The etymology of 'heart' in Nahuatl and Southern Uto-Aztecan

Much of the study of "Aztec philosophy" builds on the notion that Nahuatl words have meanings that go beyond mere reference, and which if we analyze them, can tell us about deep underlying philosophical systems of Nahuatl speakers. This blog post is not about the validity of this way of thinking in general - but about one specific word and how its etymology may or may not be able to tell us something about how ancient Nahua and other Uto-Aztecans conceived of one specific, and highly important, aspect of the world. 

Heart removal depicted in the Codex Tudela.

In his 1956 Filosofía Náhuatl, Miguel León-Portilla supports several key arguments by reference to such deeper meanings, which he finds not only in the way the words are used in the Nahuatl texts that he analyzes, but also in his ideas about word origins. One of those words is the word yōllōtl 'heart' and its related verb yōli 'to live' (and the noun derived from this verb yōliliztli 'life').

 León-Portilla wrote:

" YOLLOTL: corazón. Como derivado de ollin: "movimiento", significa literalmente en su forma abstracta yollo-otl "su movilidad, o la razón de su movimiento" (se entiende del viviente). Consideraban, por tanto, los nahuas al corazón como el aspecto dinámico, vital del ser humano. De aquí que la persona sea "rostro, corazón". Posiblemente por esto mismo en la concepción místico-militarista de los aztecas se ofrecía al Sol el corazón, el órgano dinámico por excelencia, que produce y conserva el movimiento y la vida." (León-Portilla 1956:396)

Here, Miguél León-Portilla claims that word for 'heart' is derived from another Nahuatl word, ōlin 'movement' so that the heart is "the mover" (he adds an l erroneously, the root is ōl- and when it takes the -in absolutive suffix it does not add another l, and the form with the -tli absolutive suffix ōlli means not movement but 'rubber'). And he builds a good deal of his understanding of the root yōl- on the idea that its origin expresses dynamic movement, the palpitation of the heart and the locomotion of the creatures that it animates. It seems reasonable and not at all odd that the notion of life and living should be associated with the ability to move about - but there are several problems with the claim that the root yōl- 'life' is "derived" from the root ōl- 'movement'.

Etymology and its Evil Twin

A major problem is that there is no known process in the grammar of Nahuatl that would allow such a derivation. There is no general proces that allows one to take a root beginning in a vowel and us it to produce another related root by simply adding a y- to the beginning. In linguistics, the concept of derivation is usually used for describing such grammatical process by which a word can be coined by applying different grammatical processes. But a broader understanding of words "deriving" from other words, refers to etymology, that is the process through which words change their form and meaning over time, so that words can be related to each other because they share an origin in the same historical root. Positing that the root ōl- could have been the etymological origin for the root yōl- is a different claim, and does not require us to explain the change by reference to a grammatical rule. But etymology has its own rules, and its own standards of argumentation - León-Portilla does not make any argument, he merely claims the relation. 

The most fundamental principle of etymological argumentation, is that the mere fact of two words having similar forms is not a valid argument for their being etymologically related. Languages all work with a relatively small number of phonological building blocks, and by sheer mathematical necessity words will end up looking alike without being related. 

The second fundamental principle is that even though we can imagine a meaningful semantic association between the meanings of two words that look alike, this is also not in it self a sufficient argument for positing an etymological relation. The human mind is a machine made for creating relations, it is what it is best at, it is what makes both language and thinking possible. But it also means that we cannot trust our ability to find associations that are objectively meaningful, precisely because we are so enormously good at making them up ourselves. 

The eternal enemy of the etymologist is the folk etymology, the evil twin of the true etymology; and the eternal fear of the etymologist is to fall into the embarrassing trap of producing or reproducing folk etymologies ourselves. A folk etymology, is of course a popular explanation of a word's origin that makes sense because it says that two words that sound alike are related through some relatively reasonable semantic association. We could make one right now, just as an example: We could say that  the english word 'female' arose because women are like males except they are feeble, and that therefore ancient English-speakers began distinguishing between 'males' and 'feeble males', which over time was contracted to become females. This is of course patently wrong, but it is easier to understand and more socially meaningful in our gender-difference obsessed world than the real explanation, which is that female is a loanword into English from the Old French femelle, which is derived from Latin fēmina 'woman' (whereas male is loaned from Old French masle, derived from Latin masculus 'manly'). The strength of folk etymologies is that they tend to play into preexisting ideas about the world (such as gender stereotypes in this example), so people are inclined to want to believe them, and if they are offered through some sort of authority, they will be even more so inclined. This makes folk etymologies obnoxious to the etymologist, and very hard to eradicate, assuring that etymologists will never run out of work. 

Valid Etymological argumentation

To make a valid etymological argument, showing that words are similar in form and have meanings that can be associated is not enough. Instead the etymologist has different ways of making an argument that actually produces likely etymologies. I say 'likely' here rather than 'true', because it is important to realize that etymology is not a "hard science" but rather more of an art form, which like history writing produces stories about the past that are always just one interpretation of the set of historical facts at hand, and which is limited by the always imperfect and partial nature of these facts. A good etymology begins as a hypothesis, which must be formulated in such a way that it can either be supported or falsified by the facts at hand. For languages with a long writing tradition, we may have many facts at hand; we may be able to see the change of meaning or form from one word to another documented in real time in the historical record (e.g. we can see the development from wȳfman 'wife-man' to 'woman' in the Old English record), or we may even see when a word is coined and have an explanation of why it was coined by the person coining it (For example the word 'utopia' coined by Thomas More in 1515). This would make for a very strong, almost unassailable (almost, because sometimes several people claim to have coined a word, that is documented before either of them having used it), support for an etymological hypothesis . This, however, is almost never the case - and when dealing with languages that have a shallow history of written documentation, it tends to get considerably more difficult to support hypotheses. And consequently more attractive to posit etymologies based on synchronic analysis (i.e. interpreting words as compounds of roots we already know).  

The best method for making etymologies in languages without long histories of writing, is by using the comparative method to reconstruct the developments of words from an earlier stage into their current forms. This method only works however, when we have cognate forms in several related languages. We can say for example, even without the written record, that the English word 'deer' is not derived from the word "dear", because we know that in related languages there is a cognate word that means 'animal', for example Swedish djur and Danish dyr. And we know that in English the earlier form was dēor, and we know that following regular sound changes these could all come from an original form *deur. This supports a hypothesis that in English the original word was restricted to mean 'game animal' and then subsequently the most common or prototypical game animal - the deer. If we ignore the textual evidence it would be possible that the opposing hypothesis could be true, that because they are so cute and valuable, they were 'dear' to the old Anglo-Saxons - but to contradict the evidence from cognate forms in other languages, we would need an even stronger piece of evidence to assert this. 

Good etymology in this way, poses multiple hypotheses for a word's origin, and collects evidence for and against all of them, until it finds the strongest one. 

A good etymology, in this way can be supported by being linked to cognate forms in other related languages through systematic sound changes, and principles of change that we have established independently. Interestingly, precisely because languages change, many times the best etymology is a counterintuitive one, that requires specialized knowledge about language change to see. This is especially the case when dealing with languages that have many homonyms and near homonyms because they have undergone a change that reduced the number of sounds in the language. Nahuatl is such a language - where many contrasts that previously existed in the ancestor language were neutralized. 

The ōl-yōl folk etymology

Se yōllōtl ōlintika
"a heart is moving"

Having now pontificated a bit about etymology, the point I am trying to make is that the ōl-yōl proposal is most likely a folk etymology, and one of a particularly pernicious sort that has been repeated so many times that it becomes self-perpetuating. 

In his magnum opus on Nahua views of the body, Lopez Austin follows León-Portilla's idea that the yōl- words are derived from ōl- words, and expands it into an elaborate "etymological" family tree that adds most of the words in Nahuatl that include the syllables ol or yol - including words such as ōlōtl 'corncob', ololli 'ball, sphere', ololoa 'roll something, wrap something, gather something'. But where León-Portilla sees movement as the conceptual "root" from which the other concepts are derived, López Austin sees the concept of something round, ball-shaped as the original meaning. So that to López Austin the heart is not the "mover" but the "ball" of the body. This allows him to add round maize kernels and the round corn cob center to the set of words he considers to be related. López Austin also does not provide actual etymological arguments, other than the way the words fit together with his view of how the Nahua concepts about the body are related. 

Here is a diagram contrasting how the two proposals of León-Portilla and López Austin each link together the chain of semantic associations of this folk etymological edifice: 

In an influential 1982 article on Tlazolteotl, Thelma Sullivan perhaps followed Lopez Austin (though she cites neither him nor León-Portilla),  (1982:29) and made the same claim and derived yōllōtl and tlaōlli 'maize kernels' from the root ōl-. She also made no argument for how this "derivation" came about or what evidence we have for it having occurred. 

Later again Maffie (2014:188-196) cites Sullivan and Lopez Austin as authorities for the etymological relation between ōl- and yōl- and likewise makes it a key element in his understanding of how ancient Nahua people conceives of the world and of life as characterized by movement and change following a curved movement, which he sees as central and unique to Aztec metaphysical conceptualizations. Interestingly Maffie sort of fuses the López Austin and León-Portilla proposals so that instead of either roundedness or movement at the root, he says that a particular kind of circular or curved or spiral movement that he calls "olin-movement". Maffie also tells us that this corresponds with a proposal by Eduard Seler in 1904 who argued that the root ol- meant rolling movement, also combining the notion of round and movement at the same time. 

It is easy to see why these hypothesis of word relatedness are attractive for seeking deeper meanings of Nahua thought, progressing from a basic shape or processes to physical objects to metaphorical abstractions that include culturally salient concepts and motivate philosophically interesting syllogisms. But nowhere is there an attempt to see whether these words have cognates in other Uto-Aztecan languages that could either support or falsify these etymological hypotheses. 

So now the task of evaluating the hypothesis that these words are related falls to me. 

Words for Heart and Breath in Uto-Aztecan languages

In this section, I go through the cognate sets for the different Southern Uto-Aztecan subgroups one group at a time, reconstructing in this way the intermediate forms. Then in the end, I see how they fit together. 

One important thing to note before hand is that apparently the words for 'heart' and 'breath' are closely related in Uto-Aztecan, so that in many branches we see the same root meaning both or heart in one language and breath in the other. This suggests to me that it really refers more to what Lopez Austin calls an animating principle or force, and that this could be alternately seen as being more closely related to the breath or to the heart.  

Corachol-Nahua: 

In a previous post in 2017, I wrote about the relation between words for 'to live' 'heart',  'life' and 'life force' in Corachol and Nahuatl. This was one of my very first engagements with the comparing Corachol and Nahua. I reconstructed the original form for the words for 'hart' and 'life' as *yauri, but I have revised this reconstruction a bit. First of all the u was a mistake, it would have to be an o, since Corachol u comes from *o. Secondly, I have realized that even though it looks like the Cora and Nahua forms would not have had the initial i- syllable found in Huichol iyaari, but really it could well have been there, but fused to the following glide (in Nahua, the sequence iya often becomes ia, which can then become ya). So it would be possible to reconstruct the shared corachol-Nahuan root as *iyaori, giving Wixárika iyaari, Náayeri rúuri and Nahua yōl-. Recently it dawned on me that in addition to Nahua yōl- in fact the root for 'breath' ihiyō could be derived regularly from the same root, simly by assuming that when used in the sense 'breath' it had the accent on the first syllable *íyao(ri) (then reduplicated to *ihíyao(ri)), and when used in the sense of 'heart' it had the accent on the second syllable *iyáori which in fact works really well with the suggestion that the initial i fused to y when not stressed and was kept when stressed. 

proto-Corachol-Nahuan: *iyáori 'heart'

Wixárika: iyaari 'heart, spirit, soul' (here it levels the *ao diphthong to aa)

Wixárika -yuuri 'be alive'  (here it levels the *ao diphthong to oo) f

Náayeri: ruuri  (in Corachol *y becomes cora /r/ before back vowels)

Nahuatl: yo:li  'life'  (from *yoori)

Nahuatl: yollo- 'heart' (from *yoori-yo)

proto-Corachol-Nahuan: *íyao (or perhaps *íyaɨ) 'breath'

Nahuatl: ihiyo- 'breath' 

Wixárika: iiyáte 'lungs, breath'

Náayeri: í'iyeh 'breathe havily'

Cahitan: 

In the Cahitan languages, Yoeme/Yaqui and Yoreme/Mayo, the word for 'heart' and 'life force' or 'spirit' can be reconstructed as *hiyapsi - a somewhat weird form with the ps cluster, that is highly uncommon in Uto-Aztecan and in Cahita. Clusters like that usually only occur when a vowel is lost by syncope, so probably we should really reconstruct as *hiyapVsi. The same root, but with a final a means 'to be alive'.

Finally, there is an interesting possibility that the s-sound really reflects a previous *r, because we have examples of r devoicing to s - precisely when occuring before another consonant. For example in Corachol 'plum' Náayeri kwaspwá, Wixárika kwarɨpa, where the original *r becomes s in Náayeri after the loss of the unstressed vowel ɨ. If we admit this possibility, we could also reconstruct *hiyapVri for proto-Cahitan (and as we will see when comparing further, we might even have reason to think the p itself comes from an original w). As for the quality of the lost vowel we can assume it was an o, since Corachol-Nahuan shows an o in that position, and either a *p or a *w would have disappeared in this position in proto-corachol-Nahua.

pre-Cahitan:  *hiyaposi/*hiyapori, *hiyawori/*hiyawosi

Yoeme: hiapsi

Yoreme: hiyepsi

Ópatan:

We have two forms for Ópatan languages, hida in Teguima and hibés in Eudeve. In Ópatan d comes from previous y and Eudeve b comes from previous kw - so apart from the initial hi- these forms do not look cognate. It is interesting that the Teguima form accords with the first two syllables of the Coracholan and Cahitan forms begining in *hiya. So perhaps a truncated version of the same form as in Cahita? The Eudeve form could suggest coming from a root *hikwɨ - which accords both with the Hopi and Tepiman forms I give further below, but could also potentially be a loan form something like Yoreme hiyepsi.

Pre-Ópatan: *hiya - (a truncated form of hiyawo(ri)?)

Ópata/Teguima: hida 'heart' (Ópatan /d/ comes from previous *y, so perhaps related to Raramuri: hiya 'be in a hurry, pay attention to something')

Eudeve: hibés (Perhaps a Cahitan loan or cognate to the Tepiman/Hopi forms with hikws?)

Tepiman: 

The Tepiman forms are very interesting because they look like a compound of two roots (and because they display the cool sound changes in this branch of SUA). In Tepiman b comes from previous *kw, d comes from previous *y, and g comes from previous *w (and initial h is lost, and word final vowels are lost). So the form *ii'bɨdaga reconstructed by Burt Bascom, suggest a pre-Tepiman form *hi'kwɨyawV. This is intriguingly close to the hiya forms, but with the weird kwɨ-syllable in the middle. As it happens several other UA languages have the form *hikwV or *hikVw in the word for 'breath', and this does seem to be the basic meaning of that root also in Tepiman as evidenced by the forms ʔiibhu and ibɨ-kɨi. 

One might propose that the original form for all of the words for 'heart'  was *hikwɨ-yawa-(ri) where the first element is the root for 'breath', but that the non-tepiman languages dropped the kwɨ- syllable. This kind of syllable-dropping usually happens in two steps 1. Syncope of the weak vowel > *hikw-yawa-ri and 2. Deletion of the first consonant in the resulting cluster > *hi-yawa-ri. If this is the case then all of the SUA branches share a single ancestral form, and a basic split between Tepiman and non-Tepiman sub-groups, and the Corachol-Nahuan languages nested as a subgroup under the non-Tepiman SUA languages. 

Proto-Tepiman: *ii'bɨdaga 'heart' 

Tohono O'odham: iibɨdag  'heart'

North Tepehuán: ʔiibdag 'heart'

Tepiman loses initial h, and drops final vowels, Tepiman /b/ comes from  PSUA *kw, /d/ comes from PSUA *y, /g/ comes from PSUA *w - so the sequence /dag/ comes from *yawV, and the sequence ib from *hikw. 

Tepiman: *iibɨ 'breathe' < pre-Tepiman *hikwɨ 'breathe'

Tohono O'odham ʔiibhu 'breathe'

Nporth Tepehuán: ibɨḱɨi  'breathe'

Other UA:

Below are some of the cognates found outside of SUA in the languages of California and in Hopi. The root does not seem to appear in Numic languages. Interestingly  Warihío seems to share the root *hika with the Californian languages in the meaning 'heart/spirit'. Hopi has a root that is quite close to what looked like the pre-Tepiman root for 'breath' *hikwV with a suffix -si. 

Californian  *hika

Serrano: hik  'breathe' be alive'

Warihío: hiʔká, iʔká 'heart, spirit'

Tübatülabal: ihk(-ɨt) 'breathe'

Californian: *hikawis

Kitanemuk: hikaw 'breathe'

Luiseño: hakwís  'breathe, be alive'

Cahuilla: hikus 'breathe be alive'

Proto-Hopi: *hikwsi  (from *hikawis?)

Hopi: hikwsi 'breath'

Stubbs (2011) reconstructs this set as #301 *hikwis, and also reconstructs #1166 *ikwiyawa for proto-Tepiman, and #13 *yoLi for Cahitan, Corachol and Nahuan. Hill (2020) has a set of words meaning 'breathe' or 'heart' under Hi-03 - but doesn't include the Nahua and Corachol words. I believe I am the first to propose a connection between these sets, and to demonstrate how they fit together. Also both Stubbs and Hill propose Cahitan *hiyapsi as a relative only of the words meaning breath iyaari/ihiyotl, and they do not explain the -psi ending. 

Intermediary Reconstructions

This gives us the following intermediate forms: 

Corachol+Nahuan shares the form: *yoori

Wixárika+Nahua shares the form *iyaori

Corachol+ Cahitan (and perhaps Ópatan and Raramuri): shares the protoform *hiyawori 'heart'

Pre-Tepiman *hikwɨ-yaw suggests that the form *hiyawori in non-Tepiman SUA has evolved from that same form through syncope, 

giving us a shared form for all of SUA:

PSUA: Shares the proto-form *hikwɨ-yawo 'heart'

The first element of that form appears to be a general PUA form *hikVwVs 'breath'.

Below, I show the development I propose in diagram form:

This comparative evidence alone strongly suggests that the Nahua word for 'heart' is not derived in any way from a root ol- either in the meaning 'movement' or 'ball'. The word is inherited from earlier stages of Uto-Aztecan and developed following a path of systematic sound changes. 

The fact that we can also reconstruct the Nahua words for maize/corn cob tlaōlli/ōlōtl to the SUA verb *hora 'de-grain corn' and the word for movement ōlini and probably also rubber ōlli to a Corachol nahuan root *oro 'to move', strongly suggests that these are independent roots, going back at least to the split of Southern Uto-Aztecan, which would be several millennia before there was anything such as Nahuatl. 

Conclusion: 

The main point of this long etymological exercise is to caution the student of Nahuatl against supporting their arguments about cultural conceptualization with folk etymologies. Being very proficient at Nahuatl and understanding its grammar is not a guarantee against falling into this trap, León-Portilla, López Austin and Sullivan, were all very accomplished nahuatlahtoque. Being good at Nahuatl grammar and knowing all the ways in which words can be constructed, may mislead us to think that all Nahuatl words are constructed through the grammatical principles of for word construction that we know. But this assumes that Nahuas somehow started with the grammar and then began cosntrucitng words with the rules. When of course the reality must have been the reverse - the ancient Uto-Aztecans had a set of lexical roots and a set of grammatical rules that they passed down through the generations, changing the phonological and grammatical rules. Preexisting words where then changed to fit to the new phonological grammatical rules (for example adding suffixes). Nahuas have always had a word meaning 'heart' (or 'breath' or 'life-force'), they never had a need to construct such a word from the word for 'ball' or 'movement'. 

Etymology is a specialized field of knowledge, and building one's big theories of Nahua culture on etymologies without using this knowledge, amounts to constructing fancy castles on sand.

Bibliography:

León-Portilla, M., 1956. La filosofía náhuatl. Instituto Indigenista Interamericano.

López, Austin, Alfredo, 1980. Cuerpo humano e ideología. Las concepciones de los antiguos nahuas. Vol 2. México: Universidad Nacional Autónoma de México, Instituto de Investigaciones Antropológicas.

Hill, Kenneth. 2020. Wick Miller's Uto-Aztecan Cognate Sets. UC Berkeley: Publications of the Survey of California and Other Indian Languages https://escholarship.org/uc/item/9px6p8h8

Maffie, James. 2014. Aztec philosophy: Understanding a world in motion. University Press of Colorado.

Stubbs, Brian D. Uto-Aztecan: A comparative vocabulary. Flower Mound, TX: Shumway Family History Services, 2011.

Sullivan, Thelma. 1982. "Tlazolteotl-Ixcuina: The Great Spinner and." In The Art and Iconography of Late Post-Classic Central Mexico: A Conference at Dumbarton Oaks, October 22nd and 23rd, 1977, p. 7. Dumbarton Oaks.