consonant deletion, obligatory synharmony, typical suffixing: an explanation of spelling practices...

8/13/2019 Consonant deletion, obligatory synharmony, typical suffixing: An explanation of spelling practices in Mayan writin

1/63

Tis is a contribution from Written Language & Literacy 13:1 2010. John Benjamins Publishing Company

Tis electronic file may not be altered in any way.Te author(s) of this article is/are permitted to use this PDF file to generate printed copies tobe used by way of offprints, for their personal use only.Permission is granted by the publishers to post this file on a closed server which is accessible

to members (students and staff) only of the authors/s institute, it is not permitted to postthis PDF on the open internet.For any other use of this material prior written permission should be obtained from thepublishers or through the Copyright Clearance Center (for USA: www.copyright.com).Please contact [email protected] or consult our website: www.benjamins.com

ables of Contents, abstracts and guidelines are available at www.benjamins.com

John Benjamins Publishing Company


2/63

Written Language & Literacy: (), . ./wll...mor

/ - John Benjamins Publishing Company

Consonant deletion, obligatory synharmony,typical suffixing

An explanation o spelling practices in Mayan writing*

David Mora-MarnUniversity o North Carolina, Chapel Hill, USA

Te Pre-Columbian Mayan hieroglyphic script utilized logograms, representing

CVC roots or CVCVC stems, and CV syllabograms. Starting with Knorozovs

(1952 etc.) initial breakthroughs in applying a Mayan linguistic model to account

or the scripts spelling practices, most scholars have assumed that synharmonic

spellings o roots or stems, those in which the final consonant is complemented

by means o a CV syllabogram whose vowel is identical in quality to that o

the root (e.g. C1

V1

C2

C2

(V1

)), exhibited a linguistically fictitious (or silent)

vowel; such synharmonic spellings were commonly assumed to be deault. E-

orts were then aimed at determining the motivation o spellings in which the fi-

nal syllabogram is instead disharmonic (e.g. C1V1C2 C2(V2)). Recently, it has

been proposed that the vowels o disharmonic spellings were utilized as diacrit-

ics applied to the vowels o the preceding syllables in order to convey that such

vowels were complex, while maintaining, generally, that synharmonic spellings

were deault. Te present paper offers a thorough review o these proposals and

gives arguments against their persuasiveness, abiding instead by our phonologi-

cal contexts that call or the insertion o fictitious synharmonic vowels, supple-

mented by morphological conditioning and consonant deletion that account or

additional cases o synharmonic spellings, and the vast majority o disharmonic

spellings. Tese principles allow or a major refinement o the definition o con-

ventionalized and deault spellings, a new avenue or determining the nature o

pseudologographic or morphosyllabic signs based on common syllabograms,

and a new cognitive ramework or addressing the question o the nature o

logograms and syllabograms, as well as the origin and development o Mayan

spelling practices.

Keywords: Mayan, epigraphy, hieroglyph, disharmony, synharmony,

morphosyllable, phonological conditioning, morphological conditioning, typical

suffixing, conventional spellings


3/63

Consonant deletion, obligatory synharmony, typical suffixing

. Introduction

Knorozovs (1952, 1955, 1958, 1965) Principle o Synharmony1

states that the pho-netic spelling o the second consonant o a C1V1C2root would involve a phonetic

sign o the shape C2V1, whose vowel echoed the vowel o the root, but was not

meant to be read, i.e. it was fictitious. Tis principle is illustrated by the spell-

ings in Figures 1a and b, KIN-nior Pre-Cholan *kiin sun, day and ka-kaor

Pre-Cholan *kahkfire, respectively.2However, Kelley (1962) emphasized that a

Principle o Disharmony, whereby the word-closing consonant o a C1V1C2root

are spelled with a C2V2sign whose vowel differed rom the vowel o the root, was

much more common than Knorozov had originally thought, and thus posed a

great challenge to decipherment.3Tis principle is illustrated by the spellings in

Figures 1c and d, TUN-nior Pre-Cholan *tuunstone, rom Proto-Mayan *too,

and cha-kior Late Proto-Cholan *chahk (~ *chahak)lightning, rom Proto-Ma-

yan *kahoq. Te question is, then, what determines whether a spelling is synhar-

monic or disharmonic?

a b c d

Figure 1. Synharmonic and disharmonic spellings o CVC roots.

a. KIN-ni. Excerpt rom Chichen Itza Four Lintels, Lintel 2. Afer Beyer (1936: 235).

b. ka-ka. Excerpt rom Chichen Itza Monjas Lintel 4. Afer Beyer (1936: 236).

c. TUN-ni. Excerpt rom onina Monument 95. Afer drawing by Ian Graham.

d. cha-ki. Excerpt rom Caracol Ballcourt Marker 3. Afer drawing by Nikolai Grube inChase et al. (1991: 5, Figure 3).

Tree types o approaches have been utilized to address this question, and are given

the ollowing labels: (1) an Orthographic Approach, based on Orthographic Arbi-

trariness; (2) a Phonological Approach, based on Phonological Conditioning; and

(3) a Morphological Approach, based on Morphological Conditioning, with Actu-

al Suffixing and ypical Suffixing subtypes. Tese approaches, their subtypes, and

their major proponents are listed in able 1. Te present paper reviews and tests

some o these proposals in detail: Bricker (1989), Justeson (1989), Hopkins (1997),Houston et al. (1998, 2004), and Lacadena and Wichmann ( 2004). Part o the im-

petus or this study lies in the wide acceptance o two recent proposals by Houston

et al. (1998, 2004) and Lacadena and Wichmann (2004), who have argued that


4/63

David Mora-Marn

disharmonic spellings served as diacritics to indicate that the vowel o a preceding

syllable was complex (Vh, V7, VV) rather than simplex (V). Tis wide acceptance

is unortunate or the ollowing reasons: (1) many specialists and non-specialistsalike have begun to utilize the predictions o such hypothesis as tools or historical

linguistic reconstruction even when linguistic evidence or such reconstructions is

otherwise lacking or contradictory; (2) these authors have not systematically and

exhaustively tested previous approaches, but have simply dismissed them out o

hand; (3) the two teams o scholars have ollowed a purely etymological approach,

Table 1. Approaches to synharmony and disharmony.

Scholars Synharmony Disharmony Knorozov 1952, 1958,

1965

Deault Motivated morphologically

Kelley 1962, 1976 Deault Remnant o ancient phono-

logical root structure

Bricker 1989 Deault Consonant Deletion or Vowel

Insertion

Hofling 1989 Deault Actual Suffixing

Justeson 1989 Generally deault, some are due toPhonological Conditioning

Consonant Deletion, ypi-cal Suffixing, some are due to

Phonological Conditioning

Hopkins 1997 Actual Suffixing should be the null hypothesis to be tested; Also,

Phonological Conditioning, due to echo vowels, also a possible

source or some o the fictitious vowels

Houston et al. 1998, 2004 Deault (1998), Deault and Pho-

nological Conditioning (2004)


Lacadena and Wichmann

2004

Deault initially, later due to Pho-

nological Conditioning


Possibly Orthographic Arbitrariness initially based on neutral

use o Caword-closing syllabograms

Justeson 2000 ypical Suffixing, except cases

o CC roots, which requently

exhibit synharmony with Ca signs

ypical Suffixing

Mora-Marn 2001, 2002,

2004, 2005a

Deault, Consonant Deletion ypical Suffixing, Consonant

Deletion, Actual Suffixing

Kauman 2003a, 2003b Morphological Conditioning and ypical SuffixingAnderson 2004 Phonological Conditioning based on homorganicity

Boot 2004 Neutral use o word-closing syllabograms: Ce, first, and Ca, later;

Value o Cewas based on ypical Suffixing


5/63


studying roots in isolation rom their morphological properties, whether inflec-

tional or derivational, and their syntactic contexts; and (4) one o the proposals,

the one by Houston et al. (1998, 2004), is analytically interrelated with a secondproposal by Houston et al. (2001), dubbed here the Morphosyllable Hypothesis,

which is shown to exhibit important analytical and empirical flaws.

Furthermore, Houston et al. (2001) have put orth a model o the analysis

and interpretation o the spellings o morphological suffixes in Mayan texts that is

directly linked to the implications o their approach to disharmonic spellings, and

thereore requires careul examination as well.

With this background in mind, the goals o this paper are the ollowing:

a. o test the proposal by Houston et al. (1998, 2004), reerred to below as Dis-harmony Hypothesis I (DH I), and its more refined version by Lacadena and

Wichmann (2004), reerred to as Disharmony Hypothesis II (DH II), by as-

sessing whether alternative approaches that have not been properly tested can

account or the same data.

b. o determine whether morphosyllables in the Morphosyllable Hypothesis

(MH) by Houston et al. (2001) are necessary constructs or analyzing spelling

conventions.

c. o assess whether synharmonic spellings are really unmarked or deault, or

whether they are themselves guided by more complex principles.

d. o arrive at a definition o conventionalization in terms o what it meant in

the practical everyday use o the script.

e. o consider the implications o the analysis or the history o the script.

Te first section o the papers provides the basic ramework or the epigraphic and

linguistic study o Classic Mayan writing, as well as the rationale or the methodol-

ogy employed.

. Preliminaries

. Linguistic assumptions

Classic Lowland Mayan (CLM) writing represented a standard Cholan language,

one o the subgroups o the Mayan language amily (Figure 2), as evinced by ex-

clusive Cholan innovations present in the conventional spellings o the most con-

servative components o the script, such as the time period and month names(Campbell 1984; Campbell & Kauman 1985; Josserand et al. 1985; Justeson et

al. 1985; Justeson & Fox 1989; Justeson & Campbell 1997; Houston et al. 2000;

Lacadena & Wichmann 2002). Te script was adopted by Yucatecan speakers,


6/63

David Mora-Marn

probably during the Early Classic period i not earlier, who began to incorporate

Yucatecan innovations gradually at first, but at an accelerated pace by the middle

o the Late Classic period (Bricker 1986; Justeson & Fox 1989; Lacadena & Wich-

mann 2002).4Nevertheless, a significant number o standard traits present in the

script represent linguistic materials no longer preserved in any o the Cholan or

Yucatecan languages, but still present in other branches o the Mayan languageamily, especially the zeltalan branch, the languages most closely related to the

Cholan languages (Mora-Marn 2003, 2009a; Mora-Marn et al. 2009).

Te basic root shape in Mayan languages, including the languages o relevance

to this paper, is CVC, or Consonant-Vowel-Consonant, where the vowel, V, may

be simplex (V) or complex (VV, V7, Vh).5Syllabification, or example o a CVCVC

word ollows the pattern CV.CVC, as with 7ah.kotdance, and that o a CVCVC-

VC word consisting o a root and a suffix ollows the pattern CV.CV.CVC, as with

7ah.ko.t-aj s/he danced, which is realized as CVC.CVC due to vowel syncope

(deletion) o o, 7ahk.t-aj. Te most common suffixes exhibit the shapes -V (-i

completive status o positionals), -VC (-elinalienable possession) and -CVC (-laj

intransitivizer o positionals).

Proto-Cholan has been reconstructed with a six-vowel system (/*i, *e, *a, *,

*o, *u/) by Kauman and Norman (1984), with only Vh remaining as a complex

vowel, as a result o a merger o *V7 and *VV into VV, possibly during Proto-

Cholan-zeltalan times, but very likely by Pre-Proto-Cholan times, and the merg-

er o *VV and *V into V, with a concomitant shif o *a to , by Proto-Cholan

times. Houston et al. (1998, 2004) argue instead that Proto-Cholan, and in act, theancestor o Chorti and Cholti, the two Eastern Cholan languages in Kauman and

Normans (1984) classification, still retained contrastive vowel length, using evi-

dence rom disharmonic spelling patterns in support o this hypothesis.6Similarly,

Late Proto-Mayan Huastecan

Central Mayan

Western Mayan

Cholan-Tzeltalan

(Greater Tzeltalan)

Greater

Qanjobalan

Greater

Mamean

Greater

Kichean

Eastern Mayan

Yucatecan

Proto-Mayan

Figure 2. Major subgroups o the Mayan language amily (afer Kauman 1976, 1990).


7/63


Lacadena and Wichmann (2004) have argued, also based on their own analysis o

disharmonic spelling patterns, that Cholan still retained *V7 and *VV during the

Classic period, but soon began to loose such distinctions, leading to a period odifferentiation in disharmonic spelling patterns. Tus, a resolution to these ques-

tions may be crucial to the analysis o the linguistic history o Cholan. Part o this

controversy is somewhat o a matter o relative chronology. Justeson et al. (1985)

have observed that Pre-Cholan likely retained phonemic *VV:*V contrasts, based

on the evidence o diffused Cholan vocabulary in Yucatecan.7 I the script was

innovated and standardized, say, by Pre-Cholan speakers, then it is possible that

some convention or conveying vowel length contrasts could have been incorpo-

rated at some point. Te present paper addresses whether this was the case, not

whether it was Pre-Cholan, Proto-Cholan, or some post-differentiation branch o

Cholan that standardized the script.

. Epigraphic assumptions

Te Maya script is pictorial and logosyllabic: it uses CVC and CVCVC logograms

(e.g. 544 KINor *kinsun, day, BALAMor *bahlm), as well as CV syllabo-

grams (e.g. 25 ka, 130 wa, as in 25.25: 130 ka-ka-w(a)). It also uses semanto-

grams (determinatives, classifiers), discussed in Hopkins and Josserand (1999) andMora-Marn (2008), but these are o no consequence in the present paper. Logo-

syllabic spellings, those combining logograms and syllabograms, were very com-

mon or roots and their suffixes (e.g. HUL-li or hul-i-arrive.here--3s

s/he/it arrived,CHUM-wa-nior chum-wan-isit-- s/he sat down), but

they were also used in cases where a suffix was not obviously present or intended,

leaving a word-closing syllabogram with a fictitious vowel a vowel that was not

to be read. Tis is the vowel that is at the crux o the problem, or this vowel is the

one proposed by some scholars to be a diacritic o the complexity o the vowel o

the preceding syllable.

. Methods

In contrast to Houston et al.s (1998) approach, in which previous proposals, such

as consonant deletion (Bricker 1989) and typical suffixing (Justeson 1989), are

simply claimed to have been proven inadequate to account or the nature o dis-

harmonic spellings, the present paper will test such ideas directly, in a much more

systematic and thorough ashion than was attempted by their original authors. Forthe act is that these early proposals have not received the detailed treatment they

deserve. o test them, two kinds o evidence are needed: (1) multiple spellings o

the same word in similar grammatical environments, and (2) multiple spellings o


8/63

David Mora-Marn

the same word in different grammatical environments. Also, there are two kinds o

evidence used by Houston et al. (1998) and Lacadena and Wichmann (2004) that

are not useul in testing the consonant deletion and typical suffixing approaches:(1) unique spellings, and (2) spellings o words with unknown etymologies or

phonological shapes. In other words, such approaches can only be tested by means

o contextual analysis, not isolated analysis o the spelling o the roots alone, with-

out regard or their suffixes and grammatical contexts.

. Testing the disharmony hypotheses

. Disharmony Hypothesis I

Houston et al.s (1998, 2004) hypothesis, dubbed here Disharmony Hypothesis I

(DH I), proposes that disharmonic word-closing syllabograms serve as diacrit-

ics indicative o complex syllable nuclei (i.e. CVVC, CV7C, CVhC, as opposed

to plain CVC). Tus, in their approach, an isolated spelling such as ba-kiwould

be read as ba-k(i) or baak bone, rom Proto-Mayan *baaq, and another such

as cha-kiwould be read as cha-k(i) or chahk lightning, thunder, rom Proto-

Cholan *chahuk, rom Proto-Mayan *kahoq. Te DH I also claims that: (1) syn-

harmonic spellings are unmarked/deault, in which case wi-ts(i)would give us

Proto-Cholan *witsmountain rom Proto-Mayan *wits, and tsi-b(i)would give

us Proto-Cholan *tsihb writing rom Proto-Mayan *tsihb; (2) there are no ex-

ceptions (cases where a disharmonic spelling yields a simple syllable nucleus); and

(3) whenever a disharmonic syllabogram is used to represent vowel complexity,

but the root it is ound with requires a suffix o some sort, an additional sign or

sequence o signs, called morphosyllables (Houston et al. 2001), is necessary to

spell that suffix. For example, a spelling such as BAH-hi-jawould be interpreted

as BAAH(-hi)-AJ (Figure 3a), or baah-ajhead- head/ace (o someone)rather than BAH-hi-j(a) or ba(:)h-ij, under the assumption that the hi syllabo-

gram is used in this context to indicate, with its disharmonic ivowel (which di-

ers rom the avowel o the preceding syllable), that the preceding syllable has a

complex vowel, aa. At the same time, i the io the syllabogram hiis being used

solely or its diacritic unction, then it cannot serve in the spelling o the required

suffix. Instead, such suffix is spelled exclusively by the syllabogramja, which is no

longer merely a syllabogram, but a morphological logogram, or morphosyllable

hence, this is the basic ormulation o the Morphosyllable Hypothesis (MH) aspresented by Houston et al. (2001). In act, this idea has many precedents, starting

with Knorozov (1952, 1955, 1958, 1962), who analyzed 181 jaalso as -ajwhen

used to spell a suffix on a verb, and Tompson (1950), who analyzed 24 lialso


9/63


10/63

David Mora-Marn

languages the suffix is usually composed o two or three allomorphs, -aaj, -iij, -eej.

Tus, the DH I and MH approaches are negating the possibility that BAH-hi-ja

could be spelling bah-ijon purely theoretical grounds, not on empirical grounds.Te data set used by Houston et al. (1998; 2004: 93, Figure 5.1), reproduced

here as able 2, is problematic: (1) almost none o the spellings in the Houston et

al.s Appendix contains source inormation that would be necessary to examine the

orthographic and grammatical contexts o the spellings; (2) two o the examples

represent the same morpheme, the suffix -omor -oom, once as -Co-maand also

as 7u-to-maor uht-o(o)m, which contains the same suffix, bringing their useul

sample size rom 75 to 74; and (3) the authors do not mention which spellings

were used or the Cu-Cicell (i.e. uu i), leaving 72 useul examples. Tese are

the examples reproduced in this papers Addendum. Te case o the -o:msuffix

is important, or it illustrates the interdependence between the DH and MH: i

the spelling 7u-to-macalls or the use o the oo toas a diacritic or a preceding

complex syllable nucleus, uho 7uht-oom, then it is not used to spell the vowel o

the suffix -oom, which it could very easily represent under normal, syllabographic

spelling conventions; instead, such ovowel is provided by the syllable ma, which is

now assumed to be a morphosyllable -OOM.

Table 2. Houston et al.s (1998) data set.i a u

i ii (3)

iii (1)

ihi (2)

i7i ()

ia ()

iia (2)

iha ()

i7a ()

iu ()

iiu ()

ihu (1)

i7u ()

e ei ()

eei ()

ehi ()

e7i ()

ea ()

eea (2)

eha ()

e7a ()

eu ()

eeu (1)

ehu (1)

e7u ()

a ai (1)

aai (8)

ahi (4)

a7i (1)

aa (18)

aaa (2)

aha (2)

a7a ()

au ()

aau (1)

ahu (1)

a7u ()

o oi ()

ooi (2)

ohi ()

o7i ()

oa ()

ooa (5)

oha ()

o7a ()

ou ()

oou ()

ohu ()

o7u ()

u ui ()uui (2)

uhi ()

u7i ()

ua ()uua (3)

uha ()

u7a (1)

uu (5)uuu (4)

uhu (2)

u7u ()


11/63


.. Exceptions

Contra Houston et al.s (1998, 2004) claim o no exceptions, the ollowing counter-

examples do exist (able 3; Figure 4):

Table 3. Exceptions to DH I (Houston et al. 1998, 2004).

Spelling8 Reconstruction Source

1. 7o-ki-bi(7ok-ib step) PC *7och, PYu *7ok PM *7ok enter

2. 7i-chi-la(7ich-il within) PYu *7ich-il

3. 7u-tsi(7uts good) PC *7utsgood PM *7uts

4. cha-ji (chaj dripping) Yucatec chahto drip V,

chahdrop N (Some cases oPM *tyare preserved in PYu

and PC as *ch)

PM *tyajdripping, splattering

5. ha-7i(ha7+i this one)9 PC *hain+e PM *ha7 (+i/+e/+a/+o) de-

monstrative pronoun base plus

deictic enclitic

6. la-ji(-laj-i intransitivizer

o positional roots)

PreC *-laj intransitiv-

izer o positionals, PYu *-laj

intransitivizer o completive

positionals7. ta-ki(-tak pluralizer o

some nouns)

PC *-tak (Chol -tyak, Chorti

-tak), PYu *-tak

PM *tiq ~ *taq plural

8. ta-li(-tl incompletive

intransitivizer o position-

als)10

PC *-tl incompletive intran-

sitivizer o positionals

9. wa-ni (-wn/-wan in-

transitivizer o positionals)

PC *-wn/-wan

10. WINAK/WINIK(winak/

winik man, person)

PC *winik man, person PM *winaq person

11. yi-cha-ki(y-ihchak(-il)

his/her/its fingernail/claw)

PC *7ihchak PM *7iSkaq

12. yu-ki-baand 7u-ki-ba

(y-uk-ib his/her/its drinking

cup, 7uk-ib-l/al (some-

ones) drinking cup)

PC *7uch, PYu *7uk; PC -Vb,

PYu *-Vb

PM *7uk drink, PM *-a-b

(V roots), *-b (Derived V),

*-i-b (IV roots)

13. ka-yo-ma(kay-o(o)m

fisherman)

PC *chy, PYu *kay PM *kar fish, *-oom agentive

nominalizer o nouns14. tsi-ma-hi(tsimah gourd) PCH *tsimah PM *tsima7


12/63

David Mora-Marn

a b c

d e

Figure 4. Some examples o counterevidence to the DH I.

a. yi-cha-kion Piedras Negras Lintel 2. Afer Stuart (1987: Figure 38e).

b. 7u-tsion Vase K1453. Afer photograph o Vase K1453 in Kerr (1999).

c. yu-ki-baon Early Classic bowl. Afer photograph in Coe (1973: 110).

d. 7u-ki-baor uk-(i)b-lcup (o someone), on Late Classic pottery vessel. Afer photo-

graph in Coe (1973).

e. cha-jion Quirigua Stela E. Afer drawing in Stuart (1987: Figure 54d).

. 7i-chi-laon Lintel 4A at Chichen Itza, Las Monjas. Drawing by the author based on draw-ing by Graham (1977a) in Bricker (1986: 98).

Houston et al. (1998, 2004) do acknowledge the case oyi-cha-kias a possible ex-

ception, but claim that its reconstruction is uncertain; in other words, they suggest

that it could just as easily have beeny-ihchaak instead oy-ihchak. However, the

evidence or this reconstruction is relatively straightorward, even i one considers

vowel-lengthening morphophonemic processes that result rom possession o cer-

tain nouns: in some Mayan languages the vowel o a noun root is lengthened when

the root is possessed, e.g. Mam n-xaaq+amy rock, based on xaqrock. A survey opossessed orms o this particular root in languages that exhibit vowel lengthening

with possession, seen in able 4, suggests this root does not experience such pro-

cess (Kauman 2003a). I have also added the likely Proto-Yucatecan orm *7ihchak,

given the Yucatec and Mopan data rom Bricker et al. (1998) and Cohuoj et al.

(2001: 33),, as well as the Chontal orm ichc/7ichk/, discussed below.

In the Chontal term 7ichc/7ichk/ (Keller & Luciano 1997: 116), the al-

most certainly descends rom Pre-Proto-Cholan *a(Kauman & Norman 1984),

or this language, unlike the other extant Cholan languages (Chol, Chorti), retainsvowel length contrasts in morphophonemic contexts, particularly in the deriva-

tion o nouns rom verbs: mnto buy (something) contrasts with manpurchase

(Knowles 1984).


13/63


Another case is the spelling ka-yo-maor kay-o(o)mfisherman. I we abide by

the MH we would get ka-y(o)-OM, in which case a orm kaay-o(o)mwould be rep-

resented. However, the term or fish comes rom Proto-Mayan *kar, with reflexes

*chyin Proto-Cholan (rom Pre-Cholan *chay) and *kayin Proto-Yucatecan. Te

suffix is reconstructible as *-oomin Proto-Mayan or Central Mayan, as indicated

by other cases o nouns derived rom other nouns provided by Kauman (2003a: 69,

361): *beh.oom rich; wealthy seemingly rom *behroad, and *baaq-eel jol-oom

skull (bone o the head) romjolhead.11I we abide by simple spelling rules, weobtain ka-yo-m(a), where only one fictitious vowel the ao ma is required,

instead o two the ooyoand the ao ma, as in the ka-y(o)-OManalysis under

DH and MH. Earlier it was noted that the inclusion o both -Co-ma and 7u-to-ma

in the data set poses a majorproblem: i intended to suggest that the ao main

-Co-maserves as a disharmonic diacritic o vowel complexity or the preceding

syllable nucleus, i.e. -Co-m(a) or -o:m, and at the same time, that the oo toin

7u-to-ma serves as a disharmonic diacritic o vowel complexity or the preceding

syllable nucleus, i.e. 7u-t(o)-maor 7uht-om, then 7u-to-macannot represent bothcomplex vowelsat the same time. Tis is a major inconsistency that could compli-

cate a large number o readings, essentially any spelling o a word exhibiting these

general traits: two complex vowels, one in the root, and one in the suffix.

Tese approaches problematize spellings that are completely unproblematic

under simple spelling rules (syllabographic conventions), leading, in some cases,

to constructions that go against the grain o linguistic evidence, but generally and

unnecessarily multiplying the number o entities that require explanation. For in-

stance, in Figures 5a and b, two spellings o CHUM[mu]-li or chum-ul- it is

seated appear. Following the DH I and MH approaches we would analyze this

spelling as CHUM([mu])-IL, which would suggest an -ilsuffix, rather than an -ul

suffix, or the stative marker o positionals, which is without a doubt a -V1lsuffix

Table 4. Comparative data or claw, fingernail (primarily rom Kauman 2003a: 364365).

Language Attested form of his/her/its claw, fingernail

Popti (Jakaltek) y-iskaqQanjobal y-isqaq chej

Chontal 7ichk

Mam t-xkyaq

Poqom -ixkaq

Kaqchikel r-ixkaq

Yucatec 7u y-ichak

Mopan 7u y-ichak


14/63

David Mora-Marn

whose vowel echoes the vowel o the positional root. O course, under the MH

approach, Houston et al. (2001) call or the existence o two kinds o morpho-

syllables: (1) regular morphosyllables, involving cases where a syllabogram was

used as a morphosyllable with a -VCshape whose vowel varied according to vowel

harmony rules, as with 130 wa, used in the spelling o the -V1wsuffix o active

transitive verbs; and (2) irregular morphosyllables, those with a -VCshape whose

vowels were invariable, as in the case o 24 lias -ILor the possessive suffix -il.

Given such conceptualization, one could simply say that 24 li, which they define

as an irregular, invariable, morphosyllable -IL, is actually regular in this instance,and thus it can stand or any -V1lwith any vocalic value. But this is not necessary:

i we simply abide by simple spelling rules, the uo muwould spell the vowel o the

-V1lsuffix, and the lo lispells its consonant: CHUM[mu]-lior chum-ul.

. Disharmony Hypothesis II

urning now to the DH by Lacadena and Wichmann (2004), henceorth DH II, it

consists o the ollowing ormulation o rules: (1) Harmony Rule 1, under which

synharmonic spellings are not deault, but serve instead as diacritics o simplex

syllabic nuclei (CV1C-CV1/CV1-CV1 CV1C); (2) Harmony Rule 2, under which

two kinds o disharmonic spelling contexts are diacritics o long vowels in the

preceding syllable (CVC-Ci/CV-Ci CVVC, where V = {e, a, u, o}, and CVC-Ca/

CV-Ca CVVC, where V = {i}); and (3) Harmony Rule 3, under which two kinds

o disharmonc spelling contexts are diactritics o glottalized vowels in the preced-

ing syllable (CV(C)-Cu CV(V)C, where V = {i, a }, and CV(C)-Ca CV(V)

C, where V = {e, u, o}). Lacadena and Wichmann qualiy their first rule by say-

ing that several synharmonic examples are actually used to spell words with a Vhnucleus, and so they do not argue or the use o disharmonic spellings to represent

Vh nuclei. In act, they also note that some synharmonic spellings are known to

occur or roots with VV or V7 vowel nuclei: CHAN-naor chan< Proto-Mayan

a b

Figure 5.

a. CHUM[mu]-lior chum-ulsit-SA seated rom Copan Hieroglyphic Stair Step 8. Afer

drawing by Barbara Fash in Schele (1989: 81).

b. ti-CHUM-mu-li or ti chum-ulseated. Afer drawing in Stuart (2005: 73).


15/63


*kaansnake, KIN-nior kin< Proto-Mayan *qiisun, day, and 7AJAW-waor

7ajaw< Proto-Mayan *7aajaawruler. Te authors argue that spellings like these

are conservative, harking all the way back to an Age o Synharmony, when thespelling rules or representing VV and V7 had not yet developed (Lacadena &

Wichmann 2004: 132). Te authors also note that [w]hen looking through the

earliest texts we notice that Cacomplements dominate the picture, and suggest

that prior to the Age o Synharmony the scribes used the a-vowel as a deault in

complements, concluding that [t]he competing principles o synharmony and

Table 5. Counterexamples to DH II (Lacadena & Wichmann 2004).

Rule Counterexamples

Harmony Rule 1:

CV1C-CV1/CV1-CV1

CV1C

bu-kuor *buhkclothes

tsi-bi or *tsihbwriting

ka-kaor *kahkfire

ta-naor *tahnchest

cha-pa-taor chapahtcentipede

KABA77aor *kaba7< Lowland Mayan *kaabaa7name

7AJAW-waor *7ajaw< Proto-Mayan *7aajaawlord, ruler

CHAN-na or *chansky (PYu. *kan)


16/63

David Mora-Marn

Ca-complementation were each assigned their own unction and coordinated into

a larger system (Lacadena & Wichmann 2004: 132). As is shown below, this notion

o an Age o Synharmony is not supported by the data; this act, and several ad-ditional counterexamples to DH II, shown in able 5, argue against its validity.12

It is also worth observing that while Lacadena and Wichmann (2004) call or

the existence o an early period o time during which scribes applied synharmony

quite regularly and generally, Houston et al. (1998, 2004) argue or a similar period

o time, only a late one, during the Late Classic period at the sites o Caracol and

Copan, during which scribes applied synharmony in a very general ashion. Te

ormer scholars suggest that the disharmony rules had not yet devised during that

early period, while the latter scholars propose that the late period o synharmony

reveals language in progress, particularly in the mergers o VV and V7 to V. While

these proposals are not necessarily contradictory, the act is that neither is neces-

sary, as is argued below.

. Section summary

Te above examples illustrate the urgency o evaluating the DH I, DH II, and MH

approaches, which have been widely accepted as valid by many specialists. Martin

and Grube (2000: 22), or instance, analyze the names o several kings two atikal, one at Calakmul, one at onina that contained the CLAW logogram as

Ichaak, an analysis that exhibits two errors: it omits the preconsonantal ho the

first syllable o Proto-Cholan 7ihchak, which has never been in question, and it

shows a long vowel, aa, in the second syllable, which the comparative and histori-

cal evidence demonstrates not to be accurate. Many epigraphers who have adopted

DH II have routinely incurred in the practice o reconstructing the phonological

shapes o glyphic roots and affixes based on this hypothesis, even when outside

evidence or their meaning or etymology is unavailable (e.g. 7a-nu) or contradic-

tory: the spelling cha-ho-mais now ofen transliterated as chaj-om, or chah-o7m

dripper, ollowing the present orthographic conventions, even though no outside

evidence or o7, only or oo, which would yield chah-oomdripper, is known. Also,

some scholars arbitrarily reconstruct VV or any disharmonic term: thus, 7a-na-

biis reconstructed as 7a[j]-naabhe o lakes, even though the root in question is

nahblake, and Vh did not merge with VV in Proto-Cholan-zeltalan or Proto-

Cholan. Tus, transcriptions and reconstructions, in an attempt to show more lin-

guistic details called or by the DH I or DH II, have become highly unreliable and

inconsistent.13Some o the remaining violations to the DH (I/II) likely involveActual Suffixing, such as the case o the independent demonstrative pronouns (e.g.

ha-7i, likely or *ha7-+i that one and ha-7o-ba, likely or *ha7-o7b+a these

ones), and are urther discussed below.


17/63


. Testing alternative approaches

. Consonant deletionNeither o the DH approaches takes into account the convention o Consonant

Deletion (C-deletion), which leads to underrepresentation, most commonly o

root- and word-final consonants, especially /l/ (Mathews & Justeson 1984; Brick-

er 1989, 2000; Justeson 1989; Zender 1999). Examples o such practices are very

common in two contexts: (1) the second consonant o the first root in a compound

word, such as ta-ja-MO7(-7o), attested in more complete spellings as ta-ja-l(a)-

a b

c d

e g

Figure 6. Principle o Consonant Deletion.

a. ta-ja-MO77oor taj-a(l)=mo7orch Macaw. Yaxchilan Hieroglyphic Stairway 5, Glyph

84. Afer drawing by Graham (1982: 179).

b. ta-ja-la-MO-oor taj-al=mo7orch Macaw. Machaquila, Structure 4, Glyphs V2-V3.

Afer drawing by Graham (1967: Figure 39).

c. ti-7AJAW-le or ti 7ajaw-lelin lordship. Piedras Negras Stela 3, Glyph F5a. Afer drawing

by Sylvanus G. Morley in Bricker (1989: 48).

d. ti-7AJAW-le-le or ti 7ajaw-lelin lordship. Piedras Negras Trone 1, Glyph H3. Aferdrawing by Sylvanus G. Morley in Bricker (1989: 48).

e. ka. Excerpt rom vessel K532. Afer photograph in Kerr (1999).

. ka-ka. Excerpt rom pottery vessel. Afer photograph in Robicsek and Hales (1981: 200).

g. ka-ka-wa. Excerpt rom pottery vessel K1837. Afer photograph in Kerr (1999).


18/63

David Mora-Marn

MO7, or taj-a(l)=mo7torchy macaw (a species o macaw?) (Figures 6a and b),

and also ku-MO7(-7o)~ ku-ku-MO7(-7o) or kuk=mo7 quetzal macaw; and

(2) the final consonant o a root or word, such as ka-se, also ka-se-wa, or caseufifh month, 7AJAW-le, also 7AJAW-le-leand 7AJAW-2leor 7ajaw-(a)l-elking-

ship, and cha-jo, also cha-jo-m(a), or chaj-o(7)mdripper. Given such practice,

one could ask whether yi-cha-ki is an underspelling o y-ihchak-il (c. Chol iy-

ejcach-il). O interest here, Bricker (personal communication 2002) has suggested

that the C-deletion and synharmony (see below or more discussion) evinced in

cases like the spelling o the suffix -lelcould be more phonetically generalized, and

perhaps used in other contexts, i.e. examples where C1VC1sequences. Evidence

or this phenomenon is presented below.

In any case, this tells us that it is not difficult to test or C-deletion. It is simply

necessary to find spellings o the same word in identical morphosyntactic envi-

ronments. Te ollowing spellings, among many others, are ound in ree varia-

tion in identical morphosyntactic environments that require a -Vlsuffix (able 6;

Figure 7).

A ew o these examples are illustrated in Figures 7 and 8, where the title 7a(j)-

bi-ki-laappears in ree variation with 7a-bi-ki, which is missing the lasign rom

the first example (see arrow), in the spelling o the the same scribes names. In

Figure 7, two different examples attest to C-deletion: first, ree variation in thespellings ma-su-la and ma-su in the context o the same title, and second, ree

variation in the spellings TUN-niand TUN-ni-li, in both cases or stone,

which appears in identical morphosyntactic contexts, as the head o a possessed

noun phrase that unctions as the subject o an intransitivized verb.

Using the database o Primary Standard Sequence texts prepared by Mora-

Marn (2004), which currently includes 805 entries, it is possible to show that out

o 348 entries with the drinking cup expression, yu-ki-bi or y-uk-ib (Mora-

Marn 2000), only 16 contain a final lVsyllabogram, 14 o them with laand 2 with

li; these point to the generally implicit presence o a possessive -Vlsuffix that is

partly spelled by the io 585 bi. In the same database, out o 47 entries o the

possessed nominalization 7u-tsihb-naj-al his/her painted thing, which exbibits

the suffix -najpassivizer o derived transitives ollowed by the suffix -al nomi-

nalizer, 23 examples show an explicit syllabogram la, while 24 show no such syl-

labogram, even though the possessed nominalization context requires a possessive

-Vlsuffix. Similarly, o the 99 examples o the possessed verbal noun 7u-tsihb-al

his/her writing as 7u-tsi-ba-li, 80 occur with an explicit li syllabogram, while

19 occur without it, also suggesting that these 19 were intended to be read as 7u-tsihb-al. C-deletion, in conclusion, can account or a large number o so-called

disharmonic and synharmonic spellings.


19/63


Table 6. Variable spellings occurring in equivalent morphosyntactic contexts.14

Morphosyntactic

Context

Spelling Variation for -VC

suffixes

Sources

[Possessed Noun]-il+

Possessor

7u-BAH ~

7u-BAH-hi ~

7u-BAH-hi-li

Yaxchilan Lt 24: B1b, G1a

Yaxchilan Lt 1: E1

Yaxchilan St 1

7u-TUN-ni ~

7u-TUN-ni-li

7u-LAKAM-TUN-ni ~

7u-LAKAM-TUN-ni-li

Nim Li Punit St 15: E1

Copan HStairway

Quirigua St 1: C3-D3

ikal St 12

yu-ki-bi ~

yu-ki-bi-la

[Hundreds o cases; e.g. Colas et al.

(2002)]

[Possessed.Noun]-al+

Possessor

ya-ja-wa ~

ya-ja-wa-la

onina Mn 31: B3

Rio Azul Plaque

[Possessed.Noun]-ul+

Possessor

ye-bu ~

ye-bu-li

Pusilha Stela H: A13

Copan Hieroglyphic Step

itles

7aj-bi-ki ~

7aj-bi-ki-la

Piedras Negras Trone 1

Piedras Negras St 12

ma-su ~

ma-su-la

Rio Azul jade earring

Early Classic incised bowl

7IX-sa-ja~

7IX-sa-ja-la

Yaxchilan Lt 14: C3

Yaxchilan Lt 54: C2

7a(j)-?-ni ~

7a(j)-?-ni-la

El Chorro Altar 2: D5, St: A915

El Chorro St: B3, Itzan St 17: I11

ma-ta-wi ~

ma-ta-wi-la

Looted Panel: D3

Palenque FC Alarda: B2

Proper Names

7a-7u-ku ~

7a-7u-ku-m(a)

Yaxchilan Lt 42: G3, Lt 3: G1, Lt4:

D5b

Yaxchilan 54: C2

7a-ku ~

7a-ku-la

Palenque I, West ablet: O1

Bonampak Stela 1: I, Stela 2: D7

APIR-la ~

ti-la ~

ti-la-la

Palenque Eves o House C and

Yaxchilan Lt 35: A7

Vase K1253

Vase K1442

sa-ku~

sa-ku-ma/la

Quirigua St E: D13

Quirigua St C, onina Mn 69


20/63

David Mora-Marn

. Te allacy o deault synharmony

Te DH, particularly DH I, also assumes that synharmonic spellings are unmarked

or deault. Tis is not the case. Justeson (1989) demonstrated that two types o

phonological contexts, C1VC1and CV7 roots, call or obligatory synharmony (a-ble 7; Figures 9ad):

Cases that appear to be exceptions to this rule can be explained as an anticipat-

ed spelling o the initial vowel o a ollowing word, C-deletion, or morphological

a b

c d

Figure 7. Presence or absence o explicit -Vl suffix in names o same scribes. Name o

same scribe illustrating C-deletion.

a. Scribal signature on Piedras Negras Stela 12. Drawing by John Montgomery at http://re-

search.amsi.org/montgomery.html.

b. Scribal signature on Piedras Negras Trone 1. Drawing by John Montgomery at http://

research.amsi.org/montgomery.html.

c. 7a(j)-bi-kil(a)title demonstrating presence o an -ilsuffix afer root bik. Detail o draw-

ing by John Montgomery at http://research.amsi.org/montgomery.html.d. 7a(j)-bi-kititle demonstrating underspelling o -ilsuffix as simply -i(l). Detail redrawn by

this author afer drawing by John Montgomery at http://research.amsi.org/montgomery.

html.
http://research.famsi.org/montgomery.htmlhttp://research.famsi.org/montgomery.htmlhttp://research.famsi.org/montgomery.htmlhttp://research.famsi.org/montgomery.htmlhttp://research.famsi.org/montgomery.htmlhttp://research.famsi.org/montgomery.htmlhttp://research.famsi.org/montgomery.htmlhttp://research.famsi.org/montgomery.htmlhttp://research.famsi.org/montgomery.htmlhttp://research.famsi.org/montgomery.htmlhttp://research.famsi.org/montgomery.htmlhttp://research.famsi.org/montgomery.htmlhttp://research.famsi.org/montgomery.htmlhttp://research.famsi.org/montgomery.html


21/63


22/63

David Mora-Marn

tsutscoati with some sort o -ihsuffix, possibly a weakened version o -il, as in

cases where final lis weakened word-finally in the Lowland Mayan languages.16

Tis rule, I would argue, can be extended to phonological sequences, not just

roots, as first suggested to me by Bricker (personal communication 2002).17 I

have ound supporting evidence or this hypothesis (able 8). Te evidence also

Table 7. Contexts or phonologically-conditioned synharmony (according to Justeson

1989).

C1VC1roots CV7 rootska-kaor *kahkfire

po-po or *pohpmat

ku-kuor *kuk(< *qu7q)

ta-ta or *ttthick (o liquids)

mo-7oor *momacaw

te-7e or *tetree

ti-7ior *timouth; edge

tsi-7ior *tsi7dog

a b c

d e

Figure 9. Examples o C1VC1roots.

a. Spelling ka-kaor *kahkfire. Excerpt rom Chichen Itza Monjas Lintel 4. Afer drawing

by Graham (1977a).

b. po-po-tsa-ma orpohp-ol, tsam-a(l)mat and throne. From Dresden Codex page 46. Afer

drawing by Markus Eberl in Knowlton (2002: 10).

c. c. te-7eor te7tree. From Madrid Codex page 42c. Afer drawing in Bricker (1995: Fig-ure 3a).

d. 7u-tsi-ior 7u tsihis dog. Afer drawing by Stuart (1987: Figure 13).

e. ka-ku-pa-ka-l(a)or kahk 7u pakalhis shield is fire. Chichen Itza Monjas Lintel 2 (at B1).

Afer drawing in Graham (1977a: 269).

. tsu-tsi-hior tsutsi(h)coati. Excerpt rom Kerr Vase 8076. Afer drawing in Houston et

al. (2001: Figure 2b).


23/63


shows that scribes were ree to spell C1VC

1sequences with a single C

1V sign quite

requently. Tis was the case with yu-ne, the more common spelling o y-unen

his child, which was occasionally rendered as yu-2

ne, with a reduplication dia-critic (D. Stuart 1985, 1988), as well as tsu-nu~ tsu-2nuor Pre-Cholan *tsunu7n

hummingbird, and even ka ~ ka-wa ~ 2ka-wa ~ ka-ka ~ ka-ka-waor Pre-Cholan

*kakawcocoa.

Table 8. More general phonological contexts or obligatory synharmony.

C1VC1Sequences CV7 Sequences

-le-l(e) or *-lelabstractive

yu-ne ~ yu-2neory-unenhis child

tsu-nu-n(u)or *tsunun(< PM *tsuunu7n)

hummingbird

si-na-n(a) or Yu. sn7anscorpion

(ka-)ka-wa ~ 2ka-waor Pre-Cholan *kakaw

cocoa

KABA77aor *kaabaa7name

a b c d

e g h

Figure 10. Examples o C1VC1non-morphemic sequences.

a. Spelling tsu-nu.

b. Spelling tsu-2nu.

c. yu-neon Copan Stela 63. Afer drawing by Marc Zender (1999: Figure 48a) afer Fash (1991).

d. yu-neon pottery vessel. Afer drawing by Marc Zender (1999: Figure 48d) afer Coe

(1973: 103).

e. yu-ne2on unprovenienced panel. Afer drawing by Marc Zender (1999: Figure 48e) afer

Mayer (1987: Plate 104).. yu-ne2on unprovenienced panel. Drawing by Marc Zender (1999: Figure 48) afer Mayer

(1987: Plate 54).

g. Spelling ka-waor *kkwcocoa.

h. Spelling kaor *kkwcocoa.


24/63

David Mora-Marn

I have urther identified two additional contexts, not previously discussed by

any author in relation to the possibility o obligatory synharmony, that demand

obligatory synharmonic vowel insertion (able 9; Figure 11). Kauman (personalcommunication, 2004) has discussed the spelling o +ichalready, yet; indeed as

a separate word in Classic texts, as in the case o KAL-la-j(a)(j)i-ch(i) , or kal-

aj-+ich it was wrapped already/indeed, and has suggested that this redundant

use ojVsyllabograms allows or the syllabographic spelling o a V-initial enclitic

as a separate word. As it turns out, this context coincides with one o the two ad-

ditional contexts or obligatory synharmony discussed next.

Te first examples, those o VCCV sequences, appear to be cases o vow-

el syncope (deletion) between consonants whenever a ollowing syllable (e.g. in

the orm o an added suffix) is present, a process well known or Proto-Cholan

(Kauman & Norman 1984). Tis rule accounts or all positional verb roots, which

typically take either a synharmonic -V1lstative suffix, or -CVCsuffixes such as -l-

aj(-i)and -w-an(-i). Te use o one o these rules with oreign words (e.g. cactonal)

supports its status as a phonologically conditioned process. In act, 12 o the 39

synharmonic spellings in the Houston et al. (1998) data set can be accounted or

in this way. Some synharmonic spellings, in particular those o nouns that do not

correspond to roots o C1VC1or CV7 shapes, cannot be readily accounted or in

this way, and I address those below. Te examples in able 9qr constitute cases oConsonant Insertion akin to the cases identified by Bricker (1985: 352) in the post-

conquest Books of Chilam Balam of Chumayel. She notes a series o spellings with

Table 9. Obligatory synharmony with vowel insertion in the spelling at morpheme

boundaries.

VCCV Sequences18 VC-V Sequences19

a. 7a-k(a)-ta or *ahk(o)t-a(j)to dance

b. yo-k(o)-bil(i)ory-ok-b-ilc. jo-ch(o)-bi-yaorjoch-b-i(y)(+a)

d. jo-l(o)-bi-yaorjol-b-i(y)(+a)

e. chu[k(u)]-ji-yaor chuk-(a)j-i(y)(+a)

. ti-l(i)-wa/wior til-w-a/i

g. cho-k(o)-wa/wior chok-w-a/i

h. 7u-pa-k(a)-buor u-pak-bu

i. 7e-k(e)-li-b(i) or ek-l-ib

j. 7e-k(e)-wa-ni or ek-wan-i

k. CHUM(-mu)-wa-nior chum-wan-i

l. pa-t(a)-wa-ni orpat-wan-im. xi-w(i)-te-ior xiutecuhtli(Nahuatl)

n. ta-wi-s(i)-ka-l(a) or huiz cal(Nahuatl)

o. ka-k(a)-tu-na-l(a)or cactonal(Nahuatl)

p. yi-l(i)(7)a-jiory-il-a-ji

q. 7u-PAT(-ta)(t)i-jior u-pat-ij(+i)r. KAL-la-j(a)(j)i-chior kal-aj+ich


25/63


a b c

d e

g h i

j k l

m

Figure 11. Newly defined contexts or obligatory synharmony.

a. 7a-k(a)-taor ahkt-ajs/he danced. Edzna Stela 18 (at A2-B2). Afer drawing in Stuart et

al. (1999: II-38).

b. Spelling tsi-b(i)-na-j(a)or tsihb-n-ajit is/was painted. Passage rom unprovenienced

ceramic vessel. Afer drawing by Stuart (1987: 52).

c. yo-k(o)-bil(i)ory-ok-b-il. Palenque emple XIX Platorm, West Side (at A3). Afer

drawing in Houston et al. (2001: Figure 6).

d. Spelling chu[k(u)]-ji-yaor chuk-(a)j-iy(+a)s/he is/was seized. From Yaxchilan Hiero-

glyphic Stairway 3, step I, tread, glyph A2. Afer drawing by Graham (1982: 166).

e. Spelling 7u-pa-ka-buor s/he placed it ace down or placed ace down (e.g. stone lintel),

on Early Classic lintel at the Nelson-Atkins Gallery in Kansas City. Afer drawing in Stuart

et al. (1999: 32).


26/63

David Mora-Marn

doubled consonants to spell -VC suffixes, as in ahau uobinstead o ahauob

or 7ajaw-oobrulers, and u mul lilinstead o u mulilor its mound. She also

compared these to the glyphic examples 7u-KIN-ni-leand ti-CHAN-na-li, which

would amount to u kin niland ti caan nali they were written in Colonial Yu-

catec orthography. In these cases, as Bricker explains, the -VC suffixes are spelled

as CVC syllables, in the case o the Chumayel, or CV-C(V) sequences, in the caseo the Pre-Columbian texts. More recently, Kauman (2004) noted his discussion

with Barbara MacLeod and Nikolai Grube o March 11, 2002, and subsequently,

in which theyi-chi ~ ji-chi collocation o the Primary Standard Sequence o dedi-

catory texts was discussed and defined as a spelling o +ichalready, indeed: tsi-

b(i) na-ja ji-chior tsihb-naj-+ich(writing-.-3s+indeed) it was

written indeed and GOD.N-yiyi-chi or GOD.N-Vy-+ich (VERB-.-

3s+indeed) it was GOD.Ned indeed. In each case, the vowel-initial enclitic

is spelled as a separate word with an initial consonant copying the immediatelypreceding consonant. Yet another example is present in Figure 11m): 7u WE7 7i-

bi, or 7uwe7-ib(3s eat-) thing or eating. Here the suffix -ibinstru-

mentalizer was spelled as its own word with an initial 7, which is also the final C

o the root we7which it inflects. Tese are also examples o Consonant Insertion

or the spelling o a -VC suffix or enclitic. Bricker notes that the same practice was

used in the Chumayelor the spelling o -V suffixes and enclitics, as with chac ce

instead o the alphabetic version chace, involving chakrain and +etopicalizer,

and muyal leinstead o muyale, or muyalcloud and +etopicalizer.

Lastly, an important number o synharmonic spellings are likely the result o

C-deletion, or o morphological conditioning, particularly when the typical su-

fixes o such roots were synharmonic. Tus, or some spellings o C1V1C1roots, it

. Spelling 7e-k(e)-wa-nior 7ek-wan-iit insert vertically (onto a wall). ortuguero Monu-

ment 6. Afer drawing in Wichmann (2002: 10).

g. Spelling 7e-k(e)-li-bior 7ek-l-ibvertically inserted panel. La Corona Panel 2B. Afer

drawing in Wichmann (2002: 9).

h. Spelling CHAK-xi-wi-te-7i, or Nahuatl xiutecuhtli. Dresden 49. Afer drawing in Bricker

(2000: 98, Figure 15b).

i. Spelling ta-wi-si-ca-la, or the first hal o Nahuatl tlahuizcalpantecuhtli. Dresden 49. Afer

drawing in Bricker (2000: 98, Figure 15a).

j. Spelling ka-ka-tu-na-la, or Nahuatl cactonal. Dresden 49. Afer drawing in Bricker

(2000: 98, Figure 15c).

k. Spellingyi-l(i)-7a-ji-yaory-il--j-iy(+a)s/he it has said it. From Palenque emple o

Inscriptions, middle panel, glyph H2. Afer drawing by Stuart (1987: Figure 36d).

l. Spelling (7/Y)IL(-li)-7a-jaory-il--(a)j(+a)s/he has seen it or il--(a)j(+a)it/s/he hasbeen seen. Excerpt rom Naj unich. Drawing by Barbara MacLeod in Stone (1995).

m. 7u-we-7i-bior 7u we7-ibhis/her thing or eating. Afer drawing by Zender (2000).


27/63


is possible that their -V1Csuffixes, whose vowel harmonized with the vowel o theroot, were sufficient to condition synharmonic spellings. For example, in the spell-

ing po-po-lo cha-ya(Figure 12), where the first term likely spellspohp-ol, possibly

o the mat, the suffix was synharmonic, and thus, cases where the spelling appears

as po-pocould be the result o C-deletion or perhaps morphological conditioning,

as explained next.20

. Morphological conditioning

Tere are two types o evidence or morphological conditioning: epigraphic (Mora-

Marn 2001, 2002, 2004) and linguistic (Kauman 2003a, 2003b). Te two almost

always reinorce each other, as shown in this paper, and in some cases one approach

can fill in the gaps o the other. Tis process involves the typical or most common

suffix o a root or stem as the strongest influence in the spelling o this root or

stem even when no suffix was required in a different morphosyntactic context. For

example, Justeson (1989) argued that the typical spelling o intransitive roots with

a Ciword-closing syllabogram (e.g. 7OCH-chi, hu-li) was based on the requent

suffix -i(h/y)completive status o intransitives (i.e. 7och-is/he entered, hul-is/hearrived (here)), and likewise, that the basign o tsi-bain 7AJ/7aj-tsi-b(a)was

due to the act that tsihbwriting is a noun that requires an applicative suffix -

to be transitivized; thus, its typical spelling as tsi-baeven in contexts where it was

not a transitive verb (e.g. 7aj-tsihbwriter) could be due to its typical suffixing.21

In other words, a root or CVC sequence could be spelled with a conventionalized

CV-CV sequence in which the vowel o the second syllabogram reflects the vowel

o the most common -VCsuffix a root might take (able 10; Figures 13 and 14).

Te vowel o the typical -VCsuffixes o these terms is maintained in spell-ings where no explicit suffix is apparent or needed by the morphosyntactic con-

text. Some o these spellings supplement Kaumans (2003a, 2003b) approach. For

instance, Kauman did not find evidence or the use o suffixes with Proto-Mayan

Figure 12. Name o prisoner on Yaxchilan Hieroglyphic Stairway 3, Step I (at E1-E2).

po-po-l(o) cha-ya. Afer drawing rom Graham (1979: 166).


28/63

David Mora-Marn

Table 10. Spelling alternations in differing morphosyntactic contexts, one requiring a

-VC suffix, the other not.

No explicit suffix22

Suffix neededa. 7u-mu-t(i) *mutbird CHAK-mu-ti-l(a)or chchk mut-ilred bird23

b. 7a-na-b(i) 7a[j]+nahbhe o the pool ya-na-bil(i) ory-a[j]+nahb-ilhis 7aj-nahb

c. 7u-to-ka7u-tok(-al)his flint 7a(j)-to-ka-l(a)or 7aj+tok-alhe o the flint

d. cha-k(i) chahklightning, thunder 7u-cha-ki-l(i)or 7u-chahk-il24

e. yu-hay-uh(-al)his bead 7u-ha-j(a)or 7uh-ajbead (o someone)

. 7u-tu-pa 7u-tuuphis earring tu-pa-j(a)or tuup-ajearring (o someone)

g. yo-7OL-lay-ohlhis heart 7o-la-s(i)or 7ohl-asheart

h. 7e-bu*7ehbstep ye-bu-l(i)ory-ehb-ulhis step25

i. KIN-n(i) *kinsun, day (7u-)KIN-ni-l(i)/l(e)or kin-ilby day; special

day or 7u-ki(:)n-ilhis/her day; special day,

KIN-ni-ch(i)or kinichSun God

j. 7u-WAY(-ya)7u-wayhis alter ego WAY-ya-l(a)or way-alalter ego

WAY-ya-s(i)or way-asalter ego/sorcerer

k. chu[k(u)]-ka-j(a)chuk-aj 7u-chu-ku-w(a) or 7u-chuk-uws/he seized it

a b

c d

Figure 13. Spellings o the term mutbird; omen.

a. 7u-mu-ti. Dresden Codex, D18b. Drawing by this author afer Bricker (1989: 48, Fig-

ure 4.12).

b. ?CHAK-mu-ti-l(a). Passage rom unprovenienced Early Classic ceramic vessel in the

shape o a bird. Drawing by this author.

c. Spelling o u mu(:)t-ilas u mution page 2 o Te Book o Chilam Balam o Chan Kan.

Afer photograph in Bricker (1989: 46).

d. Spelling o u mu(:)t-ilas u mutilon page 2 o Te Book o Chilam Balam o Chan Kan.

Afer photograph in Bricker (1989: 46).


29/63


a b

cd

e

g

hi

j k l m

n o

pq

r


30/63

David Mora-Marn

*tyooq(Lowland Mayan *took flint). However, in CLM texts the term is ound

possessed as 7u-to-ka (no explicit suffix) and unpossessed as 7AJ/7aj-to-ka-la,

a spelling which exhibits a suffix -al. Also, Kauman reconstructs Proto-Mayan

*muutbird; omen and notes it is attested as mut-alin zotzil. Tis differs rom its

spelling in CLM texts as mu-ti, as in 7u-mu-ti (Figure 13a); however, the phrase

CHAK-mu-ti-la(Figure 13b) is also attested and suggests that the spelling with

timay have been motivated by a suffix -il; the Colonial period texts in Yucatec

Maya discussed by Bricker (1989) in act exhibit variation o the spellings o this

term as u mut, u muti, and u mutil, which would correspond to logographic

Figure 14.

a. 7a-na-bi or a(j)+na(h)b(-il). Naj unich Drawing 29. Drawing by Barbara MacLeod in

Stone (1995).

b. ya-na-bili ory-a(j)+na(h)b-il. Naj unich Drawing 29. Drawing by Barbara MacLeod in

Stone (1995).

c. 7u-to-ka-pa-ka-la or u-tok, u-pakalhis/her flint, his/her shield. Yaxchilan Lintel 46 (F8).

Afer drawing in Bricker (1995: Figure 6j).

d. 7a(j)-to-ka-l(a) or aj+to(:)k-alhe o the flint. Afer Stuart et al. (1999: II-44).

e. to-ka-l(a) or to(:)k-al. Yaxchilan bone bloodletter. Afer unpublished drawing by Ian

Graham.

. cha-ki or chahkrain, thunder (Rain God). Excerpt rom Caracol Ballcourt Marker 3.

Afer drawing by Nikolai Grube in Chase et al. (1991: 5, Figure 3).

g. 7u-cha-ki-l(i)or u-chahk-ilhis/her/its chahk(rainy season?). Comalcalco brick. Aferdrawing in Martin et al. (2002: II-61).

h. CHAK-kior chahklightning. From Dos Pilas Stela 8. Afer drawing by Houston (1993).

i. 7u-CHAK-ki-l(i) or u-chahk-ilhis/her/its chahk(rainy season?). Vase o the Eleven

Gods, K7750. Afer photograph in Kerr (1999).

j. 7u-KIN-ni-le or u-kiin-il(+e)or u-kiin-il-e(l). Kabah Structure I, North Jamb (B). Afer

drawing in Bricker (1989: 40).

k. 7u-KIN-ni. Chichen Itza, emple o the Four Lintels, Lintel II, glyph E5. Drawing by Ruth

Krochock.

l. KIN-li-nior kiin-il. Chichen Itza 4 Lintel 2. Afer drawing by Ruth Krochock.

m. [KIN]chi-nior kiin-ichSun God. Simojovel shell. Afer drawing by Peter Mathews in

Schele and Miller (1986).

n. 7u-tsispelling on pottery vessel or utsgood. Pottery vessel K1453. Afer photograph in

Kerr (1999).

o. yu-tsi-li ory-uts-ilgood/goodness. Dresden Codex, D14b. Afer Stuart (1987: Fig-

ure 25b).

p. 7e-buor ehbstep.

q. ye-bu ory-ehb(-ul)his/her/its step. From Calakmul.

r. ye-bu-l(i) ory-ehb-ulhis/her/its step. From Naranjo. Drawings pr. afer drawing in

Schele (2000).


31/63


(7u-MUT > 7u-MUTIL), C-deletion (7u-mu-ti), and ull spelling (7u-mu-ti-li)

strategies o hieroglyphic writing, respectively.

In some o these cases we could be dealing with C-deletion, but it is likelythat quite a ew o these spellings with no apparent suffix are in act convention-

alizations. Tis is suggested by cases where it is clear rom linguistic evidence that

no suffix is necessary: (1) yi-tsi-n(a) ory-ihtsi(i)nhis younger sibling, and (2)

yi-cha-n(i) or y-ihchan his maternal uncle. Tese kin terms probably did not

take a suffix when possessed, as is the case in modern Cholan-zeltalan languages.

However, Kauman (2003a, 2003b) has shown that the vowel o the word-closing

syllabogram o these terms agrees with the vowel o the suffix that these kin terms

take in the modern languages in unpossessed contexts.

Table 11. Examples o kin terms and glyphs that take -Vl suffix in absolute orm but not

in possessed orm.

Hieroglyphic spelling Extant language data

yi-cha-ni Proto-zeltalan *ichan-il(-ab)to materno, zeltaly-ichansu

primo, Choly-ichansu to

yi-tsi-na zeltal ihtsin-lyounger sibling, k-ihtsin-abmy younger sib-

lings,y-ihtsinhis younger sibling

7i-ka-tsi ~ yi-ka-tsi zeltal 7ihkatz-ilcarga, 7ihkatz-in-elllevar carga,y-ihkatzsucarga

yi-cha-ki Proto-zeltalan *7ihchak-ilua, Chol ejkachua iy-ejkach-ilsu

ua

Tus, morphological conditioning is supported by epigraphic and linguistic evi-

dence, and it explains conventionalized spellings in the hieroglyphic texts. Fur-

thermore, some examples, such asyi-cha-ki(able 5), can be explained as cases o

conventionalization (i.e. the kioyi-cha-kias based on the probable unpossessed

orm *7ihchak-ilfingernail, claw (o someone)) or underrepresentation (i.e. the

vowel o kias a partial spelling o -ilo iy-ejkach-il).

Another set o examples is seen in Figure 15. Tese and other examples are

discussed by Houston et al. (2001), and also, in much more detail, by Zender

(2004). Tese various authors, assuming the DH I by Houston et al. (1998), sug-

gest that a spelling like 7u-tu-pashows disharmony in order to indicate a complex

vowel nucleus o the root, hence 7u-tu-p(a) or tuupearring. Tey also argue that

tu-pa-jais to be analyzed as tu-p(a)-AJ: the ao pais there simply as a diacritic

or the complex vowel nucleus o tuup. Te same is argued oryo-7OL-laand 7o-la-si: they analyze these spellings as yo-7OL(-la) ory-ohlhis/her/its heart and

7o-l(a)-IS or ohl-is heart, respectively. And finally, they analyze 7u-BAH-hias

7u-BAH(-hi) or 7u-baah and BAH-hi-ja as BAH(-hi)-AJ or baah-aj, as well


32/63

David Mora-Marn

as 7u-si-hias 7u-si-h(i)and si-hi-jaas si-h(i)-AJ or 7u-sihhis/her/its gif and

sih-ajgif, respectively. Nevertheless, the more conventional spelling rules would

lead us to the ollowing analyses: 7u-tuupversus tuup-aaj, 7u-baahversus baah-iij, 7u-sih versus sih-iij, and y-ohl versus 7ohl-as. Interestingly, the comparative

a b

c d

e

g h

Figure 15. Spellings o absolute and possessed nouns.

a. 7u-tu-paor 7u tu(G)phis earring. Chichen Itza Cenote jade earring. Afer drawing by

atiana Proskouriakoff.b. tu-pa-jaor tu(G)p-a(:)jearring. Palenque emple o Inscriptions, Center ablet (at A9).

Afer drawing in Schele (2000).

c. yo-7OL-laory-ohl(-al)his/her/its heart. Palenque emple o Inscriptions, West ablet (at

B7). Afer drawing in Schele (2000).

d. 7o-la-sior 7ohl-a(7)sheart. Unprovenienced shell at Yale Art Gallery. Afer Stuart et al.

(1999: II).

e. 7u-BAH-hior 7u ba(:)h-ilhis/her/its image/ace/portrait. ikal Stela 5 (at D4). Afer

drawing in Zender (2004).

. BAH-hi-jaor ba(:)h-i(:)jimage. amarindito Hieroglyphic Stairway 3, Step III (at E1).Afer drawing by Stephen Houston.

g. 7u-si-hior 7u sihhis/her/its gif. Itzan Stela 17 (at C1). Afer drawing by Ian Graham.

h. si-hi-jaor sih-i(:)jgif. amarindito Hieroglyphic Stairway 3, Step II (at E1). Afer draw-

ing by Stephen Houston.


33/63


Mayan evidence supports the possibility o both -a(a)jand -i(i)j, or instance, as

well as the existence o both -asand -is. Curiously, to my knowledge, no special-

ist has adduced explicit evidence or a orm -isin the script: all explicit phoneticspellings point exclusively to -as. Tereore, the interpretation o the use o 57 si

as a morphosyllable -ISis not clearly warranted. It seems simpler to assume that

the spellings o the possessed orms were influenced or motivated by the spellings

o the absolute orms: i tu-pa-jashows an -a(a)jsuffix, then the use o pain the

possessed orm 7u-tu-p(a) could be motivated by its use in the spelling o the

absolute orm.

Houston et al. (1998, 2001, 2004) and Zender (2004) have analyzed the above

examples as invariant sets showing -a(:)jand -i(:)ssuffixes. However, the spellings,

i read conventionally, ollowing Simple Spelling Rules, yield two variable sets -a(:)

j ~ -i(:)j, on the one hand, and -a(G)sand possibly -i(G)s, on the other. Tese orms

require some discussion. Fortuitously, as the preceding authors have pointed out,

these markers are attested outside o Cholan: Mam (-baj ~ -j), Kiche (-a:j ~ -i:j),

Kaqchikel (-aj ~ -ij, -tz), zutujil (-a:j ~ -i:j ~ -e:j), and Qeqchi (-bej ~ -ej) attest

to it in Eastern Mayan, while Jakaltek (-e) and Qanjobal (-e ~ -ej) attest to it in

Western Mayan. ogether with the Classic Mayan orms, we can say that the cog-

nate orms in this set o suffixes is reconstructible to Proto-Central Mayan, possi-

bly as *-a:j ~ *-i:j ~ *-e:j. Also, the act that the Kiche, Kaqchikel, and zutujil dataattest two different vowels (Proto-Greater Kichean *-a:j~ *-i:j) could suggest that

the spellings rom Classic Mayan texts (e.g. tu-pa-j(a) and si-hi-j(a)) could also

point to two different vowels and thereore two morphophonemically or lexically-

derived suffixes, -a(:)j ~ -i(:)j. So ar no spellings support the retention o *-e:j

orms in the language o Classic Mayan texts, although the possibility that there

might be such spellings is something that should be explored.

Also, the authors alluded to above read the 57 sisign as -ISin the context

o unpossessed nouns, as with 7o-la-si, ollowing the MH approach, which would

yield 7o-l(a)-IS. Interestingly, not a single spelling o the orm 7o-li-siis known, to

my knowledge, whereas several spellings o the orm 7o-la-si are known. Indeed,

most o the spellings where Houston et al. (1998, 2001, 2004) and Zender (2004)

propose a orm -IScan be read simply as -Ca-s(i)spellings o an -a(G)ssuffix, and

none as an -i(G)ssuffix: a Simple Spelling Rules approach yields 7o-la-s(i), with an

-a(G)ssuffix. Tis is a suffix o unpossessed nouns, and may be related to the ro-

zen suffix o the Proto-Cholan term *alastoy rom Proto-Mayan *a:la7stoy. Te

orm can be analyzed as composed o *al womans offspring, rom Proto-Mayan

*aal, and an apparently rozen suffix *-as, seemingly rom a Proto-Mayan orm*-a7s. Colonial Yucatec attests to a nominal term wayaslo que pasa de presto

como sueo (that which happens quickly like a dream), and wayasbasign(al),

parable; divine by means o dreams and sign(al)s (Barrera Vsquez 1980: 916,


34/63

David Mora-Marn

917), both probably based on a reflex o Proto-Mayan *wa(h)r, itsel likely derived

rom *warto sleep (Kauman & Justeson 2004: 1260). Tis term may be attested in

Classic texts, in the spellings discussed by Zender (2004): WAY-ya-si, analyzablehere as WAY-(y)a-s(i)or way-as, but as WAY(-ya)-ISor way-isby Zender. Te

latter spelling alternates with the possessed orm 7u-WAY-ya.26Te so-called ab-

solute suffixes o Mayan languages ofen convey clear meanings, such as uncertain

possessor or generic/plural reerence. In this case, one may take *al-asas meaning

o a child (uncertain possessor), in other words, toy (i.e. thing o a child). I so,

we can posit a rozen Proto-Cholan suffix *-as, which may have been productive

in earlier times. Tis suffix may be a reflex o Proto-Mayan *-atz~ *-itzabsolute

noun orms, as reconstructed by Kauman (1989: 8, Part B). Poqom, in act, shows

-is ~ -es ~ -bes, Kaqchikel shows -tz < -atz, (Colonial) Yucatec shows -tz(-il), and

Mocho shows -itz ~ atz ~ otz. Colonial Yucatec attest to the term wayas. In any

case, these sets o uncertain possession or generic/plural markers typically show

two or three distinct, morphophonemically-determined vowels; thus, the presence

o -a(G)s ~ -i(G)s, which is in act suggested by Simple Spelling Rules, should be

no surprise.

Conventionalization could be promoted by multiple reinorcement based on

different -VCsuffixes with the same vowel:

(1) Multiple Reinorcement

7u-BAH-hi-l(i)with -iland BAH-hi-j(a)with -ijshare vowel; both reinorce

use o hi

KIN-ni-l(i)with -iland KIN-ni-ch(i)with -ichshare vowel; both reinorce

use o ni

Interestingly, it is possible that some conflicting spellings may have coexisted or

some time, but one was ultimately picked as a result o requency o associated

suffix in the genre where it commonly occured:

(2) Competing Motivations

yu-ki-b(a) ~ yu-ki-b(i)ory-uk-ib3s-drink- his/her cup

7u-ki-baor 7uk-ib-ldrink-- cup (o someone)

yu-ki-biory-uk-ib(-il)3s-drink-(-) his/her cup

Te term or cup appears possessed in 91% o its occurrences; this act prob-

ably influenced the preerence or the spelling o this term with a word-closing

syllabogram bi(i.e.yu-ki-bi), given that in many o those possessed occurrences

a possessive -ilsuffix was likely present afer the instrumentalizing suffix -ib(yu-ki-bil(a)andyu-ki-bil(i) ory-uk-ib-il), though not necessarily all (some cas-

es oyu-ki-biwere probably ory-uk-ib). It is worth pointing out that in at least

a handul o instances on Early Classic pottery vessels rom ikal and a ikal-style


35/63


vessel that is unprovenienced one findsyu-ki-bias well asyu-ki-ba, in at least one

case within the same text (Figure 16ab). Te use o 501 bais not motivated by

the presence o a suffix -il; when possessed, it is plausible that only a suffix -ilcould

be added, given the evidence rom the modern languages. An explanation or this

spelling type comes instead rom the unpossessed orm o the term or cup: Chol

uch-ib-lcup. In unpossessed orm the noun or cup requires an absolute (un-

certain possessor) suffix -l. Is it possible, then, that the use o bawas motivated

by the presence o a suffix -lwhen the noun was not possessed? A ascinating

dedicatory ormula on an unprovenienced pottery vessel suggests this was in act

the case. On this vessel one finds several sentences and clauses. Te first clause is

made up o the Initial Sign (laythis), the GOD.N glyph (a dedicatory verb) and

the unpossessed term 7u-ki-baor uk-ib-(l). Tus, given the unpossessed orm,it is likely that an absolute suffix -lwas intended, even i underspelled. Tis clause

is ollowed immediately by a clause beginning withyu-ki-bi, the possessed orm

o the same term. Tis possessed term is then ollowed by its possessor, cho-ko

a b

c

d

Figure 16. Competing motivations in spelling o cup.

a. yu-ki-biory-uk-ibory-uk-(i)b-ilhis/her cup on Early Classic pottery vessel. Afer

photograph in Coe (1973).

b. yu-ki-baory-uk-ib, noty-uk-(i)b-il, on same Early Classic pottery vessel as example a).

Afer photograph in Coe (1973).

c. Passage beginning with 7u-ki-baor uk-(i)b-lcup (o someone), on Late Classic pottery

vessel. Afer photograph in Coe (1973).

d. yu-ki-bi, on same Late Classic pottery vessel as example c). Afer photograph in Coe

(1973).


36/63


37/63


conditioning approach on a larger data set will probably yield a higher rate o con-

sistency or that approach as well.

O the remaining 8 disharmonic spellings, 6 are unique (i.e. ba-tsu, ta-ji, chi-

ku, bu-la) or invariant (i.e. 7AYIN-na,ja-yi) spellings that cannot be used to test

the typical suffixing and morphological conditioning approach, which requires

spelling and contextual variation, and or which the linguistic approach cannot

adequately account because those terms do not typically occur with suffixes in the

modern languages. In act, the term spelled byja-yiis o uncertain etymology. Tisleaves only two requent disharmonic spellings unaccounted or at this time: -Co-

ma, and -7o-ba.27Tus, i we do not take the unique spellings into account, we can

say that morphological conditioning (typical suffixing, actual suffixing, consonant

deletion) accounts or all but 5.1% o the disharmonic spellings in Houston et al.s

(1998, 2004) data set.

From another perspective, the disharmonic word-closing syllabograms in -Ci

(18,52%) and -Ca(12,34%) significantly outnumber those in -Cu(5,14%); togeth-

er, -Ci and -Caspellings add up to 86% o all disharmonic word-closing syllabo-grams. Tis corresponds well with the much higher requency in the descendant

languages o vowel-initial suffixes that begin with ior a. In Yucatec, or example,

o the 44 vowel-initial suffixes listed by Bricker et al. (1998: 408), 13 (30%) be-

gin with i, 14 (32%) begin with a, 8 (18%) begin with e, and 3 (7%) with o; the

remaining 6 (13%) vowel-initial suffixes echo the root vowel, whatever that vowel

may be. ogether, i- and a-initial suffixes amount to at least 62% o all suffixes. In

modern Chol, based on data rom Aulie and Aulie (1999: 275278), one can find 5

-VC(VC) suffixes with a/, 6 with e, 8 with i, 5 with o, and 3 with u. Tough a lower

proportion than in Yucatec, the a-and i-initial suffixes still account or 48.2% o

the vowel-initial suffixes.

Table 13. Revised data set: not accounted or (including unique and invariable spellings,

as well as spellings o unknown etyma).

i a u Totals of disharmonic

spellings unaccounted for

i 3 (syn) 1(dis) 1(dis, unique) 2

e 0 0 0

a 2(dis, unique,

uncertain)

5 (syn) 1(dis, unique) 3

o 0 2(dis) 2

u 1(dis) 2 (syn) 1

2 4 2 8/39 = 18%


38/63

David Mora-Marn

. Remaining matters

. More on morphosyllables

Te DH approach requires the existence o so-called morphosyllables, or other-

wise certain necessary suffixes would be underrepresented or not represented at

all. Tat is, given a spelling such as BAH-hi-ja, i the hio BAH-hiis needed as a

diacritic to mark vowel length o the root ba(:)hhead, then it cannot be used to

spell part o the -VCsuffix that is needed here; instead, 181jamust do this on its

own, and must thereore be read as -AJin this context. Likewise, in a spelling like

7u-TUN-ni-li, i the nio TUN-niis needed as a diacritic to mark vowel length o

the root tu(:)nstone, then it cannot be used to spell part o the -VCsuffix needed

here; instead 24 limust do this on its own, and must thereore be read as -ILin

this context. Interestingly, in this last case, the suffix rendered would be the same

regardless o the spelling principle involved: 7u-TUN-(n)il(i)yields a suffix -il

(simple spelling rules); and 7u-TUN(-ni)-ILyields a suffix -il(morphosyllable and

disharmony approach). However, in the first case, the suffix rendered will have di-erent shapes depending on the approach: BAH-hi-j(a)yields a suffix -i(:)j(simple

spelling rules); and BAH(-hi)-AJ yields a suffix -a(:)j (morphosyllable and dis-

harmony approach). Tus, the approach invoked has substantive consequences

or the reconstruction o the grammatical structure o the language represented

in the texts.

Interestingly, in spellings like BAH-hi-ja and tu-pa-ja, the simple spelling

rules adhered to by epigraphers or the past hal-century yield two suffix orms:

-i(:)jand -a(:)j, respectively. Tis suffix, in either orm, would represent a suffixmarking uncertain possession. Tis marker is attested in Mayan languages with

both -i:jand -a:jorms. However, it is not preserved in any modern Cholan lan-

guage. Tus, the orms attested in CLM texts could simply reflect a retention o

Table 14. Revised data set: not accounted or (notincluding unique, invariable, or etymo-

logically problematic spellings).

i a u Totals of disharmonicspellings unaccounted for

i 3 (syn) 0 0 0

e 0 0 0

a 0 5 (syn) 0 0

o 0 2 2

u 1(dis) 2 (syn) 0

Disharmonic 2 2/39 = 5.1%


39/63


Central Mayan *-i:j~ *-a:j, making it unnecessary to assume, as Houston et al.

(2001) do, that only one orm, -a:j, was represented.

A crucial example given the attention paid to it by Houston et al. (2001) isthe case o the 7u-KAWIL-wi-la-lispelling on Yaxchilan Lintel 26 (Figure 17a).

Tose authors argue the spelling is analyzable, assuming both DH I and MH, as

U-KAWI:L-wi-l(a)-IL, or 7u-kawiil-il. In other words, the vowel o the la sign

would be silent and serve merely to indicate that the preceding vowel is complex

(i.e. either i:or i7or ih), while the necessary suffix, which the authors assume to

be o the orm -i(:)l, is spelled exclusively by the lisign, which they argue is a mor-

phosyllable -IL. Te point worth reiterating here is that or there to be a spelling o

the -Vlsuffix that this possessed orm o the KAWILglyph should take, the only

recourse available to these authors is to analyze 24 lias -ILin this situation. Te

reason or this is, o course, that the vowel o the preceding 178 la, by the DH I

hypothesis, is simply diacritical o the preceding vowels complexity, and thereore

fictitious. Consequently, the DH I requires the existence o morphosyllables.

Tere is, nonetheless, an alternative, one that assumes only simple spelling

rules, and thereore, that the spelling in question be analyzed as 7u-KAWIL-wi-

la-l(i). Tis alternative also takes into account the morphosyntactic differences be-

tween the context o this spelling and o other, more common spellings o the orm

KAWIL-la. First, it is worth noting that in contexts such as the one in Figure 17b),

ab

c

Figure 17.

a. Spelling 7u-KAWIL-wi-la-li. Excerpt rom Yaxchilan Lintel 15. Afer drawing in Schele

(2000).b. Spelling TSAK-ji-ya KAWIL-la. Excerpt rom Yaxchilan Lintel 25. Afer drawing rom the

Corpus o Maya Hieroglyphic Inscriptions by Ian Graham (1977b).

c. Spelling o ull sentence: 7u-TSAK-wa 7u-KAWIL-wi-la-li 7u-to-ka-PAKAL 7a(j)-

BUTZ 7o-CHAK-ki. Excerpt rom Yaxchilan Lintel 15. Afer drawing in Schele (2000).


40/63

David Mora-Marn

KAWIL-la appears unpossessed as the subject o a passivized or mediopassiv-

ized verb, TSAK-ji-ya, or tsa[-h-]k-[i]j-iy-(+a) conjure/cure[PASS]---

3s(+) it was/got conjured, while in the Yaxchilan Lintel 26 context, theterm appears possessed as the subject o the active transitive verb 7u-TSAK-waor

7u tsak-aw-s/he grasped (conjured) it, seen more ully in Figure 17c). In Mayan

languages in general, the most common orms o the possessive suffixes are -i(:)

land -a(:)l, and suffixes o different orms could be used depending on the type

o possession (i.e. ownership, intimate, inalienable, part-o-the-whole, inanimate/

associative, etc.), as well as abstractive suffixes o similar orms. Applying simple

spelling rules would yield an -a(:)l suffix, a hypothesis worth entertaining: 7u-

KAWIL(-wi)(l)a-l(i). Tis hypothesis, urthermore, would be consistent with the

typical spelling o the unpossessed term, in accordance with the ypical Suffixing

Principle: KAWIL-la. In the present example, the possessor o kawilin this con-

text is not human, but an inanimate entity 7u tok 7u pak-althe flint and the shield

o , which in turn is possessed by the spirit reerred to as 7o Chahk.28Te name

o the person who probably carried out the action o conjuring is not mentioned

until the next clause, and there that person appears as the possessor o a nominal-

ized verb (Mora-Marn 2004: 357358). Even i taken concretely, the possessor o

kawilis a material object. Inanimate or impersonal possession in Chol is carried

out with our allomorphs (Aulie & Aulie 1999): -al(e.g. 7i yixm-al chol-elel mazde la milpa), -el(e.g. 7i tye7-el 7otyotthe wood o the house), -il(e.g. 7i ytzm-il

tyumutla sal de los huevos), and -lel(e.g. 7i yotyot-lel 7iximla casa del maz vs. 7i

yotyothis/her house). Te hieroglyphic example, i analyzed as 7u kawil-al, where

-alollows a syllable with ias the vowel nucleus, would be consistent with the Chol

data, where -alollows a syllable with (underlying) 7i:7i yixm-al chol-el, i.e. /iy-

ixim-al chol-el/. Tus, taking -liin this context to be simply a morphosyllable -IL

may be a bit premature, and it may lead epigraphers and linguists down a blind

alley, one that would prevent them rom refining the morphophonemic analysis o

numerous grammatical affixes. Quite simply, simple spelling rules yield linguistic

orms that accord well with expectations rom the modern Mayan languages. It is

thereore not necessary to revise them at this point.

. Logosyllabic spellings, polymorphemic logography, phonetic

complementation

I morphosyllables are unnecessary, what can we make o spellings like BAH-ja,

ound in variation with BAH-hi-ja, and spellings like 7u-TUN-li, ound in varia-tion with 7u-TUN-ni-li? In other words, BAH-hi-jaand BAH-jaare somehow

spelling the same word; assuming so, and assuming too that BAH-hi-jawas meant

to be read BAH-hi-j(a)or ba(:)h-i(:)j, how are we to read BAH-ja?


41/63


Te answer lies in polymorphemic logography and the practice o phonetic

complementation.29First, logograms in the script could represent CVC roots or

CVC-VC stems (Justeson 1986, 1989). Day signs, or example, ofen representpolymorphemic words. Te day sign or the first day name is not supposed to be

read DAY7AJAWbut DAY7AJWAL, as shown by Mathews and Justeson (1

consonant deletion, obligatory synharmony, typical suffixing: an explanation of spelling practices...

Documents