Review of Focus and Background Marking in Mandarin Chinese
Date: Mon, 30 May 2005 21:30:01 -0700
From: Shiaowei Tham
Subject: Focus and Background Marking in Mandarin Chinese
AUTHOR: Hole, Daniel
TITLE: Focus and Background Marking in Mandarin Chinese
SUBTITLE: System and Theory behind cai, jiu, dou and ye
SERIES: RoutledgeCurzon Asian Linguistics Series
PUBLISHER: Routledge (Taylor and Francis)
Shiao Wei Tham, Asian School I, Defense Language Institute Foreign
Language Centre, Monterey, CA.
I am grateful to Daniel Hole and Peter Sells for their help with this
review. Any errors are of course, my responsibility.
Focus and Background Marking in Mandarin Chinese is an extremely
comprehensive, detailed, and highly convincing study of four of the most
elusive words in Mandarin that continually frustrate second-language
learners and linguists alike. CAI, JIU, DOU, and YE occur in sentences
with a range of interpretations, including "ONLY", "EVEN", and other less
directly describable readings. This study provides individual analyses of
each of these words, then synthesises them into a coherent system of focus-
background marking in terms of quantification over alternatives in a focus-
In providing an overarching systemic analysis of focus-background marking,
this study is no longer an analysis of individual focus marking words in a
particular language, but rather yields predictions on possible analogous
systems in other languages, and on the kinds of information structural
relationships that may be lexicalized across languages. The study makes
dexterous use of up-to-date findings from formal semantics, couched in
precise, highly readable prose, but with minimal formal notation,
combining insightful theoretical proposals with copious amounts of data,
detailed analysis, and careful empirical motivation. As a result, this
book is not only a valuable reference for researchers of Mandarin Chinese,
but would also be of interest to theoretical linguists with varying
degrees of affinity towards formal semantics.
To give an idea of the complexity and depth of the study, I first provide
a brief glimpse into the range of meanings associated with CAI, JIU, DOU,
and YE. CAI and JIU feature in a range of sentences with seemingly
diverse meanings: for instance, CAI in (1) is interpreted as suggesting
that Xiao Wang came rather late, yet in (2) it is interpreted as
suggesting that 8 o'clock is early for now, which appears somewhat
contradictory. CAI also occurs in sentences containing ONLY-foci (3).
Xiao Wang ba dian CAI lai
Xiao Wang 8 o'clock CAI come
Xiao Wang came only at 8 o'clock.
xianzai CAI ba dian
now only eight o'clock
It's only 8 o'clock.
Lao Wang zhiyou cha CAI he
Lao Wang only tea CAI drink
Lao Wang drinks only tea.
The meaning of JIU is similarly difficult to pin down: indeed, it
sometimes seems untranslatable: while one might try to translate JIU
with "then" in (4), it is not clear what JIU in (5) would correspond to.
The author notes that this translation gap is not specific to English, but
is also true of other "common European languages".
Ruguo tianqi hao, wo JIU qu
if weather good I JIU go
If the weather is good, I will go there.
Women zai zher JIU neng wanr
we at here JIU can play
We can play here.
In contrast, the interpretation of DOU and YE sentences is more obvious:
these items appear in sentences with a meaning of EVEN (6). Yet it is
hardly obvious what effect DOU and YE achieve by their presence.
Lian Lao Wang DOU/YE lai.
even Lao Wang DOU/YE come
Even Lao Wang came.
The puzzling aspects of these words have not gone unnoticed, of course,
and many researchers have studied them in different combinations and to
varying degrees. Convention links CAI with JIU (for reasons to be
discussed below) (Biq 1984, 1988, Lai 1999); DOU is often studied in the
context of EVEN-sentences (e.g. Shyu 1995), although the author notes that
YE is relatively under-studied. Even where all of these words have been
studied (the author cites Alleton 1972), however, no attempt has been made
to relate these four words into a system of items that could be mutually
complementary or conflicting. The current work provides a largely
successful proposal for just such a synthesis. As a result, its
conclusions raise implications for focus-background marking devices across
languages, and are not limited in relevance to Chinese linguistics per se.
This study treats CAI, JIU, DOU and YE (henceforth CJDY) as information
structure devices that indicate particular relationships between the focus
and the background of a sentence. CJDY are called "parametric words" as
they are proposed to involve a choice among alternatives. The author
places these four words within a paradigm of focus-background marking that
forms a quantificational square. A major innovation of this study is the
proposal that CJDY are agreement words: they indicate (but do not bring
about) particular focus-background configurations. The viability of focus-
background agreement aside (and the author cites Yukagir as an example of
a language with such a system), the claim that Mandarin contains agreement
particles is controversial in itself, given the 'isolating' nature of the
language and the paucity of morphological devices. Whether agreement
analysis is correct or not, however, the observation that these markers
reflect, rather than create, specific focus-background relations, seems
astute and accurate.
Below, I provide a summary of each of the six chapters in the book, then
follow up with an evaluation section.
The first is a brief introductory chapter giving a cursory overview of the
phenomena to be dealt with, and provides an organization of the following
chapters (2-6). The main analysis is found in chapters 3 and 4. Four
distinct focus marking words are discussed, so each of chapters 3 and 4 is
divided into at least three major sections dealing with them relatively
independently. The first two sections are devoted to CAI and JIU
respectively. Because the uses of parametric DOU and YE overlap to a
large extent, discussion of these two items falls into the same section,
which is the third major section. In addition, chapter 4 contains a
fourth major section drawing these different threads into a paradigm of
focus marking devices in Mandarin that unites these four words as part of
Chapter 2 distinguishes parametric from other uses of CAI, JIU, DOU and YE
(henceforth CJDY). It is a very important chapter that delineates the
senses of CJDY to be dealt with in the rest of the book, separating these
from various other meanings associated with CJDY, and in so doing, it
imposes a clarity and organisation on the CJDY terrain hitherto unseen in
previous literature. Each of the various use types of CJDY is considered
in an independent section in chapter 2. This discussion yields a few
rather surprising conclusions, such as the proposal that the sense of
ONLY, which both CAI and JIU are associated with, is one of the parametric
uses of CAI, and thus part of the book's concern, but not for JIU, and
thus outside the purview of the book. Although surprising, this
conclusion is convincingly argued for via intonational, distributional,
and information structural properties of JIU to express a meaning of ONLY
(called the 'focusing' use of JIU), that contrasts with corresponding
properties of parametric CJDY. Section 2.5 then argues for identifying a
PARAMETRIC use type for each of CJDY, based on two major criteria of
obligatoriness and paradigmaticity. It is argued that parametric CJDY are
obligatorily realized in their sentence, whereas in their other uses, they
are not obligatory in their sentence. It is also proposed that CJDY form
a semantically coherent paradigm of quantification over alternatives, each
word constituting one corner of a quantificational square. This
second "criterion" is really the entire proposal laid out in chapter 4.
To argue for the existence of parametric CJDY based on an analysis of
these items as a paradigm of parametric words smacks somewhat of
circularity, but it is possible to view this assertion as a manifesto
rather than an argument and to treat the existence of a class of
parametric CJDY as a working assumption without affecting the progress of
the rest of the book, so no real harm is done.
Chapter 3 takes on the difficult task of capturing the conditions under
which CAI, JIU, DOU, and YE are obligatorily used or disallowed, although
the main aim of the chapter, as noted below, is to argue for the analysis
of CJDY as agreement markers that indicate agreement between specific
information structural categories and a background. For each of CJDY, the
chapter carefully details (i) the relative position of the information
structural category (usually a focus, although in the case of JIU a
contrastive topic is also possible) that interacts with CJDY; (ii) the
type of focus (e.g. 'only' focus or 'even' focus) that each of CJDY is
associated with; and (iii) how CJDY interacts with quantificational
expressions. The chapter ends (section 3.4) with a discussion of the
problem of categorizing CJDY. The author notes that CJDY have been
treated in various ways in the existing literature, e.g. as adverbs (e.g.,
Li and Thompson 1981, Paris 1981, 1985), focus particles or focusing
adverbs (e.g. Biq 1988), and as functional heads (e.g. Shyu 1995), but
often with reservations by the relevant authors. Hole points out empirical
and conceptual problems with each of these treatments, thereby laying out
the groundwork for the agreement analysis of CJDY.
Before proceeding to chapter 4, it should be mentioned that chapter 3
could be frustrating for some syntacticians (as the author himself notes,
p49), as many descriptive generalizations as to word order in CJDY
constructions and their interaction with information structure are laid
out without being given an explicit syntactic analysis. On the other
hand, the generalizations are presented with extremely clarity and stated
with great precision, so that this theoretical agnosticism does not affect
the main analysis, which is after all semantic and pragmatic, rather than
syntactic, and indeed, provides a useful database for potential syntactic
The meat of the analysis (and thus the book) lies in chapter 4, which
provides not only individual analyses for CJDY, but also a synthesis of
these four items into a focus-marking system which forms a paradigm of
quantification over "contextually-relevant" alternatives. Within this
system, the quantificational properties of CJDY make up a quantificational
square: CAI reflects negated existential quantification over the domain of
alternatives (no alternative will yield a true sentence); JIU reflects
negated universal quantification over alternatives (some alternative will
not yield a true sentence); DOU reflects universal quantification over
alternatives (all alternatives true), and finally YE reflects existential
quantification over alternatives (some alternative true). The chapter
draws together the individual analyses of CJDY into a paradigm of focus
quantification. As noted above, the four types of focus quantification
indicated by CJDY are related via a quantificational square. The rest of
this section demonstrates some of the predictions made by this paradigm
about the logical relationships between CJDY sentences (e.g. because of
their proposed quantificational properties, JIU (NOT ALL) forms a
contradiction with DOU (ALL)). The final section of this chapter save for
the summary looks briefly at two other parametric words HAI and ZAI, more
as a foil to the proposed system, and as an indication of potential
extensions to the paradigm, than as an actual part of the analysis.
Chapter 5 completes the picture by discussing issues of syntactic
structure and semantic scope relations. It contains first, a solution for
a mismatch between the semantic scope of modals with universal
quantification force (i.e. "must") that have wide scope semantically, but
occur in a syntactically subordinate position in CAI sentences. The
solution takes the shape of an unconventional mapping from semantics to
syntax: essentially, it allows the restrictor in a quantificational
structure to map to a subordinate position while the nuclear scope maps to
a matrix position. The author then takes on a relatively understudied
class of CAI and JIU sentences which encode an evaluation of the
desirability of affairs (e.g. zhe yang CAI dui 'this style CAI right':
roughly "it's only correct if (done in) this way"; zhe yang JIU dui
le 'this style JIU right ASP: roughly "that's right". These sentences are
described as providing a modal ordering on a set of worlds. A third task
of the chapter contrasts the multiply ambiguous English sentence "Three
people can move the piano" to the counterparts of each of its
intepretations in Mandarin. It is argued that each reading in English is
encoded by a different construction in Mandarin, distinguished by the use
of different parametric words (either CAI or JIU). Finally, the author
takes the analysis of CJDY a step further by considering the combinations
of two different parametric words in the same sentence.
Chapter 6, the concluding chapter, places the analysis of CJDY within the
wider context of focus marking in Mandarin, contrasting the position of
focus in CJDY sentences (typically pre-verbal), to the canonical post-
verbal position of focus in other sentences. It also traces a potential
historical source for CJDY-style focus agreement marking to contact with
This study provides a comprehensive, insightful, and responsible scrutiny
on information structure and semantics in Mandarin grammar, completed by a
careful consideration of the syntax-semantics interface. The system of
focus-background marking proposed raises typological implications for the
possibilities of analagous systems in other languages. Below, I raise
three points about the book that the reader might wish to be aware of, and
then proceed to a slightly more in-depth discussion of potential
shortcomings in the analyses. This discussion may perhaps be of greater
interest to researchers/speakers of Mandarin.
First, one notable aspect of the book is the important role given to
intonational factors in delineating sentence interpretations. Almost all
example sentences contain an indication of where intonational prominence
falls. In some cases, the intonational contour has the effect of
excluding certain readings. This is especially the case in the discussion
of JIU, and the native speaker reader should keep this point in mind when
sentence interpretations begin to appear rather too restrictive (e.g. p148
Second, it is somewhat surprising that there is no clear definition of the
term "parametric", given its importance in the study. The closest to an
independent definition is that of Biq's (1984), from where the term is
adopted (p.13 fn.1). On the other hand, there is sufficient descriptive
criteria that these parametric uses can be clearly identified, and if we
accept the major claims of the analysis, these parametric words do not
have any easily statable meaning but rely on the information structure of
the sentences they occur in for their "meaning".
Third, a brief note about the data is in order. This study provides vast
amounts of data, many which are attested, or adapted from attested data.
Elicited data is used, but not overwhelmingly, and the majority of
relevant examples are in agreement with my own judgements. I simply wish
to point out that various examples are taken from earlier works and thus
appear somewhat dated or stilted (e.g. p31 ex.52c would have been much
improved by adding an aspect marker GUO to both verbs in the sentence, but
this example is presumably adapted from a naturally-occurring source).
There are also occasional (very occasional) cases in which the data are
slightly inaccurate. But these are only very minor infelicities (some are
simply tone marking inaccuracies), and none of the very few inaccuracies
affect the argument they are intended to make. For instance, ex.31 on
p135 is, for myself, infelicitous because of the presence of RUGUO 'if' in
the conditional clause, but the sentence would be perfect without RUGUO.
Both in ex.31 and in the parallel ex.30, RUGUO may be removed without
affecting the judgements and the argument at hand. A rather more
potentially confusing typo is found on p165 (the first paragraph), in the
discussion of two contexts for ex.70. In this paragraph, Context 1 and
Context 2 should be substituted for one another.
Below, I discuss the analyses of CAI and JIU in a little detail. Among the
analyses of CJDY, the treatments of DOU and YE are relatively
uncontroversial. This is in part because DOU shares the same shape as the
distributive marker in sentences containing universal quantification, and
YE corresponds in form to an adverb meaning "also", that it seems
relatively unsurprising for DOU to be associated with universal
quantification and YE with existential quantification. The author takes
pains, however, to show that parametric DOU is distinct from distributive
DOU. Moreover, parametric YE is also treated differently from the YE that
means "also", called "focusing YE" (p42-44).
The analyses of CAI and JIU deserve greater attention partly because their
meanings, or at least the meanings of their non-parametric counterparts,
are far less easily statable than those of YE and DOU. In previous works,
analyses of CAI and JIU have been to a great extent intertwined, because
they seem to be synonymous in some contexts, and strangely antonymous in
others. For instance, both appear to mean "only" in (7), but while CAI in
(8) suggests that the leaving at 8 is felt to be later than expected, JIU
in (9) suggests that this event is earlier than expected. To add to
further confusion, CAI also seems to have another interpretation similar
to JIU in (9), exemplified by (10) below.
Wo CAI/JIU he le liang bei cha.
I CAI/JIU drink ASP two cups tea
I only drank two cups of tea.
Ta ba dian CAI zou
(S)he 8 o'clock CAI come
(S)he leaves only at 8 o'clock.
Ta ba dian JIU zou
(S)he 8 o'clock JIU leave
(S)he leaves (immediately) at 8 o'clock.
Xianzai CAI 8 dian.
Now CAI 8 o'clock
It's only 8 o'clock now.
Earlier analyses (e.g. Biq 1988, Lai 1999) typically go to great lengths
to unify these meanings, but the present work shows convincingly (chapter
2) that JIU in (7) is distinct from JIU in (9), and that only JIU in (8)
(but not in (7)) can be counted as "parametric", thus distinguishing
further the uses of CAI and JIU, and imposing unprecedented clarity on
their classifications. With this new classification, the analyses of CAI
and JIU are successfully extricated from one another. Interestingly, CAI
in (10) is unified with CAI in (8) --- this analysis of CAI with temporal
phrases I find in particular to be the most convincing and complete, not
only in this book, but also among previous work on CAI. Much of the
results of the proposed analysis of CAI are achieved from the precise
definition of "contextually-relevant domain of alternatives". This
definition is based on the notion of an "eventuality bag", which is the
set of (discourse-constrained) eventualities that have occurred up to the
point in time at which a sentence is evaluated. Exploiting this notion,
the analysis treats temporal adverbials occurring with CAI as uniformly
indicating an "until" interpretation, ingeniously unifying the apparently
contradictory (8) and (10). There is a slight shortcoming in the analysis
of CAI sentences, however. In conditional sentences, CAI has often been
noted to be associated with expressing a necessary condition. The
analysis of conditional CAI sentences is based on a detailed compositional
analysis of "only if" sentences that derives their semantics by first
reversing restrictor/scope relations to a universally quantified sentence
and applying existential quantification to this structure. This part of
the analysis is compositional and very explicitly laid out, but I would
have found it more useful to see a clear link between conditional CAI
sentences and temporal CAI sentences, and between conditional CAI
sentences and the negated existential meaning claimed to be associated
The most challenging task in the book is presented by the analysis of
JIU, which is also potentially the most controversial part of the proposed
system. JIU is argued to reflect "negated universal quantification" over
contextually-relevant alternatives (i.e. at least one alternative to the
focus will yield a false sentence). There are two aspects to the analysis
of JIU that are potential complicating factors. First, negated universal
quantification (NOT ALL) takes up that corner of the quantificational
square that is typically assumed to be non-lexicalized (potentially
because its meaning is expressible by existential quantification combined
with analytically encoded inner negation: EXIST NOT (Horn 1989/2001
p252ff, Hole in press p8). Second, JIU is also argued to interact not
only with focus but also with a notion of contrastive topic. The section
on JIU does an admirable job of presenting evidence for both of these
assumptions, especially in providing convincing examples and in yielding
the correct prediction that JIU cannot occur with the focus marker
LIAN "even", which the author suggests encodes universal quantification
over alternatives and is thus directly contradictory to JIU. I wish to
point out, however, that the analysis of JIU, while in large part
convincing, seems to miss one set of interpretations for JIU sentences.
The problem is raised by an example such as (160) (p238), which has two
possible interpretations. One is that given by the author, and predicted
by his account of JIU (see below). The other, however, goes against the
idea of negated universal quantification over alternatives.
zai ZHER women JIU neng wanr
be.at here we JIU can play
We can play here (there are places where we can't play)
In addition to the interpretation given, the sentence may also be
interpreted as providing an exemplar for a universal generalization; i.e.
it can be preceded by a statement of "we can play anywhere", to mean "here
is one of the places we can play at", which seems to contradict the
proposed hypothesis. This is perhaps the only real empirical problem that
one can point to in the entire book. Even so, the analysis of JIU remains
convincing for the examples cited towards it, and the prediction it makes
concerning the incompatibility of LIAN and JIU is very attractive. It is
conceivable that the current analysis could be maintained for at least one
set of JIU sentences, and the solution might ultimately lie in a non-
unified analysis of JIU.
To sum up, this is an outstanding piece of work that I would recommend
both to Chinese linguists, and to linguists interested in information
structure. By postulating a coherent system of focus-background marking,
and by positing CJDY as agreement words, the analysis underscores the core
status of information structural factors in Mandarin grammar, converging
with previous works such as La Polla (1995) and Li and Thompson (1976).
Alleton, Viviane (1972). Les adverbes en chinois moderne. Den Haag &
Paris: Mouton & Co.
Biq, Yung-O (1984). The semantics and pragmatics of CAI and JIU in
Mandarin Chinese. Cornell University dissertation.
Biq, Yung-O (1988). From focus in proposition to focus in speech
situation: CAI and JIU in Mandarin Chinese. Journal of Chinese
Linguistics 16 (1), 72-108.
Downing, Pamela and Noonan, Michael eds. (1995). Word order in discourse.
von Heusinger, Klaus, Pregrin, Jaroslav, and Turner, Ken P. eds. (in
press). Where Semantics Meets Pragmatics. Elsevier.
Hole, Daniel (in press) Mapping VPs to restrictors: anti-Diesing effects
in Mandarin Chinese. In: Klaus von Heusinger, Jaroslav Peregrin & Ken P.
Horn, Laurence R. (1989/2001) A Natural History of Negation. CSLI
Publications: Stanford, CA. Originally published: Chicago University
Press: Chicago, 1989.
Lai, Huei-Ling (1999). Rejected expectations: the two time-related scalar
particles CAI and JIU in Mandarin Chinese. Linguistics 37 (4), 625-51.
La Polla, Randy (1995). Pragmatic relations and word order in Chinese. in
Downing, Pamela and Noonan, Michael (eds.)
Li, Charles N. ed. (1976) Subject and Topic. Academic Press: New York.
Li, Charles N. and Sandra Thompson (1976) Subject and topic: a new
typology of language. in Li, Charles N. (ed.)
Li, Charles N. and Sandra Thompson (1981). Mandarin Chinese: a Functional
Reference Grammar. 2nd edition. University of California Press Berkeley,
Paris, Marie Claude (1981). Problèmes de syntaxe et de sémantique en
linguistique chinoise. (Mémoires de l'Institut des Hautes E;tudes
Chinoise XX.) Paris: Collège de France.
Paris, Marie Claude (1985). The semantics of CAI and JIU in Mandarin
Chinese. Ajia Afurika go no keisu kenkyu (Computational Analyses of Asian
and African Languages) 24, 181-96.
Shyu, Shu-ing (1995) The syntax of topic and focus in Mandarin Chinese.
University of Southern California dissertation.
| ABOUT THE REVIEWER:
ABOUT THE REVIEWER
Shiao Wei Tham is an assistant professor of Chinese at the Defense
Language Institute Foreign Language Center, Asian School I. Her research
interests lie in lexical semantics, information structure, and their
interaction. She has studied information structure in Chinese linguistics
through its different locative constructions, and through cleft-like
constructions. Her research interests include the encoding of locative,
possessive, and existential meanings across languages, and the
implications of such constructions for information structure.