syllable-initial aspiration

[Frances Karttunen]

Dear John,

I don't have any solution to offer off the bat about how to represent the
problematical syllable-initial aspiration in Nahuatl of which you give an
example (except maybe to ask whether it's segmental or prosodic).

But I do have some other thoughts I'd like to share.

As long as a language is unwritten, or is written mainly with ideographic
characters (like Chinese), or--as in Mesoamerica--with a mixed system of
ideographs and rebus-writing (word x, in full or in part, sounds like word
y), people don't seem to get hung up on spelling.

Once an alphabetic writing system comes into play, it seems inevitable that
people confuse the language with the elements of the writing system.  They
talk about "the letters of the language" instead of about the letters being
used to represent the sounds of the language.  Purism arises, so that some
ways of representation are considered "correct" and others "wrong" or
"corrupt."  If you think this only happens to layfolk, have a look at the
censorious things Andrews has to say in the second edition of his Intro. to
Classical Nahuatl.  Yikes!

There's no virtue inherent in written characters.  For the sound system of
any given language there are many ways (actually, an infinity of ways) of
representing utterances in that language that are fairly internally
consistent. Representing a sound with the letter "c" or  the letter "k" is
not a matter of correctness, as you know so well.  But context (historical,
social, conventional) can make people take umbrage about one as over against
another.  Think, for instance, of the connotations for English speakers of
"Amerika" spelled with a "k."

Most languages have some consonants and/or vowels that don't correspond to
the usual uses of the letters of the Roman (or Greek or Cyrillic) alphabet
(especially as they exist in boxes of type).  Many/most languages have
series of distinctions for which these alphabets offer no resources.  So
what is to be done in creating an alphabetical system of writing for such a
language?

One device is to add diacritics to the existing letters: a bar here, an
accent there, a cedilla underneath. In Swedish there is a little "o" set on
top of "a" to represent the same sound that Danish writes as "aa," which in
turn is different from what "aa" represents in Finnish.  In Norway an "o"
with a slash written through it represents the equivalent to what in Swedish
is written with an umlaut over the "o." Any and all these solutions work
equally well as long as nobody gets his or her nose out of joint about
nationality.

Another is to create digraphs: two letters together to represent a single
sound.  There are plenty of digraphs used in the traditional writing of
Nahuatl. Even the writing systems for Nahuatl that use "k" and "w" usually
also use "tl" and "ch."

Another way to create more letters to represent sounds is to turn some
letters and even numbers around (in the old days, taking them out of the box
of type and inverting them or setting them sideways).  So you get the number
"8" on its side in some printed Algonquian.  In Yucatec May, there used to
be a backwards "c" that has since been replaced with the digraph "dz."  Old
writing systems for other Mayan languages were studded with numbers used as
letters to represent glottalized consonants.

The International Phonetic Alphabet and the American system of phonetic
notation (that differs in some details from IPA) were created as tools to
represent pronunciation independently of language. A [k] in IPA is [k] no
matter what language is being represented. There are no digraphs, but
instead "one character for one sound." This is where the support for "k" and
"w" in Nahuatl comes from.

Naturally this runs into all sorts of problems, because utterances in
particular languages exist as the product not of a concatenated string of
unitary sound units but within the realm of a set of contrasts
(alike/different) that varies from language to language.  Fine phonetic
transcription is loaded with little squiggles and flourishes, most of which
are unnecessary and irritating for speakers of the language in question.
But until a fieldworker approaches native command of a language, s/he
doesn't know what is important and what is redundant.  That's a field
linguist's job, and the friars who adapted the the letters of the Roman
alphabet to represent Mesoamerican languages were proto-field linguists of
very high caliber.

The thing that everyone tends to lose sight of is that the letters,
including special characters, are just symbols.  They are not inherent in
any particular language.

But there are things that ARE inherent in a particular language.  If
language X has a contrasting series of voiceless and voiced consonants, then
fine. The Roman alphabet has p/b t/d k/g etc. to represent those contrasts.
But if the language in question has a series of contrasting
plain/glottalized consonants, for instance, then some convention (a
diacritic, digraphs, special characters) is needed.

Suppose a language has a contrast of plain/aspirated consonants.  They could
be written as p/ph t/th k/kh.  Or p/bh t/dh k/gh.  Or even the same old p/b
t/d k/g.  As long as everyone is in on what is being used and what it means
in this language.  BUT, using p/b t/d k/g could lead someone to think that
the contrast is voiceless/voiced. To the ear of a person who is a native
speaker of a language that contrasts voiceless/voiced, unaspirated/aspirated
sounds just the same.  Even spectrograms don't reliably show a difference,
because it is the systematic phonology of each language that is in question.

Another issue is, as you have said so eloquently, tradition. Once a language
has used one writing system for a while and has a body of written/printed
literature, a spelling reform (like the one so actively promoted for English
by George Bernard Shaw), no matter how "logical," separates the post-reform
reader from the pre-reform literature.  Most linguists move nearly
unconsciously from IPA to the American phonetic system to whatever
alphabetic conventions a particular language uses, but most nonlinguists get
frustrated.

(Linguists aren't immune to frustration. I never got the hang of Cyrillic,
so in Russia I carried around a small notebook and laboriously
transliterated what was on street signs in order to find out where I was.
It would take me ages to get to places.)

Likewise, in a situation of daily bilingualism, the difference between, say,
the conventions of written Spanish and the conventions of writing Nahuatl
with "k"s and "w"s can be counterproductive when the teaching of reading and
writing is going on in school in the dominant language, Spanish.

And, of course, there are matters of ideology. The new spelling conventions
adopted for the Mayan languages were based in part on decisions that have
less to do with the sound systems of the languages involved than with
matters of identity.  This was equally true of the spelling conventions for
Nahuatl proposed and adopted at the Aztec Congress held in Milpa Alta during
the presidency of Lázaro Cárdenas.

When one struggles with the type of questions to which you are addressing
yourself, there are multiple dimensions including at least the following:

The way the language actually works.

The way the language has been written in the past and the existing body of
literature so written/printed.

The subjective feelings speakers of the language have about one particular
writing system over another. For instance, do Nahuatl speakers wish to
maximize the difference in appearance from the way written Spanish looks on
the page? Or do they find a sense of unity with their past in the way
colonial-period written/printed Nahuatl looks on the page?

In this matter there are bound to be majority and minority opinions. I have
heard some Nahuatl speakers reject the Aztec Congress spelling system as
looking "bolshevik." (It's not coincidental that Diego Rivera and his
buddies hung out in Milpa Alta during the 1930s.) On the other hand, some
people might associate the Carochi spelling (or the Molina spelling, the
Olmos spelling, etc.) with the evangelization of the indigenous worlds in a
profoundly negative way.

So what it comes down to is that it must be up to the speakers of a language
how they will write it (or even IF they will write it).  It's up to the rest
of us to ask, listen, and respect even if we disagree and choose to use a
different system for clarity among ourselves.

Personally, I think the 16th and 17th century friars did a superb job of
field linguistics and created very durable writing systems in which a huge
corpus of writing in Nahuatl exists.  There were a couple of things they
under-represented, and I prefer to simply augment the traditional system, as
Carochi did, but differing in a couple of minor details, one of them having
to do with the "saltillo" (which, then as now, was apparently variously
pronounced in different areas but had the same FUNCTION whatever its
pronunciation).

By writing the saltillo as a diacritic over the preceding character, Carochi
gave the impression that the saltillo was a characteristic of the vowel: a
glottalization of the vowel or a rise in the pitch of the vowel.  Yet we
know that these changes to the vowel were contextual. They happened when the
vowel was followed by a particular consonant, albeit a consonant Spanish
speakers had difficulty hearing or describing.

One consequence of this was that in the 1930s Benjamin Lee Whorf, a
perfectly good linguist, went off on a tangent and described Nahuatl as a
tone language.  (Lyle Campbell and I have published an essay on this as an
introduction to our edition of Whorf's Milpa Alta field notes.)

In Nahuatl, vowels followed by saltillo behave systematically like vowels
followed by other consonants.  Since the distinction between open and closed
syllables is a crucial one in Nahuatl morphology, I have admired the choice
Andrews made to follow the sporadic usage of Molina, Sahagún, and others in
writing the saltillo with the letter "h."

This consideration seems to me to trump other arguments, such as wanting to
look like Carochi because Carochi is somehow the best and purest example of
written and printed Nahuatl. (After all, Carochi was unsuccessful in his own
time in getting Nahuatl speakers to adopt his package of orthographic
improvements, so it's hardly sacred.)

Likewise, I find it strange that some of my colleagues are so fond of the
cedilla, as though writing c-cedilla is better or clearer or more authentic
than "z."

I really hope that for older Nahuatl, we non-Nahuah follow the lead of
Andrews and doesn't further proliferate writing conventions. What
Nahuatl-speakers chose to do is entirely up to them, and we need to observe
and learn from them.

That said, in the course of time and over the geography of indigenous
Mexico, there has been a lot of variation and some historical change leading
to mergers of other consonants and consonant clusters with the realization
of what we call "saltillo."  This is a real challenge to modern writing.
Does one write how a word sounds (as one would in IPA) or as the word is
morphologically and has been historically? The latter strategy has been the
source of immense problems in English spelling, which often reflects a past
pronunciation/derivation of something that has since undergone change- being
pushed in one end and out the other of the Great Vowel Shift, for example.

Since as you describe it, you are engaging in a language revitalization
project to:

"re-elevate nahuatl to an academic stature at the
university. We will teach native-speakers to read and write in nahuatl.
They will read older texts, and comment on them in nahuatl verbally and
in written form. They will discuss important issues and write about
them in nahuatl, and participate in the production of dictionaries,
grammars, and other texts."

I'd like to recommend an article about decisions about representation of
specific languages:

"Issues of Standardization and Community in Aboriginal Language
Lexicography" by Keren Rice and Leslie Saxon.

This appears in Making Dictionaries: Preserving Indigenous Languages of the
Americas, edited by William Frawley, Kenneth C. Hill, and Pamela Munro.
Published in 2002 by the University of California Press.

I recommend the whole book to listeros with an interest in these matters.
Skip the theoretical section and cut directly to the "Indigenous
Communities" and the "Personal Accounts" sections. There are two (!)
articles in there about Nahuatl.


Fran