fredag den 12. september 2014

Nahuatl Oligosynthesis and the etymology of ”pachtli”


One of the amazing features of Nahuatl is the many different ways that words can be modified and derived from each other. This feature has amazed and baffled linguists since the earliest studies of the language. But since there are so many possible ways of deriving words from other words sometimes the question arises which of two word is derived from the other.

In this blog post I examine a question posted to the Nahuatl-l listserv regarding the etymologies of the word pachtli meaning ”moss, hay, mistletoe, rags, patch” and the verbs pachoa ”to move two things close together/to press two surfaces against eachother” (and seemingly metaphorically for a hen to sit on her eggs, and to govern or oppress someone) and the intransitive version pachiwi meaning ”to move close to one another/to collapse into itself/to settle (physically or emotionally)”. The noun and the verbs, seem to be obviously related. One possibility is that the verbs are derived from the noun so that the original literal meaning of the verbs is to ”to do like moss” or something similar to that. The other possibility is that the noun is derived from the verbs so that the original literal meaning of pachtli was ”something that is pressed against a surface”. Both options are possible since in Nahuatl it is possible to derive verbs from nouns by adding the ending -oa or -iwi, but it is also sometimes possible to derive nouns from verbs by removing the verbal ending and adding the absolutive ending -tli.

I believe the neither is the case, and in the following I will argue for a third interpretation – namely that pachtli and pachoa/pachiwi are all derived from a proto-Nahuan locative morpheme /*pa/ meaning something like “on top of”, and a proto-Nahuan verb root /čiwa/čiwi/ meaning basically “to move in X way”. (Note that in this post a phonemic representations are between //, an asterisk * before a word means that it is a proposed reconstruction,   in phonemic writing I use the symbol č for ch and w for hu/uh)

The Idea of Oligosynthesis

First I have to mention the concept of oligosynthesis. This concept was formulated by the American linguist and engineer Benjamin Lee Whorf who was among the first to begin to study Nahuatl as a living language, by interviewing speakers rather than just reading the old grammars. Whorf also had an expansive mind and was inclined to abstract thinking and seeing logical patterns and associatioins where others did not. When studying Nahuatl in the 1930s he found that many words with similar meanings looked similar, and that it seemed to be possible to analyze many longer words into short single syllabic morphemes with very abstract and general meanings. For example the way that the syllable /a:/ can refer to pretty much anything having to do with water or liquid. He argued that probably it was possible to analyze Nahuatl as having only a small number (perhaps in the tens) of basic morphemes from which all other etymological roots could be derived. He proposed defining such languages with small numbers of morphemes and a plethora of dervational processes as ”oligosynthetic” (from the Greek oligo = few, and ”synthesis” ”put together”). As Whorf continued to work on Uto-Aztecan historical linguistics he seems to have abandoned this idea and he never published anything about oligosynthesis. The only reason we know about it is because a manuscript about it exists on microfilm among the Whorf papers at Yale University. Whorf probably abandoned the idea of oligosynthesis as a basic typological category with good cause, it does seem unlikely that any natural language would ever be based on just a couple of hundred morphemes. Nonetheless, I do think the feeling of words being eerily similar and that they might be related by some abstract concept that just seems to escape us is familiar to most people who work with Nahuatl. When for example one notes that ”an arrow” is mitl and ”to pierce something” is mina it feels as if there is some relation, although we do not know of any current morphological processes that could derive mitl from mina or vice versa.

I personally believe that what we look at when we note these similarities is a sign of something that is not too far from what Whorf proposed. I think that it showed that the proto-Nahuan language must have had more derivational processes than it has currently, and that words could be formed by combining short stems. And not only was this possible, it must have been practiced very avidly by the speakers of the language, so that they generated families of words based on the same morphemes – words that stayed in the language after the grammatical processes they used to form them were no longer productive.

The problem with this belief is that it is very difficult to argue for, because it requires a lot of speculation about possible semantic and derivational relations between words, and what looks similar to me may not look that similar to someone else. It is hard to be convincing. So I hope that in the following I will be pardoned for speculating a bit, and that my reader will be prepared to entertain some of my speculations before dismissing them. My chosen example of an ”oligosynthetic” and historical analysis of Nahuatl, will be the etymology of the word pachtli ”moss”.

The Etymology of pachtli ”moss/hay/patch/rags

The argument that I will advance is that pachtli, pachoa and pachihui both come from a single source namely an original verb with the general meaning ”for two surfaces to come in contact one over the other” with an intransitive version meaning ”to put two surfaces in contact one over the other”. These verbs would have been *pachiwi and *pachiwa, respectively. It should be uncontroversial to posit the existence of these two verbs, pachiwi is attested with the range of meanings given in the first paragraph, which can clearly be abstractly reduced to the concept of two surface being in contact one over the other. In Nahuatl we find many pairs of transitive and intransitive verbs where the intransitive ends in /-i/ and the transitive in /-a/, and among these pairs we find a subgroup where the intransitive ends in /iwi/ and the transitive in /-oa/. Linguist Una Canger argued convincingly in 1980 in her book “Five Studies of Nahuatl Verbs in -Oa” that the verbs in /-oa/ come historically from the sequence /-iwa/ where the presence of the /w/ fused with the preceding /i/ causing it to become rounded and the glide to disappear from the phonemic representation of the word (this soundchange has also caused the noun /si:watl/ “woman” to become /so:atl/ in some dialects, and similar changes are common both in Nahuatl in and in the languages of the world in general). This makes a lot of sense and explains whi the -iwi/-oa pair are the only ones where the transitive and intransitive versions differ in more than the final vowel. So far so good, pachoa comes from *pachiwa, but that doesnt mean that *pachiwa didn't come from a construction such as /*pach-iwa/ and was derived from a root “*pač” meaning “moss”.

Here it is good to make a little point about what we know about the history of Nahuatl syllable structure. Nahuatl today allows both open and closed syllables of the shapes /CV/ /VC/ and /CVC/. But historically this was not the case. Most historical linguists would agree that Proto-Nahuatl probably allowed only /CV/ and /V/ syllables (or at least that they strongly favored them), and that most closed syllables can be understood either as the result of sequences of /CVCV/ where the last vowel has been lost through interaction with the stress pattern, or as the result of metatheses where a /CV/ syllable has been inverted to /VC/ in certain phonological environments.

This in turn means that /pač/ was not a possible syllable in proto-Nahuatl, the root would have to be reconstructed as */pačV/ the last vowel probably being an /i/ since often Nahuan /č/ comes from the sequence /tsi/. Since the original form of the absolutive suffix -tli was /*ta/, the noun pachtli must originally have been /patsita/ or /pačita/. So that means in turn that in order to derive pachoa from pachtli would have to see the original word as composed of two morphemes /patsi-wa/ where /wa/ is a suffix deriving a transitive verb from the noun. This is a possible etymology.

However lets look a little wider for related words. If we start looking at words that are similar to pachoa and pachihui we soon find /patska/ “to squeeze out a liquid” and the pair /patsiwi/ “to become crushed/mashed/bruised” and /patsoa/ ”to crush/mash/bruise something”. To me these words seem to be semantically related to pachihui and pachoa, through the meaning of pressing two surfaces together, and their phonological form is also strikingly similar. But they don't seem very similar to the word “moss”. 

If we look aside from the verbal endings, in -iwi and -oa what makes these words look similar is only the syllable /pa-/ and the fact that there is a phonetic and possibly historical relation between /ts/ and /č/. So what could /pa/ mean?

Whorf actually had the syllable /pa/ in his notes on oligosynthesis and he gave it the following possible meaning: “pervade, go through, across, beyond, and especially over.” We see the syllable /pa/ connection with a meaning of superposition in words such as -pan “on”, pamitl “banner” “-ikpak” “over”, ikpalli “seat”, panoa “cross over” and in the verbs patzihui, patzca, pachoa, pachihui if we accept the conjecture that all of these words are related through the meaning “direct contact between two surfaces one above the other”. If we postulate such an old morpheme /pa/ with the suggested meaning then we could see all these words as related, and the vebs as being formed by the morpheme /pa/ and a verbal root or a verbalizing suffix (here either -tsiwi/-tsiwa or -chiwi/-chiwa).


The verbs patzihui and patzoa would come from:
  • *pa-tsi-wa “to forcefully move a surface on top of another” > /patsoa/
  • *pa-tsi-wi “for two surfaces to forcefully move on top of eachother” > /patsiwi/
Here the verbal ending -tsiwi/tsiwa would give a meaning of “moving something forcefully/loudly” and the proto-morpheme /*pa/ would be a modifier describing the shape into which something is moved. We have other verbs where it is possible to see a similar construction, for example /ilakatsiwi/ “to twist something”/ilakatsoa/ to wrap something/to roll up something“, tlatsoa (from /*tlatsiwa/) “to beat something” (unfortunately “tlatsiwi” has a completely different meaning “to be lazy”), tlatsini “give a loud explosive sound” also suggests the relation between “tsi” and some kind of forceful motion/sound.

Pachihui and pachoa would then come from:
  • /pačoa/ < *pa-či-wa “to move a surface on top of something else”
  • /pačiwi/ *pa-či-wi “for two surfaces to move on top of each other”
Here the verbal ending čiwi/čiwa could give the meaning of “to move/be moved on top of something else” (possibly related čiwi/čiwa could also be related to /či:wa/ “to do X”).

We do have some verbs ending in /-choa/-chiwi/ that suggests a derivation through compounding with a root that indicactes movement. For example /malakačoa/ “to spin something round” and /malakačiwi/ “to spin round” obviously related to the noun root malakatl “spindle”. Here again I would propose that the evidence suggests that malakatl and malakachoa comes from a root /malaka/ meaning “move in a spinning pattern” and that the noun simply takes the absolutive suffix to become malakatl “thing that spins” and the verbs are compounded as malaka+čiwi “to do the malaka-movement”, in this way pačiwi would be “to do the pa-movement”. Really I think that malakachoa is a very good sign that we are on the right path with our analysis, because apart from malakatl “spindle whorl” another word malakachtli “something that has been spun” (for example thread on a spindle whorl) also exists, which clearly seems to be derived from malakachoa which in turn seems derived from malakatl. Note also the clear relation between ilaka and malaka in ilakatsoa and malakachoa both of which have to do with a turning or twisting motion.

And then finally the noun pachtli would come from


  • *pa-či-ta “something that is on top of something else/ something that has done the pa-movement”
The adjective pachtik would come from

  • *pa-či-ti-ka “to be like something that is on top of something else/to be like something that has done the pa-movement”
If this argumentation holds it suggests that proto-Nahuan had a class of verbs composed of a morpheme that described a specific type of movement (turning, superposition, etc.) or a position composed with verbal stems with relatively general meanings. Interestingly this is exactly what we find in the Uto-Aztecan language Cora, which has a class of 16 different verbal prefixes that describe position and movement. These prefixes can even be combined to create complex types of movement (Casad 1982). So basically it seems that proto-Nahuan may have shared this form of verbal morphology with Cora and maybe also Huichol, Cora's sister language.

So finally, instead of deriving pachoa from pachtli or vice versa, we derive both from a hypothetical root “pa” with a highly abstract meaning, and by positing the use of derivtaional processes that no longer exist in the language.

How is that for oligosynthesis?

Let me know if you find the proposal convincing or not.


Cited works:

  • Canger, Una. 1980. Five studies inspired by Nahuatl verbs in -oa. Travaux du Cercle Linguistique de Copenhague, Vol. XIX. Copenhagen: The Linguistic Circle of Copenhagen
  • Casad, Eugene. 1982. Cora Locationals and Structured Imagery. Doctoral disssertation. University of California San Diego. 

Ingen kommentarer:

Send en kommentar