LINGUIST List 6.964

Tue Jul 11 1995

Disc: Parameter of aspect

Editor for this issue: Ann Dizdar <dizdartam2000.tamu.edu>

Directory

1. Vincent DeCaen, Disc: The Parameter of Aspect

"On the Parameter of Aspect"

This posting continues the fruitful exchange subsequent to the initial
book review of C. Smith's 1991 "The Parameter of Aspect". In a
follow-up to that review, I noted an alternative to Smith's proposal
involving a simpler binary parameter for aspect that was in part the
basis for my 1995 dissertation on the Biblical Hebrew verbal system (U
of Toronto). Smith's reply pointed out the apparent empirical
disconfirmation by Mandarin and Navajo (treated in her work among
others); and this is where we pick up the thread.
 I do not grant the traditional analysis of Navajo and especially
Mandarin Chinese that is the basis of the objection; nor do I grant
the traditional analysis of the "Oriental" languages including
Burmese, Japanese, and above all, Biblical Hebrew representing the
classical Semitic systems. The earliest layer of the quasi-consensus
on tense, mood and aspect (TMA), flowing from Comrie 1976, 1985 to
present, is the early 19th century "Orientalist" framework that posits
(a) "tenseless" languages or "inflectional aspect" systems, and (b) a
definition of "perfectivity" based on the ambiguity inherent in the
concept of "completion" (global view as well as relative past tense);
on "Oriental" and "Orientalism", see e.g., E. Said's "Orientalism"
1978. I reject the Orientalist framework and the analysis of Semitic
systems which was extended to other "Oriental" systems, which was
extended worldwide to at least half of the world's systems; I
indirectly reject the traditional analysis of Mandarin and Navajo.

To clarify and make the discussion more concrete, I offer an analysis
of English in the Orientalist framework (*please* do not take this
seriously as an analysis of English). English, as all reputable
authorities agree, is "tenseless": it encodes only "aspect" (cf. Slavic
systems). Not only do we find arrested development, but actual
regression from the robust classical Aryan TMA systems. English has
but one distinction: -ed vs -s/-0, respectively perfective and
imperfective. The perfective signals "completion" vs the
non-completion of the imperfective. The severely impoverished system
of English signals the regression of the Anglo-American mind (witness
Reagan and Thatcher, etc., etc.).

We note several key elements that underlie all traditional analyses of
so-called "tenseless" languages. 1) TMA is encoded morphologically on
the Greek model or not at all: the "morphocentric fallacy." 2) The
perfective is defined as relative past tense (bypassing the revolution
in aspectology in the second half of the 1800s): the "aorist fallacy."
3) The relative hierarchy of languages, with Greek and Sanskrit on
top, Semitic near the bottom, and now the creoles on the bottom,
related to the relative development of "mind" and tied to a particular
pre-Darwinian interpretation of evolutionary theory.

English is not tenseless, nor is Biblical Hebrew or Quranic Arabic,
nor is Japanese, nor Burmese, nor Turkish, nor indeed Mohawk or
Haitian creole. They do differ from the standard European systems:
they systematically differ with respect to aspect in mirror-image
fashion. This difference is insightfully captured by Cowper's strong
claim for a simple binary parameter for aspect: a system default's for
an aspectual interpretation of its simple tense system, either
perfective or non-perfective. The aspect not "defaulted for" is
separately encoded (cross-linguistically in a limited number of ways).
European "tense" systems default for the non-perfective (apparently
the minority or marked option); non-European "tenseless" systems
default for the perfective as does English.

I simply extrapolate to a strong claim for Universal Grammar: all
systems encode tense (past vs non-past; I do not grant a "future
tense"), and at least irrealis/realis; and in addition the basic
system is configured by the binary parameter for aspect. Virtually all
systems outside of the European sphere default for the perfective
according to my studies; and I assume that is the unmarked setting for
UG. The major diagnostic among many is the (non-)obligatory
expression of the progressive: the perfective default must separately
encode the progressive. E.g., Mandarin must express the progressive
by the zai V construction (lit. "at V"): it defaults for the
perfective. Navajo and apparently the Athabaskan family as a whole
default for the non-perfective (with Algonquian systems, the only
real pocket of European-like non-perfective defaulters). This work is
summarized in my "Tenseless Languages in Light of an Aspectual
Parameter for Universal Grammar: A Preliminary Cross-Linguistic
Survey", forthcoming fall 1995, Toronto Working Papers in Linguistics.

This does not mean that individual lexica cannot increase the
complexity of the system, nor that the many interactions between TMA
and lexical classes are not also parameterized. I'm only talking about
the basic configuration of the TMA system for UG. Smith is right to
point to great complexity inherent in the systems she described in her
last posting.

re Mandarin. The difficulty is what counts as "inflection", and
whether "inflection" is not INFL of standard theorizing. I recognize
Mandarin -le as inflection; but not -guo, -zhe (Smith does not mention
V-(yi-)V, nor does she include V1-V2 compounds). The literature I use
is divided on the morphosyntactic status of -guo. But -zhe, I think,
does not in any way behave as "inflection": see among many sources,
Li, Thompson, "The Meaning and Structure of Complex Sentences with
-zhe in Mandarin Chinese" JAOS 96.4 (1976). It certainly is not a
"progressive"; but is perhaps some sort of "adjectivalization".

re Navajo. I think the jury is still out on these systems. But Rice's
work on Slave indicates the lexical nature of several classes of
morphemes in the Athabaskan verb complex. There is indeed a great deal
of complexity here. And since Athabaskan is really the only system
that does not fall out naturally from my proposal, it deserves great
attention. Either the whole project falls through, or we learn
something interesting about Athabaskan systems, or the proposal gets
modified in an interesting way: who knows. We learn nothing without
strong claims.

In summary: I reject any analysis that is "tenseless"; and/or defines
perfectivity in terms of relative past tense (with Comrie 1976). These
views embedded in the current consensus on TMA can be easily traced to
the early 1800s and the work on Hebrew and Arabic. I reject that early
framework (and the not-so-pretty cultural baggage that goes with it).
In its place I place Cowper's binary aspectual parameter default, and
extend the proposal not only to the other "Oriental" systems but to a
strong claim for UG's TMA system. According to my surveys, the only
real problem is what to do with Athabaskan systems: a project for the
near future (I hope).

I note that many proposals under other approaches would benefit from
the reduction and simplification of terms, concepts, etc. E.g., Bybee
et al's system can be streamlined; and symmetries emerge that appear
amenable to explanation. Descriptive/typological work is put on a
better footing as well.

respectfully,
Vincent DeCaen

c/o Near Eastern Studies Dept
4 Bancroft Ave., 3d floor
University of Toronto
Toronto ON M5S 1A1
CANADA

or

decaenepas.utoronto.ca