@article{oai:minpaku.repo.nii.ac.jp:00004308,
 author = {八杉, 佳穂 and Yasugi, Yoshiho},
 issue = {2},
 journal = {国立民族学博物館研究報告, Bulletin of the National Museum of Ethnology},
 month = {Oct},
 note = {This paper provides typological data on 73 Middle American
Indian languages (including some dialects). The study focuses
on: (1) the order of S(ubject)/O(bject)/V(erb); (2) adposition
order, i.e., the existence of Pr(epositions) vs Po(stpositions);
(3) the placement of the G(enitive) and N(oun) ; (4) the
A(djective) and N; (5) the P(ronoun) and N; (6) the D-(eterminative)
and N; (7) the Q(uantative) and N. The data
are summarized in Table 2.
I first discuss genetic classifications from a typological
standpoint. Major difference from the previous classifications
is the position of Tol. Since Tol shows typologically a mirror
image to Tequistlatec, it is separated from the Tequistlatec
complex. Interestingly, Seri has the same word order types as
Tol, and therefore both are worth comparing. Tlapanec is
classified as an Otomanguean, but Huave is typologically
different from the Otomanguean and is isolated. The
Otomanguean are typical VSO languages, and the word orders
are head-modifier. But Chichimec is a SOV language, and
Otomi and Mazahua are VOS. If Chichimec belongs to
Otomanguean, the Otomangean family is divided into three
major sub-branches, that is, Chichimec, the Northern (Otomi,
Mazahua) and the Southern (others). The Mayan are VOS
languages, but the Western Highland sub-branches are VSO,
the order of which is supposed to have spread from Mam.
Although the Uto-Aztecan and Mixe-Zoquean are postpositional
languages, /SO/V order varies. It is possible that they might
have been SOV languages but later changed word order under
the influence of Ianguages in a culturely defined area of Mesoamerlca.
 Greenberg set up 45 universal statements, one of which is
that Ianguages with dominant VSO order are always prepositional.
 He also recognized three basic positions of the verb:
VSO, SVO, SOV. Therefore it is worth noting that VSO-Po
languages, such as Northern Tepehuan, Cora, Huasteca Nahuatl,
and Black Carib, do exist, and there are many VOS Ianguages in
Mesoamerica.
 Next, word order diversity is examined, based on both genetic
classifications and areal distribution. Almost all Middle American
 Indian Ianguages are so-called incorporated languages.
The sentence nucleus, which consists of verb stem[v] and obligatory
affxes indicating subject [s], object [o], etc., is a kind of microsentence,
 a miniaturc of the Iarger construction. To explain the
diversity of S/O/V order, I compared the sentence nucleus with
S/O/V order. [S/O/V] order of the sentence nucleus was found to
be irrelevant to the change of S/O/V order, but it seems to be
more related to the other word orders than is the S/O/V order.
Further study of this should be rewarding.
 Mesoamerican languages share a number of linguistic
features that might have resultcd from diffusion or mutual
interaction. Areal distribution of word order types shows that
languages which have variant orders are bordered by the other
types of languages. This means that contact borrowing is
important to language change. In Mesoamerica there is no
SOV language, although the area itself is sandwichcd between
areas with SOV languages. Central Mesoamerica is occupied
by the Southern Otomanguean, which are VSO, head-modifier
languages. They are surrounded by SVO or VOS, modifierhead
 languages. Some word order challges are supposed to
have resulted from interaction between modifier-head languages
and head-modifier languages. For example, the change of
GN-AN to NG-NA is observed in Nahuan, and Classical Nahuatl
has Po/Pr, GN/NG and AN/NA. These co-occurrence sets are
assumed to refiect the transition from one historical stage, i.e.,
Po-GN-AN, to another, Pr-NG-NA, under the influence of
head-modifier languages. In regard to the P/N-D/N orders,
the Otomanguean, influence is weak, whereas on the G/N-A/N
orders, the Otomanguean are powerful. The word orders of
S/O/V,Po/Pr,N/A and N/G seem to be more sensitivc to diffusion
than are those of P/N, D/N, and Q/N.
The Southern Otomanguean are head-modifier languages,
but they have QN order. In the Zapotecan writing system of
Monte Alban I to III, from ca. 500 B.C. to A.D. 700, glyphs
precede numerals, which might reflect the ancient Zapotec NQ
order. If so, it is possible that the Southern Otomanguean were
totally head-modifier languages. Co-occurrence of Po/Pr,
GN/NG and AN/NA in Classical Nahuatl is supposed to have
occurred between the 14th and early-16th centuries, judging
from ethnohistorical data. It then changed to NG-NA order.
This might be mere guesswork, but there is no doubt that typological
studies of Middle American languages will contribute
not only to general linguistics (including areal linguistics, language
change, etc.), but also to the understanding of Mesoamerican
history.},
 pages = {259--378},
 title = {中米の言語の語順の類型論的研究},
 volume = {14},
 year = {1989},
 yomi = {ヤスギ, ヨシホ}
}