| AUTHOR: Del Saz Rubio, Maria Milagros
TITLE: English Discourse Markers of Reformulation
SUBTITLE: A Classification and Description
SERIES: Linguistic Insights. Studies in Language and Communication. Vol. 60
PUBLISHER: Peter Lang AG
Katja Jasinskaja, University of Heidelberg / University of Stuttgart
This book by Maria Milagros Del Saz Rubio presents a descriptive study of the
English discourse markers of reformulation, e.g. 'that is to say', 'in other
words', 'or rather', etc. Its main goal is to provide a detailed linguistic
account and a classification of this group of linguistic units ''which help the
speaker and/or writer to go beyond the first formulation and produce a new one,
more in accordance with his/her communicative intentions'' (p. 15). The book
provides a thorough survey of previous descriptive and theoretical approaches to
discourse markers in general and to reformulation in particular, but the study
itself is theory-neutral, ''inductive'' and ''interpretive'', as the author
characterizes it, in that it starts with theoretically unprejudiced observations
on the usage of reformulation markers in corpora and draws generalizations from
them, rather than testing specific theoretically motivated hypotheses.
Apart from an introduction and a short concluding chapter containing final
remarks, the book consists of four chapters.
Chapter 1 presents a detailed survey of literature on discourse markers and the
notion of reformulation. It starts with a discussion of what the author calls
pre-1985 descriptive studies in English linguistics comprising Halliday and
Hasan (1976) and the Comprehensive grammar of the English Language by Quirk et
al., where reformulation is traditionally addressed under the label of
apposition. This is followed by the discussion of theoretical work on discourse
markers from the discourse coherence approach (work by Schiffrin, Redeker, Knott
and colleagues), work by Bruce Fraser, and relevance theoretic studies by Diane
Blakemore. Contrary to Blakemore's proposal and more in line with the position
of Fraser the author decides to include markers of reformulation in the category
of discourse markers, more specifically, the sub-class of *elaborative* markers
in Fraser's classification.
Chapter 2 is concerned with the question of what are the English discourse
markers of reformulation. First the author once again takes up the question what
a discourse marker is. Rather than defining discourse markers by their
grammatical properties, she considers the functional or pragmatic criterion to
be decisive, which allows discourse markers to be drawn from different
grammatical classes. Two properties are taken to be defining for the category of
discourse markers: (1) connectivity; and (2) non-truth-conditionality.
Connectivity is understood as the ability to relate two (usually adjacent)
utterances, an utterance to a preceding longer stretch of discourse, an
utterance to a speech act, an utterance to the non-verbal context. Discourse
markers are non-truth-conditional in the sense that, in accordance with a common
assumption in the literature, they do not contribute to the truth conditions of
the proposition expressed by the utterance they are part of, but rather have a
discourse organizing function. Further, a number of non-definitional properties
of discourse markers are discussed including the type of meaning encoded
(conceptual vs. procedural, cf. Relevance Theory), multi-categoriality, weak
clause association, sentence position (initial vs. non-initial), optionality,
etc., which the author does not take to be necessary conditions for discourse
markers, while some other proposals among the literature discussed do. Next, the
activity of reformulation is defined as ''reinterpretation of what is conveyed by
the previous discourse segment S1, or one of its constituents, in terms of what
is said, meant or implied'' (p. 82). This is intended to encompass paraphrasing
an expression, complete recasting the intended meaning, as well as revision of
an implication of a prior message. Finally, the author gives a list of about 60
English discourse markers of reformulation, which she divides into four major
classes: Expansion (e.g. 'that is to say', 'in other words', 'namely', 'for
example'), Compression (e.g. 'in sum', 'to recap', 'to conclude', 'in brief'),
Modification (e.g. 'or rather', 'more precisely/accurately'), Reassessment (e.g.
Chapter 3 gives a brief presentation of the research methodology of this study.
The preliminary list and classification of reformulation markers from Chapter 2
is the result of collecting relevant instances from the surveyed literature as
well as from corpora, texts and the internet. A further sub-classification of
the Expansion group is proposed on this basis: Elaboration, Identification,
Illustration and Exemplification. On the basis of this refined classification
the author restricts the set of discourse markers for the detailed analysis to
the most representative markers in each group. For these the author makes global
queries in the British National Corpus, and if the resulting set is two large a
random selection of 150-200 instances is taken for further investigation.
Finally, to study the relationship between various reformulation markers, the
author uses the substitutability test of Knott and Sanders (1998): a marker in
its natural environment of use is replaced by a different one and submitted to
native speaker judgments. Depending on whether this gives rise to an acceptable
sequence, synonymy, hyperonymy and hyponymy relations can be identified between
Finally, Chapter 4 presents detailed analyses of the selected reformulation
markers, in particular: expansion-elaboration ('that is to say', 'in other
words'), markers of simplicity/complexity ('in simpler/more technical terms'),
expansion-identification ('namely'), expansion-illustration –('for example',
'for instance', 'such as'), modification ('or rather', '(or) more
precisely/accurately'), reassessment ('(or) better/worse (still/yet)', '(or)
better said'), compression ('to recap(itulate)', 'in sum', 'to sum up', 'to
summarize', 'in conclusion', 'to conclude', 'in a word', 'in a nutshell', 'in
All analyses follow the same scheme. For each reformulator, first its syntactic
properties are described, which comprises its syntactic composition if it is a
multiword expression, its acceptability in sentence-initial, -medial and -final
position, and the possibility to occur together with other discourse connectives
such as 'and', 'but', 'so', 'then', 'or'. Second, a section on the scope of the
reformulator describes entities of which types it can take as S1, i.e. as what
is reformulated. This can include a constituent, a whole proposition, an
entailment, a presupposition, a speech act, a felicity condition, among others.
Finally, for each marker there is a detailed description of its environments of
use - a classification of contexts in which the marker can occur which can be
mapped to its pragmatic functions. Which context types are considered depends
ultimately on the marker under consideration. Most of the generalizations are
illustrated extensively with naturally occurring examples as well as constructed
ones, where necessary. In addition, for each major class of markers there is a
section on the relationships between the markers within the group as well as
with respect to other groups of reformulators. This is where Knott and Sanders'
methodology is applied and we can see which markers have more general and more
specific meaning, which are synonymous, hyponymous or hyperonymous with respect
to another, and in which areas there is an overlap in the usage of some of the
markers. For space reasons it is difficult to give a balanced summary of the
findings in this chapter, since the observations are varied and often specific.
Therefore I will just mention those which I found most interesting.
The most multifunctional reformulation markers are 'that is to say' and 'in
other words', whose possible functions (environments of use) include
Explanation/Clarification, Definition, Identification, and Compression, where
the markers are interchangeable. They differ in that 'in other words' is not
found in Illustration and Modification environments, whereas 'that is to say'
is. Plus, there is a syntactic difference: 'in other words' can be combined with
'or', whereas 'that is to say' cannot. Although both markers are initially
placed in the Elaboration subgroup of the Expansion group, the classification of
their environments of use makes clear that they share certain functions with
reformulators from other groups and subgroups. Finally both markers are excluded
in Reassessment environments characteristic of such markers as '(or) better
(still/yet)'. A curious difference is found between the forms 'that is' and
'that is to say'. In Modification environments markers like 'or rather', 'or
more precisely/accurately' can be replaced by 'that is' if only an utterance
constituent (rather than a whole utterance) is reformulated, whereas 'that is to
say' is less natural in such contexts. An interesting observation is made on the
differences between the markers of Modification '(or) more precisely' vs. '(or)
more accurately'. '(Or) more accurately' is more natural than '(or) more
precisely' in (a) cases where the reformulation concerns gradable or
quantifiable notions, such as numbers and figures; and (b) where S1 is
explicitly negated. In contrast, '(or) more precisely' is more appropriate where
the speaker is voicing his/her opinions, beliefs or assessment of the situation.
Among the Compression markers, 'to recap' works only in contexts which present
an 'objective' recapitulation of facts and does not seem to fit well in
environments with a higher degree of speaker involvement, such as presenting
his/her own opinions.
This book has a clear structure and definitely fulfills its goal of giving a
broad coverage to the English discourse markers of reformulation. It presents a
wealth of naturally occurring examples for different uses of the markers and
makes a number of sharp observations on often rather subtle pragmatic
distinctions between them (cf. above).
Still there are a few points I would criticize. First, in the sections where
reformulators are compared according to Knott and Sanders' methodology, there
are often not enough negative examples to support the proposed claims. For
instance, on p. 161 the author suggests that markers of complexity like 'in more
technical terms' stand in relation of hyponymy with respect to 'that is to say',
but strictly speaking this is not shown, since there is no example given where
substituting 'in more technical terms' for 'that is to say' leads to infelicity.
More instances of inappropriateness of a certain marker in a certain context
would make the analysis much stronger. Second, in the discussion of the
Identification group of reformulators as well as the Identification function of
reformulators from other groups, the usage of the terms ''definite'' and
''indefinite'' is rather non-standard. For instance, numerical expressions, such
as 'two phases' (p. 144), are classified as definite, whereas the demonstrative
'just that' (same page) as indefinite. Since this usage of terminology differs
radically from what is widely accepted among linguists (see e.g. Heim 1982), a
definition of the author's notion of definiteness would have been in order.
Third, there is one part of the analysis with which I rather strongly disagree -
it is the whole discussion of the markers of Reassessment. While I agree that
markers like '(or) better (still/yet)' can in principle function as
reformulators indicating that the new formulation is better than the old one, in
the absolute majority of the examples given in the book they are in fact used in
a different function. Rather than comparing the two formulations they give a
comparative evaluation of the situations or objects as such. In many cases the
discourse unit hosting the marker is a request (e.g. (385) on p. 189) or a piece
of advice (e.g. (392) on p. 192) or presents an object of desire (e.g. (393), p.
192), a condition ((389), p. 191) or an otherwise unrealized hypothetical
situation. In these contexts the marker simply seems to indicate that one
unrealized situation is more preferable than the other. The only marker in the
Reassessment group which, I think, indicates unambiguously that a better
formulation for the same concept or idea is proposed is the marker 'better
said'. As for '(or) worse (yet/still)', I doubt that it is able to function as a
reformulator at all. It is perhaps not surprising that there is no negative
counterpart for 'better said', like *'worse said'. There is simply no point in
giving a worse formulation. Finally, the book would have profited from slightly
better editing. There are not too many typographic and cut-and-paste errors, but
just a bit too many to pass unnoticed. Also the notation for unacceptable
examples is partly unsystematic, sometimes using ''X'' and sometimes ''*''. This is
confusing since it is not clear that a different kind of unacceptability is meant.
Despite these shortcomings, the book presents a valuable contribution to the
study of discourse markers of reformulation. With its clear classification of
English reformulators and the systematic presentation of naturally occurring
examples (except for the Reassessment group) it gives a good way of orientation
in the vast space of reformulation markers and will serve as an excellent
reference source for further theoretical studies, as well as for related work in
applied and computational linguistics.
Halliday, Michael A. K. and Ruqaiya Hasan. (1976) _Cohesion in English_. Longman.
Heim, Irene R. (1982) _The Semantics of Definite and Indefinite Noun Phrases_.
PhD Thesis. University of Massachusetts. Amherst.
Knott, Alistair and Ted Sanders (1998) The classification of coherence relations
and their linguistic markers. An exploration of two languages. _Journal of
Pragmatics_ 39, 135-175.
Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech, Jan Svartvik. (1985) _A
Comprehensive Grammar of the English Language_. Longman.
ABOUT THE REVIEWER
Katja Jasinskaja has written her PhD Thesis on the pragmatics and prosody of
implicit discourse relations, concentrating on the case of restatement
(reformulation). She is currently working as a researcher in a project
investigating the generation of narrative discourse. Her research interests
include pragmatics, discourse structure and semantics, the function of
intonation, and information structure.