LINGUIST List 12.1786

Tue Jul 10 2001

Review: Frajzyngier & Curl, Reflexives; Reciprocals

Editor for this issue: Terence Langendoen <>

What follows is another discussion note contributed to our Book Discussion Forum. We expect these discussions to be informal and interactive; and the author of the book discussed is cordially invited to join in. If you are interested in leading a book discussion, look for books announced on LINGUIST as "available for discussion." (This means that the publisher has sent us a review copy.) Then contact Simin Karimi at or Terry Langendoen at


  1. Daniel Buring, review of Frajzyngier & Curl, Reflexives; Reciprocals

Message 1: review of Frajzyngier & Curl, Reflexives; Reciprocals

Date: Tue, 10 Jul 2001 18:31:07 -0700
From: Daniel Buring <>
Subject: review of Frajzyngier & Curl, Reflexives; Reciprocals

Frajzyngier, Zygmunt, and Traci S. Curl, eds. (2000) Reflexives: Forms
and Functions. John Benjamins Publishing Company, hardback ISBN
1-55619-653-9, xiii+286 pp, $85.00 (Typological Studies in Language 40)

Frajzyngier, Zygmunt, and Traci S. Curl, eds. (2000) Reciprocals: Forms
and Functions. John Benjamins Publishing Company, hardback ISBN
1-55619-654-7, xii+201pp, $79.00 (Typological Studies in Language 41)

Announcement at

reviewed by Daniel B�ring, UCLA

The present two volumes contain papers presented at the International
Symposium on Reflexives and Reciprocals held at the University of
Boulder, Colorado in 1997. The first of them deals with reflexives,
the second with reciprocals. Each presents studies from a variety of
different languages and theoretical frameworks. Their aim according
to the introduction is to '...analyze selected phenomena from
different theoretical perspectives, and ... present data that have
not hitherto informed theoretical discussions.' (vii-viii).

I will not discuss the papers in the order they appear. Rather, I
have grouped them thematically, although some of them might
legitimately be included in more than one group.

My first group contains the papers by K�nig & Siemund, Lyutikova, and
Maslova. These papers are primarily concerned with meaning.
 The papers by K�nig & Siemund and Lyutikova deal primarily
with the relation between the intensifier use and the
coreference-indicating use of reflexive pronouns. They try to account
for cross-linguistically common polysemies by finding a small
(ideally one-membered) set of core meanings for the
 Ekkehard K�nig and Peter Siemund's 'Intensifiers and
reflexives: A typological perspective' offers a wide range of
observations and generalizations, centered around its main question:
Why are intensifiers often (though by far not always) identical to
reflexives (e.g. English 'the queen herself arrived' and 'the queen
hates herself')? A prerequisite to answering this question is of
course greater clarity about the meaning of the two uses. The authors
explore the meaning of intensifiers, using examples from different
Germanic languages. Based on a typologically broader sample they then
show how even in languages with formally identical markers, the
intensifying and reflexive use can be distributionally distinguished
both sychronically and diachronically. Finally, the paper discusses
the distribution of different reflexive forms, often called SELF and
SE anaphors (such as Dutch 'zich' versus 'zich zelf'), critically
examining certain proposals found in the (mostly generative)
literature (see the second group below) and drawing attention to the
distinction between (prototypically) other-directed situations (e.g.
beating, hating, talking to and many others) and non-other directed
situations (e.g. grooming, defending, being ashamed) as an important
factor for the choice between the two forms.
 Ekaterina A. Lyutikova 'Reflexives and emphasis in Tsaxur
(Nakh Dagestanian)' addresses the very same main question, focusing
on data from Tsaxur, which has a pronoun 'wuzh' (here an henceforth I
will have to replace non-ASCII characters by strings of ASCII
characters), which can be used as an ordinary pronoun, an adnominal
emphatic marker, and, when reduplicated, as a reflexive marker. L's
proposal is essentially discourse-based; it is proposed that the core
meaning of 'wuzh' is 'unexpected return in discourse' and 'high
accessibility' (pp.248f), both of which are found in the emphatic and
the reflexive use. Unlike K�nig & Siemund's paper, L restricts her
discussion to Tsaxur, with an occasional glance at English, but she
seems to propose that the kind of general meaning underlying both
uses is found in all 'those languages that combine the two functions
within the same lexical entry.' (p.248)

Elena Maslova's 'Reciprocals and set construal' investigates the
question: Why are reciprocal and comitative markers the same in Bantu
and Yukaghir? Again, it is suggested that both uses can be analyzed
using a single core meaning. Unlike Lyutikova's paper, however, M's
approach is more semantic than discourse oriented. She submits that
the notion of a 'participant set' is central to explaining this
polysemy, which, for all I can see is a plurality. To substantiate
the claim, M provides a discussion of the different forms and
meanings of comitative and reciprocal constructions, the second of
which closely resembles certain aspects of the formal study in
Dalrymple et al. 1998, which M does not seem to be aware of. The
suggestion in the end is that the meaning of the marker is
'establishing the relation of inclusion between an item in the scope
of the marker and this P-set' (p.174), and that the differences in
meaning results from the different types of its semantic arguments
(relations for the reciprocal meaning, 'a participant itself' (an
individual, I suppose) for the comitative meaning).

The papers by Maldonado, Waltereit and Liu are somewhat special in
this collection in that they deal with a single language only. The
former two are similar to the previous three in that they attempt to
find common meaning for apparently different uses of a reflexive form.
 Ricardo Maldonado's 'Conceptual distance and transitivity
increase in Spanish reflexives' investigates 'se' marking in Spanish,
in particular those cases that do not signal semantic reflexivity. It
isolates four different uses of 'se' and attempts to give one basic
'focusing schema' for them (pp.180f).
 Very similarly, Richard Waltereit in 'What it means to
deceive yourself: The semantic relation of French reflexive verbs and
their corresponding transitive verbs' attempts to give unified
account of two seemingly different uses of French 'se', namely as a
coreference marker and marker of inherent, not necessarily semantic,
reflexivity. Both papers frame their proposals within the cognitive
grammar framework.
 Meichun Liu 'Reciprocal marking with deictic verbs 'come' and
'go' in Mandarin' discusses the V-lai-V-qu ('V-come-V-go')
construction in Mandarin Chinese, which is claimed to have a
'...potential function in marking the RECIPROCAL relation' (p.122).
While this construction is normally taken to signify 'repeated
action', L claims that 'cognitive or conceptual manipulation' (...)
and 'contextual reinterpretation' yield a reciprocal meaning (p.124).

The papers in my second group concern themselves with the
distributional syntax of reflexives (and reciprocals). Three issues
are central in this group: the distinction between what are commonly
called pronominals and anaphors, e.g. English 'her' vs. 'herself';
within the latter class the distinction between simplex and complex
anaphors (or SE and SELF anaphors) such as Dutch 'zich' and
'zichself'; and finally the issue of to which of these distributional
classes, if any, reciprocal pronouns belong.

In this group are the papers by Reuland, van Gelderen and Everaert,
all of which are set within the paradigm of generative grammar's
binding theory (as represented in Chomsky 1981), as well as the
papers by Abraham, which pleads for a hybrid between the former and
more functional accounts, and the paper on reflexives by Frajzyngier.

Eric Reuland's 'The fine structure of grammar: Anaphoric relations'
starts with an extensive introduction to the basic ideas and
methodology of principled based syntactic theories, including the
influential proposal in Reinhart & Reuland 1993, which is summarized
here. R illustrates the general program with a case study from Dutch
(with reference to a number of other, partly typologically unrelated
languages) regarding the question why 1st and 2nd person pronouns can
be used locally bound or free, where 3rd person pronouns use
different forms (reflexive versus non-reflexive). Rather than
assigning 1/2 person pronouns a category of their own or arguing that
they are anaphors and pronominals at the same time, R argues for an
analysis that exploits the technical notion of feature specification,
claiming that 1/2 person pronouns are not fully specified for all
grammatical features and therefore are not subject to the condition
that requires that pronominals be unbound in their local domain.
 The very same idea is also explored in Elly van Gelderen's
paper 'Bound pronouns and non-local anaphors: The case of Early
English', which is the only diachronic study in these volumes. G
traces back the historical development of the English pronoun system.
Of specific interest to her are those diachronic steps in which a
specific lexical form appears to be re-categorized in terms of its
binding theoretic status (e.g. when English plain pronominals began
to be used in non-local configurations only). G argues that even
though the form itself remains the same, its grammatical content
(phrased here in terms of so-called phi-features) within the overall
grammatical system changes, so that different principles of the
binding theory apply to them, resulting in a different distribution.
 What both papers argue, then, is that the categorization of
pronouns into anaphors and pronominals, with the pertinent
grammatical principles applying to them, should be replaced by a
finer grained theory in which those properties that define these
notions (such as full feature specification) are directly accessed by
the various principles.
 The paper by Martin Everaert 'Types of anaphoric expressions:
Reflexives and reciprocals', too, points in this direction. E
formulates a research agenda for reciprocals, namely to account for
the fact that reciprocals and reflexives do not share the same
binding domain (contra Chomsky 1981). Though his paper merely sets
the research program, it provides stimulating data from Greek,
Albanian and Basque, which challenge the traditional insights of
generative binding theory.
 In the same vein, Werner Abraham in 'The structural and
lexical space between reflexive binding and logophorics: Sundry
paradigms of reflexives and anaphora' discusses mostly Germanic data
(including various German dialects) and a little Russian, arguing
that binding principles along the lines of Chomsky 1981 and Reinhart
& Reuland 1993 cannot in principle account for cross-linguistic data,
and that a more flexible, partly pragmatic approach along the lines
of Comrie (1997) is called for.
 Finally, in 'Domains of point of view and coreferentiality:
System interaction approach to the study of reflexives', Zygmunt
Frajzyngier studies the distribution of different reflexive forms in
a variety of languages. Central to his paper is the idea that
different meanings for these forms will be expressed by different
morphemes, provided the language has them, but will be collapsed in
one morpheme if it doesn't, and that that differentiation proceeds
along a universal scale. F heavily utilizes the notions of 'point of
view' and 'affectedness' to capture those uses of reflexive pronouns
that do not seem to express argument coreference.

My third group contains typological studies in the basic sense of the
word, including the papers by Schladt, Heine, Frajzyngier,
Lichtenberg, McGregor and Gerdts.
 Two papers in this group, Schladt and Heine address the
question of grammaticalization. Mathias Schladt's 'The typology and
grammaticalization of reflexives' examines a sample of some 150
languages with regard to the lexical sources for reflexives. He
convincingly shows that there are regional preferences for the source
of reflexives, e.g. almost exclusively body-part words in African
languages, both body-parts and 'person' or 'self' in the Americas,
and words for 'return' or 'reflection', but rarely ever body-parts in
the languages of Australia (p.110f). S distinguishes three stages in
the synchronic body-part-to-reflexive development, from full NPs with
possessive via synecdoche (i.e. a part-expression denoting the whole
person) to exclusively reflexive (which might then develop into other
functions such as reciprocals).
 A similar study within the realm of reciprocals is presented
in 'Polysemy involving reflexive and reciprocal markers in African
languages' by Bernd Heine. H compares grammars of 62 African
languages from all major regions and language families with respect
to how they code emphasis, reflexivity, reciprocity, middle, and
passive. Relevant parameters are which part of speech is used (here:
nominal, particle, affix), and how phonologically heavy a word is
used. He furthermore explores how the different forms develop or
grammaticalize synchronically, and from what kinds of lexical
meanings (if any). He finds that the functions mentioned above form a
scale of grammaticalization, with full nouns being the starting point
and passive marking being the end. Morphosyntactically they go from
bona fide nominals to particles, clitics, and ultimately verbal
affixes, with decreasing phonological weight.

Zygmunt Frajzyngier's 'Coding of the Reciprocal Function: Two
solutions' probably belongs to the same group, as it contributes more
data on grammaticalization, showing that Chadic 'reflexive' (F. uses
scare-quote because he thinks these are really point-of-view markers)
are derived from the word for 'head', whereas reciprocals are derived
from the word for 'body'. He concludes that both collective markers
and 'body' words as sources of reciprocal marking have the same
'cognitive motivation': 'the exclusion of other participants' (p.192).

The remaining three papers in this group are devoted to more detailed
typological studies of reflexives and reciprocals.
 Frantisek Lichtenberk's 'Reciprocals without reflexives',
studies reflexive and reciprocal markings in a number of Oceanic
languages (a subgroup of the Austronesian family), as defined
formally as descended from a reconstructed Proto-Oceanic form. He
finds that in this sub-family reciprocal markers are not related to
or used as reflexive markers. He then goes on to study in detail the
different uses of the reciprocal markers found in these languages,
among them reciprocity, chaining, collectivity, iterativity, repeated
action, and middles (but crucially never reflexivity); even though
none of them are found in a single language, quite extensive
polysemies are claimed to exist (pp.48ff) . L conjectures that the
notion underlying the great majority (but not all) of the described
meanings of the marker is 'plurality of relations' (which I
understand as: the predicate applies to more than one participant or
tupel of participants) and 'low degree of elaboration', meaning that
the different sub-events and their respective participants tend to be
underspecified by the morphosyntax. L argues, following Kemmer 1993
that reciprocals can be connected to or share a polysemy with either
reflexives or collectives, but not both, suggesting two distinct
possible paths of grammaticalization.
 The other path is aptly illustrated in William McGregor
'Reflexive and reciprocal constructions in Nyulnyulan languages' (a
group of languages spoken in Western Australia). Here reflexivity and
reciprocity are expressed using the same construction (a verbal
circumfix). M provides a detailed study of the morphosyntactic
properties of these constructions and their interaction with the
applicative construction. The invariant meaning he proposes is
'self-directed action' (p.118).
 The last paper in this group is Donna B. Gerdt's 'Combinatory
restrictions on Halkomelem reflexives and reciprocals'. Halkomelem, a
Salishan language, has three relevant forms: reflexive, limited
control reflexive, and reciprocal. G. explains with great lucidity
and in detail how these different forms combine in the standard case
with verb roots denoting patient-oriented unaccusatives', from which
transitives (general or limited control) and causatives, as well as
reflexives (general or limited control) can be formed. G shows that
all three markers are morphologically composed of their respective
referential suffix together with the general transitivizing suffix.
At the same time, clauses based on verbs so prefixed are clearly
intransitive (their argument is absolutive marked, they can be
causativized etc.; the same result is found by McGregor for Oceanic).
G argues that this reflects a distinction between the semantic and
syntactic notion of transitivity: 'the transitivizing element allows
for two arguments in the predicate's semantic argument structure...
the intransitivizing element decreases the valency to a single
argument in the syntactic structure.' (p.142). G's paper concludes
with a discussion of the same suffixes occurring with verbs that
cannot be transitivized: reflexive suffixes as inchoative markers
(with statives rather than process unaccusatives); limited control
reflexives meaning 'manage to' (with unergatives), and reciprocals as
collective markers (with unergatives and statives), which she argues
involve lexicalized meanings.

As mentioned at the outset, the papers in these two volumes are
varied regarding their subject, goals and methods. As such, this
collection will be of interest to a large group of scholars, though
not all of them will find all of the papers equally relevant. I would
hope that the typological papers in these volumes are going to inform
further research in the formal tradition, both on the syntax and the
semantics of binding theory, and that, similarly, the more formally
oriented papers will inspire typologist to pursue in depth and
concise studies of binding phenomena.
 From a semanticist's point of view, the papers in this volume
raise a number of intriguing and fascinating questions, many of them
about the question of vagueness, ambiguity and underspecification of
lexical meanings. While most of the papers seem to converge on the
idea that it is preferable to assume one meaning per form rather than
to assume lexical ambiguity, there does not seem to be a consensus
about just when this goal is achieved. How can we determine that the
*same* meaning figures in different constructions, rather than a
class of meanings that are merely similar from a sufficiently remote
point of view; what, for that matter, is a sufficiently precise way
of formulating meanings (English paraphrases such as Lyutikova's
'unexpected return', or Lichtenberk's 'plurality of relations';
quantified logic as in Maslova's paper; or little circles, boxes and
arrows as in Maldonado's and Waltereit's papers)? Shouldn't we aim
for a full and formal meaning specification of truth and falsity
conditions that also predicts when a given form can *not* be used?
How are we going to tell different linguistic meanings from one
general meaning that just happens to describe a number of different
situations truthfully; how for example is the statement that '[t]he
major source for its [i.e. V-lai-V-qu's; DB] reciprocal reading lies,
not in lexical semantics, but rather in 'cognitive or conceptual
manipulation' (...) and 'contextual reinterpretation'' (p.124, scare
quotes there) different from saying that its meaning has nothing to
do with reciprocity, but that, as a matter of course, an adverbial
meaning 'day in day out' can be used to describe, among other things,
reciprocal events? What is a good way of factoring in contextual
components to a meaning?
 In my view, formal semantics has a lot to offer towards
answering these question, and I hope that some of the facts discussed
in this volume will see a reinterpretation in more formal semantic

Both volumes are carefully edited (with some minor exceptions such as
an offset endnote numbering in McGregor's paper) and include an
informative introduction (by Frajzyngier in the Reflexives volume,
anonymous in the Reciprocals one) summarizing and connecting up the
papers. As might have become clear from my summary, the papers differ
in terms of their quality and clarity, and some might have profited
from a critical review.
 Regrettably each paper comes with its own bibliography,
rather than referring to a common one, but fortunately and more
importantly, both volumes are complemented by three very useful
indices (subject, author, and language).

Chomsky, Noam (1981) Lectures on Government and Binding. Dordrecht:

Comrie, Bernard (1997) 'Pragmatic binding: demonstratives as anaphors
in Dutch'. Unpublished manuscript USC.

Dalrymple, Mary & Makoto Kanazawa & Yookyung Kim & Sam Machombo &
Stanley Peters (1998) 'Reciprocal expressions and the concept of
reciprocity'. Linguistics & Philosophy 21:159-210.

Reinhart, Tanya & Eric Reuland (1993) 'Reflexivity'. Linguistic
Inquiry 24: 657-720.

Daniel B�ring works as an assistant professor of linguistics at UCLA.
His specialization is formal semantics & pragmatics, in particular
intonational meaning and the syntax/semantics interface. He is
currently writing a textbook on the syntax and semantics of binding
theory, to appear with Cambridge University Press.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue