Etymology of "Mexico"

[David Wright]

Muy estimado Michael:

I'm glad we've gotten beyond the "messy phonology" argument to something
more tangible, i.e. the possibility that one of Andrew's grammatical rules
(/kk/ > /?k/) may be incorrect and that he offers little supporting evidence
for it (only the toponym Me:xihco). The frequent lack of evidence and of
showing where the examples are from is the weakest aspect of Andrews'
Introduction to Classical Nahuatl. I often turn to this book for
morphophonological information, since his treatment of this area goes beyond
other sources I have at hand. You are right in being suspicious of the
statements he makes without displaying his evidence. Usually when one looks
for supporting evidence for Andrews' claims one finds them, but there may
still be problems here and there (in spite of the fact that the 2003 edition
was revised during a quarter of a century of constant use, although this is
no guarantee). At any rate, the problem of /kk/ > /hk/ that you bring up
needs to be solved. If no further examples turn up after an exhaustive
search, this rule should be crossed out from our collective list of possible
morphophonological changes. It occurs to me that a good place to start would
be to read through Carochi with this doubt in mind. I wish I had a
searchable digital version; that would make this task much quicker. Any
other sources that consistently use the glottal stop (there aren't that
many) should also be searched. I checked Launey's thesis but it he seems to
have spread his comments on morphophonology throughout his lengthy text, so
there's no quick fix. I don't have the time to solve this problem right now.
If any listeros have pertinent data, it would be helpful if they shared it.
The matter is of some importance.

As for vowels before glottal stops being short, you don't need page 29 of
Andrews 2003 to see this; it's basic Nahuatl phonology. Just run through any
of the grammars and dictionaries that mark vowel length and glottal stops
(Carochi, Andrews, Launey, Campbell/Karttunen, Karttunen, Wolf, and
Bierhorst) and you'll see how it works.

So we have one possible analysis of the toponym "Mexico" that works as "in
the navel of the Moon", in which the optional regressive dissimilation
proposed by Andrews (kk > hk) is not applied:

Me:xxi:cco ((me:tztli - tli) + (xi:ctli - tli) (tz + x > xx) + co).

In the latter analysis the only morphophonological change required is the
regressive assimilation tzx > xx which you have accepted as "nuts-and-bolts
Nahuatl phonetics".

The second form, which depends on the optional kk > hk dissimilation, is
essentially the same as the latter, except for the first c (/k/) becoming h
(/?/), with the required shortening of the long vowel (i: > i).

Me:xxihco ((me:tztli - tli) + (xi:ctli - tli) (i: > i) (tz + x > xx) + co (c
+ c > hc).

Either form, Me:xxi:cco or Me:xxihco, can be translated "in the navel of the
Moon". Either would have usually been written "Mexico" in standard
"Franciscan" orthography, since long vowels were not marked, glottal stops
were rarely written, and double consonants were usually written as single.
(Exceptionally, I've seen xx in Franciscan orthography, e.g. "inimexxaiac"
(in i:mexxa:yac), "su cara [máscara] de [piel de] muslo", in book 2, chapter
30 of the Florentine Codex.)

In support of Me:xxihco, we have Carochi (book 3, chapter 11) writing
Me:xihco (I've changed his macron into a colon for the long e, and his
accent over the i to an h to sneak it by the Internet gremlins; these don't
alter the underlying phonology.) This is not a typo. In book 1, chapter 2,
he writes Mexihcatl (person from the city of Mexico) and Mexihcah (people
from the city of Mexico). Here he seems to have forgotten to mark the long
vowels; in book 3, chapter 11 we find Me:xicah and Me:xicah. The usual
procedure with these gentile names, derived from toponyms ending in the
locative suffix -co, is to remove the -co and add the gentilic suffix -ca
plus -tl for singular or -h for the plural. Carochi didn't write the double
x, but this can be considered normal in colonial period Nahuatl texts.

So there we have a non-Andrews example of Me:xihco. That would tend to
reinforce Andrews' kk > hk regressive dissimilation, assuming the presence
of the root xi:c, although additional examples are still needed.

Carochi's mentor, the Jesuit priest and native Nahuatl speaker Antonio del
Rincón, descendant of the royal house of Texcoco and author of a Nahuatl
grammar (Arte Mexicana) published in 1595, has something to say on this
matter, as I mentioned briefly in a recent post, providing the citation.
(Unfortunately, although he explains how he used diacritics to mark glottal
stops in his text, the printer was unable to reproduce them and they were
omitted from the published version; as far as I know the original manuscript
has not surfaced.) Here is what Rincón says, in the first chapter of book 4
(folio 50 recto and verso of the 1595 edition):

"Nota lo primero que en qualquier composicion el nombre que pierde algo con
la composicion es el que tiene la significacion en oblico, o como adjetivo
v.g. [...] Mexico. en medio de la luna, porque perdio el tli, el nombre,
metztli y generalmente pierden los nombres la ultima en composicion, como
con los genitivos de los pronombres."

In his "Vocabulario breve", at the end of his Arte (without folio numbers),
we find this gloss:

"Mexicco: ciudad de Mexico, i. en medio de la luna."

It's pretty clear that he's thinking (me:tztli - tli) + (xi:ctli - tli) (tzx
> xx > x) + co, with xi:ctli, "navel", meaning "middle" in this context. At
least I don't see any viable alternatives.

As John Sullivan pointed out, central Mexican toponymical etymology involves
going beyond morphological analysis and searching for the meanings behind
the names. I pointed out that most toponyms passed from one central Mexican
language to another as calques. Here's the data on the Otomi name for Mexico
Tenochtitlan that I mentioned in  earlier post. It appears in the Huichapan
Codex (BNAH Testimonios Pictográficos 35-60), with two words: anbondo and
amadetzänä, sometimes written together, sometimes individually, much as we
find Mexico and Tenochtitlan together or apart. The final 'o' in Anbondo
represents a vowel midway between Spanish /o/ and /a/ and today is written
by the Otomi with an underlined 'a' (except in the Mezquital, where this
phoneme has shifted to /o/). So we have the singular nominal prefix an- with
the word 'bonda (underlined 'a'), where ' is a glottal stop. It means "the
red Opuntia fruit", the seedy fruit of the nopal cactus, of the sort that
stains your mouth bright red when you eat it. (On a deeper level, this type
of fruit is a metaphor for human hearts, food for the Sun, in Mexica
literature and iconography, e.g. cuauhnochtli, "Opuntia fruit of the eagle",
i.e. human hearts.) Anbondo is the semantic equivalent of the Nahuatl word
tenochtli. The second Otomi toponym is Amadetzänä, which can be parsed as
the singular nominal prefix (an - n) plus the adjective made ("middle") plus
the word tsänä (today zänä in Mezquital Otomi and some other variants),
"Moon". (The ä is a nasal /a/.) Thus Anbondo Amadetzänä can be translated
"the red Opuntia fruit in the middle of the Moon".

The close match between Rincon's translation of Mexico and the Otomi toponym
Amadetsänä strongly supports the former's etymology.

I guess that's all I have. I hope all this helps you to see the logic of the
navel/middle of the Moon hypothesis. Thanks for motivating me to spend an
interesting four hours looking at this problem and writing up the results.
Hopefully someone will help us resolve the kk > hk question.

Saludos respetuosos,

David Wright

********************************************************************
The problem with Andrews' explanation, David, is that it's a fallacy based
on circular logic, with no supporting evidence.

He explains the etymology of "Mexico," on page 500, on the basis of this
putative /kk/ > /?k/ shift, basing this pronouncement on an earlier note in
the explanation of Nahuatl phonology. The reader then goes to that note, on
page 35, only to find that he says, well, /kk/ > /?k/...and we see this in
the term... Mexihco. (!)  Bad reasoning. We need real evidence, many more
examples. Andrews doesn't have any. :-)

Now, on page 29 that you refer us to, Andrews simply states that the vowel
before a glottal stop has to be short. Uh...

None of the above serves to explain the etymology of /me:xihco/, as far as I
can see.

Saludos y buenos tardes,

Michael


_______________________________________________
Nahuatl mailing list
Nahuatl at lists.famsi.org
http://www.famsi.org/mailman/listinfo/nahuatl