this method was fairly succesful. And in fact in the 16th century, Spanish phonology was even more similar to that of Nahuatl - because at that time Spanish didnt have the harsh j-sound (like in scottish Loch), but instead had a soft sh-sound as in fish which also exists in Nahuatl. They wrote this sound with the letter <x>, because that is how they generally wrote the sh-sound in Spanish. Only over the next century did Spanish gradually change the sh-sound to the harsh j-sound (and eventually began writing it with a j). (This, incidentally, is why the x is pronounced harshly in words like Mexico/Mejico, Oaxaca and Xalapa/Jalapa - but not in the corresponding Nahuatl versions which are pronounced meshi'ko, washakak and shalapan).

However there are some sounds that are found in Nahuatl that do not exist in Spanish: Primarily, the Nahuatl signature sound the tl (written in the International Phonetic Alphabet as  [t͡ɬ]), but this turned out to be easy to write with the letter combination <tl>. The sound [kw]  (as in queen )likewise turned out to be easy to write, since this sound also existed in Spanish as (although in Spanish it is a combination of k and u, and not a single consonant sound) so they wrote it <qu> or <cu>. The Nahuatl consonant [ t͡s] also didnt exist in Spanoish, but the Friars knew the sound from Hebrew and wrote it in the same way they would when transliterating the scripture using the letter combination <tz>. Nahuatl also had the consonant sound [w] (as in "wat?") which was not found in Spanish - friars couldn't quite decide on how to write this one, but usually they simply represented it with the vowel letter <u> - sometimes combined with a consonant letter such as <hu> or <gu> (More about this below, under Canger's orthography).

But the most difficult sounds to write were the glottal stop (or h) neither of which existed in Spanish; and the distinction between long and short vowel duration. At first most friars didn't even realize that these sounds actually existed in Nahuatl, so they simply didnt write them! This is the main difference between the orthographies of the Franciscan friars and the Jesuits.

Franciscan style orthographies:

In the 16th century the most widely used Nahuatl orthographies were those developed by the Franciscans. The Franciscans had a highly practical approach to evangelizing, without too many theoretical considerations - they just did whatever seemed to work (which sometimes got them on the wrong side of other ecclesiastic orders such as the more orthodox Dominicans). The same approach worked in the area of orthography, where the Franciscans never pined much about being consistent or about how best to write. This pragmatic approach was probably partly what allowed the Franciscans friars and their indigenous aides to author the most extensive documentation of  any Indigenous language in the colonial Americas. The 16th century saw major Nahuatl works like Andres de Olmos' Nahuatl grammar, Bernardino de Sahagún's 12 volume encyclopedia about Indigenous Nahua culture (now called the Florentine Codex) and Alonso de Molina's vocabularies. None of these works represented the saltillo or the vowel length distinction, and they were extremely inconsistent in representing sounds like [w] and [j]  - and even so they worked fine and thousands of Nahuas learned to write using these loose conventions. Apparently they didn't loose much sleep worrying about the fact that the representation of some minimal pairs was ambiguous (e.g. [tla:tia] "to hide" and [tlatia] "to burn" both of which was written <tlatia>, or [paʔti] "to become well" or [pa:ti] "to melt" both of which were written <pati>, and even the difference between plural and singular of verbs in the present tense as [kochi] "he sleeps" and [kochiʔ] "they sleep" were both written <cochi>). When Franciscans sometimes heard the saltillo (they only ever seem to have heard it bwhen it occurred before another consonant) and decided to write it, they used the letter <h> giving <pahti> "to cure", <pati> "to melt". 

Other than these conventions Franciscans (and the vast majority of colonial 
Nahuatl authors) were very unruly in their orthographies - for example they used the letters u and o interchangeably for the vowels [o] and [o:], they used the letters <i> and <j> and <y> interchangeably both for the vowel [i] and the consonant [j], they used <hu>, <u> and <o> intechangeably for the consonant [w] and used the letters <z>, <c>  and <ç> for the sound [s]. '

Really, by modern standards the Franciscan orthography was a mess - and yet we are fully able to read it today just as they were back in the 16th century. This tells me that consistency and standardization of orthographies is vastly overrated. 

Jesuit style orthographies: 

Page from Carochi's 1645 grammar which uses macron to show long vowels
and circumflex accent to show wordfinal saltillo.
Among the catholic orders the Jesuits have a reputation for being studious and academically inclined. The jesuit orthographic tradition for Nahuatl embodies this reputation for thoughtfulness, and Jesuits were among the first scholars to have theoretical insights about how the Nahuatl language differed from Spanish and other well known languages and how this ought to influence the way the language was written. Nonetheless, most of the honor for these insights should probably be given to the first Nahuatl grammarian who was also a Nahua person and a Nahuatl native speaker: the jesuit priest Antonio del Rincon. He wrote a short grammar in which he noted the existence of the saltillo and vowel length distinction and to suggest marking it in writing to represent the language more faithfully. His suggestions were taken up a fifty years later by fellow Jesuit Horacio Carochi who introduced a fully developed system for marking vowel length and saltillo systematically. Following Rincon, Carochi used diacritical marks to show these distinctions and he marked the saltillo with an accent (grave, or circumflex) and vowel length with a macron.

Hence the words "to cure" and "to melt" he wrote <pàti> and <pāti> respectively and the difference between "he sleeps" and "they sleep" he wrote <cochi> and <cochî>. Sometimes he marked short vowels with a breve sign <ă>, but he did not do this consistently (since it is redundant to mark both long and short, he only marked short vowels when a long vowel would change the meaning).

The Andrews-Campbell-Karttunen orthography:
Karttunen's dictionary which has
popularized the ACK orthography.
In the mid-twentieth century American historians discovered the rich trove of Nahuatl language writings and began working with them. Among the earliest historians looking at these works were Arthur Anderson and Charles Dibble who translated Sahagun's Florentine Codex into English. Another scholar to take up the study of Nahuatl was the grammarian  J. Richard Andrews, who published his grammar of the "classical" language in 1975. He chose an orthography that was linguistically accurate marking all the phonemes including the vowel length distinction and the saltillo - and which combined aspects of the Franciscan and Jesuit tradition. Specifically he adopted Carochi's use of macron for marking long vowels, and the tradition penchant for marking the saltillo with <h>. He conventionalized the use of <hu> to write the sound <w> before a vowel and <uh> syllable-finally

Andrews' orthography was in turn adopted and conventionalized further by R. Joe Campbell and Frances Karttunen in their Foundation Course and in Campbell's morphological dictionary, and the important dictionary of Frances Karttunen (the first full Nahuatl-English dictionary, and the first to consistently mark the vowel length distinction). This orthography was further adopted by the school of historians trained by James Lockhart who collaborated with Karttunen in the 1970s. Today, almost all new editions of colonial Nahuatl texts adopt the Andrews-Campbell-Karttunen orthography as the standard (although many of them choose not to mark vowel length). 

One problematic feature of the ACK orthography (thanks to John Sullivan for introducing this term which i stole from one of his facebook statuses) is that it uses the letter <h> in three distinct functions - as the saltillo and as a part of the <hu>-digraph used to write [w] and as part of the <ch> digraph. This gives spellings with two consecutive h's such as  michhuacan [mit͡ʃwaʔka:n] (name of the state Michoacan - "Place of Fishowners"), or ohhui [oʔwi]"difficult". And it also creates near-ambiguity in cases where a [k] sound written with <c> precedes a [w] written with <hu>  over a syllable boundary (e.g. cachuia "to provide someone with sandals" where the reader has to realize that the <ch> is pronounced as [kw] and not [ch].) From the point of view of a proponent of a modern "efficiency based" orthography, clusters like <hhu>, <chhu> and <chu> where the letter <h> has a different value, comes across as unelegant and unnecessary - even though it is not technically ambiguous.

The use of <h> for saltillo also has the problem that it makes it impossible to distinguish in writing between varieties that pronounce the saltillo as a glottal stop, and those that pronounce it as an [h] - and it also somewhat implies that the h-pronunciation is the norm, when in fact we know that the normative pronunciation in the Nahua capital of Tenochtitlan was the glottal stop. 

The main advantages of the ACK orthography is that 1. it is very similar to the ortography used for most colonial texts and makes the transition from the study of the grammars (using Andrews and Karttunen's works) to the reading of colonial texts very easy, 2. it marks each Nahuatl phoneme with a single letter (or letter combination) and uses only symbols found on a standard American keyboard. 

Launey's orthography

About the same time that Andrews was working in the US, a French linguist was also working on a major analysis of the Nahuatl language based on the Florentine Codex and on Carochi's grammar. Michel Launey published a full didactic grammar of Nahuatl in French in 1979. He chose to use Carochi's conventions for marking saltillo with diacritic marks, standardizing them, and getting rid of the breve accent on short vowels. Since Launey's work was first published only in French and Spanish, (and a somewhat inadequate English translation in 2011) it mostly gained currency in Europe and Mexico, and among linguists more than among historians. His main work, the 1986 thèse d'etat, still exists only in French. It is to my mind the single best grammar of colonial Nahuatl written - surpassing the work of Andrews, and that of Carochi (francophone readers can check it out here). 

The Carochi-Launey orthography has the advantage that because the saltillo is marked as a diacritic it avoids the collisions of digraphs that are found in the ACK orthography, and it avoids implicitly suggesting the pronunciation <h> as the way to pronounce the saltillo.

Canger's orthography

Una Canger is a Danish linguist (and my first Nahuatl teacher) and writes in many different orthographies - this is because she works with many contemporary varieties and adopt the conventions that work best with the variety and its speakers and her own linguistic sensibilities. For the writing of Nahuatl she has made one important proposal. 

In a 2011 article, Canger described the how it happened that Nahuatl grammarians ended up writing the sound [w] with the letter combination <hu>. She shows that the tradition originates with the Franciscan Andres de Olmos - but she also shows that he did not always write the phoneme [w] as <hu> or <uh> - in fact he mostly did this when the [w] followed a consonant or preceded a word boundary or a consonant in the subsequent syllable. When the u was For example he wrote the word [siwatl] "woman" as <çiuatl>, but the word [yeʔwatl] "he/she/it" he writes <yehuatl> and the word "my wife" [nosiwaw] he writes <noçiuauh>. This leads Canger to suggest that what Olmos was doing was that he was using the <h> to show to the reader that the <w> is pronounced differently when it occurs wordfinally or before or after a consonant than when it occurs between two vowels. In fact drawing on her knowledge of contemporary Nahuatl, Canger suggests that it is exactly the aspiration that often accompanies the devoiced variant of <w> that Olmos was representing with the h (this argument is also strengthened by the fact that Olmos also writes h after the letter l in the same positions - since l also devoices under those conditions). Canger then shows that subsequent grammarians adopted Olmos convention of using hu without understanding the way that he used it, and instead of writing only devoiced w with h they used it across the board. This confusion is the ultimate reason for the problematic digraphs found in the ACK orthography and other orthographies that use the <hu> convention. Instead, Canger suggests returning to Olmos original principle - representing the [w] sound with the vowel letter <u>. Hence Canger does not write "nahuatl" but nauatl (as did Olmos and many colonial authors) or else nawatl writing the [w] as <w>. 

While it seems that Canger would prefer a more modern orthography using k and w and s instead of c/qu and u and z/c, she suggests that also scholars who prefer a classical style of orthography ought to return to writing [w] as <u>. Canger's proposal shares all the advantages of the ACK and Launey orthographies - and avoids the problematic digraphs combination found in both of them. The main drawback is that the other orthographies are already in wide usage and that by now it will be quite hard to get people to start writing Nauatl instead of Nahuatl.

Comparison of "Classical orthographies":

”we do it”

Modern orthographies: 

"Modern" orthographies also differ among themselves - but they share the principle that they aim for maximal efficiency rather than maximal continuity with colonial writing traditions. But efficiency can be measured in different ways that are not always compatible. One criterion of efficience might be simple graphic efficiency to have the smallest and most parsimonious array of graphic units  - for example following the principle of "one phoneme - one letter". This kind of "phonemic efficiency" would prefer to remove all the digraphic letters (tl, tz, ch, kw) so that no letter is used to represent two different phonemes and no phoneme is represented by two distinct typographic units. Another kind of efficiency would be to make sure that the orthography is maximally easy to learn - a kind of "pedagogical efficiency". Another kind of efficiency is to make sure that the orthography is maximally accessible to linguists - for example by using symbols for sound values that are internationally established in the linguistic community (e.g. in the same values as in the International Phonetic alphabet, or the Americanist Phonetic Alphabet). 

The Americanist orthography

The Americanist orthography stems from the earliest studies of contemporary spoken Nahuatl by American and Mexican US-trained linguists in the first half of the 20th century. They tended to use a phonetic notatoin system now known as APA (Americanist Phonetic Alphabet), which aimed towards being strictly phonemic and based on the principle of one letter per phoneme. Hence they used single letter symbols for all of the sounds that the "classical" orthographies represented with digraphs -

APA style transcription key:
  • tl = ƛ
  • tz = ¢ (or sometimes c)
  • cu/qua = kw (or sometimes q)
  • ch = č
  • x = š
  • c/qu = k
  • hu/uh = w
  • h = h
  • ʔ = '
Americanist orthography is really very efficient in this way - except that it requires a bunch fo special symbols not found on ordinary keyboards. Hence many linguists taking a practical approach retained tl, ts, ch and x avoiding unnecessary additional signs apart from those already on a standard keyboard.

Such a modified Americanist orthography was in fact adopted as the official standard by the participants in the first Aztec Congress which was held in Milpa Alta in 1940 and attended by many native speakers. They stated that they prefered this orthography exactly because it didnt use the Spanish-style digraphs que/qui ce/za etc. In this way the choice of a "modern" and "scientific" orthography was a political move towards decolonization. Today a variant of this orthography (without the special symbols, but with k and w) is used by most Nahuatl speakers in the Zongolica region where the linguist Andrés Hasler has promoted it for several decades. 

SEP and SIL's orthographies

Example of the SEP/SIL orthography from a textbook.
It uses the letter j to represent the h sound.
In the 1940s an American missionary organization called Instituto Linguistico de Verano (Summer Institute of Linguistics or SIL) began collaborating with the Mexican Ministry of Education (SEP) to develop educational materials in indigenous languages. Because the government wanted indigenous peoples to learn Spanish and primarily wanted to use indigenous language education as a way to teach Spanish, they considered that the orthographies should only use letters already found in Spanish. They developed many different orthographies for different Nahuan varieties - some of which used the "classical" Spanish digraphs, and others which used k and w (although most use hu or simply u for [w]).  The only common denominator seems to be that they use the letter j for the saltillo when it is pronounced as an [h]. This presumably is because the letter <h> is "mute" in Spanish which migh confuse the children during the gradual transition from Nahuatl to Spanish. Today SIL still consider the ease of acquisition for students who are already literate in Spanish as the main criteria for efficiency. Most SEP/SIl orthographies do not mark vowel length, because most Nahuatl speakers are not actually aware of this feature of their language, and vowel length is not very important in distinguishing words from eachother. Some of them however do and when they do they tend to use either double vowels (aa/ee/ii/oo)  or underlining (a, e, i, o) to mark long vowels.

SEP and SIL style orthographies are extremely influential in Mexico and those Nahuatl speakers who have been lucky enough to have classes in their languages in school are likely to have learned them. Also most Nahuatl language authors tend to use these orthographies (because most of them are trained as bilingual teachers through SEP). Many SEP and SIL orthographies also do not write the double [ll] sound which is very common in Nahuatl but instead writes it as a single l. This is of course because Mexican Spanish pronounces the double l as the [j] sound (which is written with y in Nahuatl).

The drawback of using <j> to write the sound [h] is that it often causes non-Nahuatl speakers to erroneously pronounce it as the harsh Spanish j-sound and not as a soft h-sound. Writing the double l as a single-l is problematic from a grammatical viewpoint because the double-l is what happens when the absolutive suffix -tli occurs after a root ending in -l. So by writing only a single l, the grammatical structure of the language is obscured making it harder to teach the grammar. 

"Intuitive orthographies"

Example of intuitive orthpography form a kindergarten in Hueyapan. It says
"xi nech ate ki an xinech pojpua nochipan kion kual le ni koponis"
which usually would be written as
"xinechateki an xinechpojpoa nochipan kion kualle nikoponis"
which means
"Water me and weed me, that way I will always bloom"
Most Nahuatl speakers have not received any education in Nahuatl and many have never even been aware that their language can be written down. Usually Mexican schools teach only Spanish. This means that they have to invent new conventions almost from scratch (or based on Spanish) when they start writing their language since they havent been taught any of the existing orthographic conventions. These new Nahuatl-writers tend to adopt ways of writing that are intuitive to them based on their knowledge of Spanish and sometimes English orthography. Such intuitive orthographies can be seen on the internet where Nahuatl speakers sometimes converse in writing withouth ever having been taught how to write their language. These orthographies are often  sinmilar to the SEP orthographies (using j for h) - but two new features that are not used in any of the established orthographies are often found in intuitive Nahuatl writing. One is that they often use <sh> instead of the traditional <x> to write the sh-sound. This is probably because most Mexicans are associate this sound with English, and know that it is written sh in the English orthography. The second is that they often write grammatical prefixes as separate words, instead of fusing them together as Nahuatl grammarians do. This is probably because they often think in Spanish when they write in Nahuatl translating from Spanish into Nahuatl and therefore isolating elements of meaning the way it is done in Spanish.

Comparison of Modern Orthographies:

”we do it”
tic chiua
shi tsecuini

Note that the SEP1 and SEP2 and the "intuitive" orthography are just possible examples, but many different combinations of the different choices exist.

So Which One Should You Use?

There is no objective answer. Each orthography comes from different ideas about what is important, and is used by certain communities working within specific genealogies and traditions.

That fact of the matter is that regardless of which orthography you use someone will inevitably tell you that you are using the wrong one. I think the best approach is to learn to read all of them and to use one consistently. But really as I noted consistency is overrated. Shakespeare and Chaucer and Cervantes were able to found their national literatures without using standardized orthographies. Molina and Sahagun were able to found Nahuatl literature without one. The important part is that we keep writing and reading in Nahuatl.

Now I have told how one writes in the Nahuatl language
aʃka:n onikiɁtoɁ ke:nin se:  t͡ɬakwilo:s i:ka nawat͡ɬaɁtolli
aška:n oniki’to’ ke:nin se: ƛa’kwilo:s i:ka nawaƛa’tolli
axcan oniquito quenin ce tlacuiloz ica nahuatlahtolli
axcān oniquìtô quēnin cē tlàcuilōz īca nahuatlàtolli
axcān oniquihtoh quēnin cē tlahcuilōz īca nahuatlahtolli

axcan oniquijtoj quenin se tlajcuilos ica nahuatlajtoli

axkan onikijtoj kenin se tlajkuilos ika nahuatlahtoli
axkan onikihtoh kenin se tlahkuilos ika nawatlahtolli
ashkan onik itoh kenin se tlacuilos ica nahua tlajtoli

(Note: This post was edited on July 31st 2016 to make some minor corrections to the section on the ACK orthography based on comments from Frances Karttunen on the Nahuatl-l listserver)