Publishing Partner: Cambridge University Press CUP Extra Publisher Login
amazon logo
More Info

New from Oxford University Press!


It's Been Said Before

By Orin Hargraves

It's Been Said Before "examines why certain phrases become clichés and why they should be avoided -- or why they still have life left in them."

New from Cambridge University Press!


Sounds Fascinating

By J. C. Wells

How do you pronounce biopic, synod, and Breughel? - and why? Do our cake and archaic sound the same? Where does the stress go in stalagmite? What's odd about the word epergne? As a finale, the author writes a letter to his 16-year-old self.

Review of  New Reflections on Grammaticalization

Reviewer: Yury A. Lander
Book Title: New Reflections on Grammaticalization
Book Author: Gabriele Diewald Ilse Wischer
Publisher: John Benjamins
Linguistic Field(s): Typology
Issue Number: 13.2236

Discuss this Review
Help on Posting

Date: Thu, 5 Sep 2002 01:29:10 +0400 (MSD)
From: Yura Lander
Subject: Syntax: Wischer & Diewald, eds. (2002)

Wischer, Ilse and Gabriele Diewald, eds. (2002) New
Reflections on Grammaticalization, John Benjamins,
xiv+437 pp., Typological Studies in Language 49.

This book was announced on LINGUIST at:

Reviewer: Yury A. Lander, Institute of Oriental Studies, Moscow

Grammaticalization is a popular topic in current
linguistic investigations, and the volumes published
by John Benjamins do not play a secondary role here.
Thus, one can consider the volume reviewed here to be
a sequel of such collections as Traugott & Heine
(eds.) 1991; Pagliuca (ed.) 1994; Giacalone Ramat &
Hopper (eds.) 1998b.

At the same time, the debates about
grammaticalization obviously have reached a new stage
- where
(i) the very process of grammaticalization does not
seem to be so primitively uncontroversial;
(ii) new data concerning the grammatical development
may revise some of the well-established theses of the
grammaticalization theory;
(iii) the need to include grammaticalization into the
more broad theoretical picture has become apparent
(cf. Giacalone Ramat & Hopper (1998a: 1) on the
"growing reflexive interest in integrating
grammaticalization with theoretical work in
descriptive and historical linguistics").

All this is well reflected in "New Reflections on
Grammaticalization", which grew out of an
international symposium of the same name held at
Potsdam University in 1999. The book consists of an
introduction outlining the main topics to be
discussed, 24 papers, an appendix containing
bibliographical data of papers presented in the
symposium but published elsewhere, and a subject


The body of the volume begins with Christian
Lehmann's seminal paper "New reflections on
grammaticalization and lexicalization" (pp. 1-18),
which discusses the previously postulated opposition
between the two processes included in the title in
light of the contrast between analytic (roughly,
constructing) and holistic (roughly, fixing)
approaches to linguistic items. Lehmann argues that
lexicalization is a result of a holistic approach
whereby a sign is inventorized, while
grammaticalization is simply a sign's acquiring
"functions in the analytical formation of more
comprehensive signs". The two processes are therefore
cannot be opposed to each other, the more so that
grammaticalization can involve lexicalization.
Importantly, according to this view, one should not
merge degrammaticalization and lexicalization.

Although following Lehmann's paper in order, "More
thoughts on degrammaticalization" by Johan van der
Auwera (pp. 19-29) does not follow it in its basic
theses. Here "lexicalization is the making of a
lexical item out of something other than a lexical
item" (p. 21); hence, this notion has an
"extensional" overlap with "degrammaticalization", of
which a particular case is "the undoing of a
grammatical formative into something other than a
grammatical formative" (ibid.), e.g., into a separate
lexeme. This view allows, however, to separate a
"narrow concept of degrammaticalization" (with
grammatical formatives undoing into grammatical
formatives "with a weaker degree of grammatical
function" (ibid.)) from a "wide concept" which
includes cases of lexicalization. In addition, van
der Auwera touches upon some possible directions for
the research of degrammaticalization.

Jurgen Klausenburger ("Grammaticalization within a
theory of morphocentricity"; pp. 31-43) proposes a
theoretical scheme for accounting for various changes
occurring between the domains of syntax, morphology
and phonology from the position, according to which
it is the morphology that constitutes the nuclear of
universal grammar. After illustrating (with the data
of Romance languages) the "morpho-centripetal"
processes of "de-syntacticization" (which is equal
for him to "grammaticalization") and "de-
phonologization" and suggesting alternative analyses
for a few cases usually thought as reversal
processes, Klausenburger argues that most of what he
discusses should be thought as instances of "the
removal of characteristics". This concerns in fact
mainly the issue of unidirectionality of grammatical
change from less to more grammatical forms, which -
as Klausenburger declares - "has been rendered moot"
(p. 41).

The debates about unidirectionality also serve as a
background for Muriel Norde's "The final stages of
grammaticalization: Affixhood and beyond" (pp. 45-
65), which is devoted to the functional reduction
and/or change of function ("exaptation") of several
affixes in Swedish. It is shown that although a
number of such changes could be analyzed as
violations of the unidirectionality principle
(providing the cases of the development, say, from
inflection to derivation), they need a further
analysis which should take into account the overall
diachronic changes in a language (such as the loss of
case as an inflectional category in Swedish).

A similar point is given by Aydan Doyle in his
"Yesterday's affixes as today's clitics: A case study
in degrammaticalization" (pp. 67-81), who discusses
the evolution of Conomara Irish subject clitics from
Early Modern Irish agreement affixes and relates it
to the resetting of syntactic parameters (although
unlike Norde, in the generative sense) as well as to
the phonological change.

Bernd Heine ("On the role of context in
grammaticalization"; pp. 83-101) deals with the
problem of how a grammatical item comes to "create"
new (target) meanings from "original" (source)
meanings. Heine describes this process as a series of
stages whereby the use of a given grammatical item
extends, first, to linguistic contexts where the
target meaning is implicated, second to contexts
where the source meaning is canceled; the last stage
includes conventionalization of the target meaning.
Such a representation makes it possible to build a
typology of languages based on "how far" a concrete
grammaticalization path is passed.

The interrelations between grammaticalization and
linguistic contexts are further discussed in Gabriele
Diewald's paper "A model for relevant types of
contexts in grammaticalization" (pp. 103-120). While
using the development of the German modals as a
primary illustration, Diewald singles out a few
stages of a grammatical change which are very similar
to those suggested by Heine, but which include also
the initial expansion of a linguistic item to some
untypical contexts. Further, as Diewald emphasizes,
not only the source meaning is responsible for a
given diachronic development, but also changes of the
"older" lexical meaning on a par with restructuring
of entire grammatical subsystems.

Soteria Svorou ("Semantic constraints in the
grammaticalization of locative constructions"; pp.
121-142) is also concerned with the dependence of
grammaticalization on the source - but from a
different perspective. In particular, several
locative constructions are compared, and it is argued
that some of them are more likely to show advanced
grammaticalization (that is, "morphophonological
reduction, loss of autonomy, loss of obligatory
marking, and [semantic - YL] schematization" (p.
120)) than others.

As one can see from the very title of Gunter Lorenz's
corpus-based approach to the delexicalization and
grammaticalization of intensifiers in Modern English"
(pp. 143-161), this study deals with English
adjective intensification constructions, among which
several types are distinguished. Lorenz demonstrates
that intensifiers of some three types - namely
COMPARATIVE (e.g., 'extraordinary'), EVALUATIVE (such
as 'horribly') and MODAL (like 'certainly') - can and
do develop in English into the most unmarked SCALAR
intensifiers, while loosing their original meaning
and increasing in possible collocations. Finally, the
development of the intensifier 'really' to a "fully
grammaticalized adjective intensifier" (like 'very')
is discussed in details.

Gerda Hassler's "Crosslinguistic and diachronic
remarks on the grammaticalization of aspect in
Romance languages: Location and motion verbs" (pp.
163-179) deals with the expression of aspect by means
of periphrastic constructions including special forms
of locative and motion verbs of several types. A few
arguments are given for different degrees of
grammaticalization of these constructions, and this
is related to other functions of auxiliaries.

The close topic is touched upon by Philippe Bourdin
("The grammaticalization of deictic directionals into
modulators of temporal distance"; pp. 181-199), who
observes the diversity of "logics" governing the
development of ventive ('come') and itive ('go')
markers into the markers of "interval contraction"
and/or "interval expansion". This diversity, he
argues, makes it rather difficult to establish
UNIVERSAL grammaticalization paths, the more so that
in many cases it is even more reliable to speak about
some "notional ambivalence" rather than about

The central claim of Concepcion Company Company's
paper "Grammaticalization and category weakness" (pp.
201-215) is that grammaticalization is based on the
prototypicality of the human language categories in
that it "advances from the margins towards the focal
zone of the category" (p. 212), although not
obligatory reaching the latter. This idea is
illustrated with the expansion of overt direct-object
marking in Spanish.

Carol Lord, Foong Ha Yap and Shoichi Iwasaki in their
"Grammaticalization of 'give': African and Asian
perspectives" (pp. 217-235) discuss evolution of the
verb 'give' (with the particular attention to serial
verb constructions). A few grammaticalization paths
are established, which include not only the well-
known development of 'give' morphemes into
benefactive or permissive markers, but also the
subsequent grammaticalization into markers of
causative, reason, and some other categories.

Marianne Mithun's "An invisible hand at the root of
causation: The role of lexicalization in the
grammaticalization of causatives" (pp. 237-257) is
concerned with the grammaticalization of causative
markers from body-part morphemes in languages of
North America. The point is that such a path required
lexicalization of noun-verb complexes, providing a
possibility for the reanalysis of "original nouns"
into means/manner prefixes and then, to causative

Colette Grinevald ("Making sense of nominal
classification systems: Noun classifiers and the
grammaticalization variable"; pp. 259-275) takes up
the degree of grammaticalization of "noun
classifiers" (as opposed to, say, numeral classifiers
or genitive classifiers). It turns out that although
there are a number of different phenomena which are
sometimes covered by this term, they can be
represented as occupying different places on some
"grammaticalization axis", so that this parameter
should always be in mind when one compares nominal
classification systems.

"Phono-syntactic conspiracy and beyond:
Grammaticalization in spoken Beijing Mandarin" (pp.
277-292) by Liang Tao provides a detailed description
of the arising of a new pattern in Beijing Mandarin
Chinese, where the numeral 'one', which turns out to
be able to be used without a numeral classifier
(contrary to grammatical norms), nevertheless retains
some phonological traces of this classifier. It is
emphasized that this pattern became possible due to
the chunking of the sequence of the numeral and a
certain classifier.

Taru Salminen's "Retention of abstract meaning: The
essive case and grammaticalization of polyphony in
Finnish" (pp. 293-307) discusses the so-called
"quasi-construction" in Finnish and its semantic
relation to one of its original components, namely
the essive case marker. In common with some of
Lehmann's ideas, the importance of the
grammaticalization of a construction is underlined.

Sung-Ock S. Sohn ("The grammaticalization of
honorific particles in Korean"; pp. 309-325)
investigates the origin of the subject and dative
honorifics in Korean relating them to some locational
noun in support to the localist hypothesis.

The paper "From logophoric pronoun to discourse
particle: A case study of Finnish and Saami" (pp.
327-344) by Lea Laitinen opens the "discourse
markers" part of the volume. This paper is an attempt
to trace the history of the items 'hAn' and 'sun' in
Finnish and Saami correspondingly and to find the
contexts where these items could turn into discourse
particles as well as preconditions for such an

Dagmar Barth-Weingarten and Elizabeth Couper-Kuhlen
in their "On the development of final THOUGH: A case
of grammaticalization?" (pp. 345-361) discuss the
functioning of 'though' in (seemingly American)
English as a concessive marker and as a discourse
marker and give an analysis of the development of the
latter function from the former. In respect to
certain properties of this development, the authors
notice that it can be classified as an instance of
grammaticalization only if the very notion of
grammaticalization is considered to be organized
prototypically, so that the evolution of 'though' (as
well as of some other discourse markers) should be
"related to more prototypical cases of
grammaticalization in terms of family resemblance"
(p. 357).

The paper "Grammaticalization, subjectification and
the origin of phatic markers" (pp. 363-378) by Jose
Pinto de Lima explores the genealogy of phatic
markers by the example of Portuguese 'pois', which
passed a long way from a temporal conjunction to a
particle with a "metacommunicative" function. After
this case study, other possible sources for phatic
markers are briefly touched.

Heide Wegener's paper "The evolution of the German
modal particle DENN" (pp. 379-394) deals with a
similar grammaticalization story of a discourse
marker 'denn', which presumably developed from a
deictic morpheme through the stages of being a
temporal and causative adverb.

The collection ends with two theoretical papers. The
first is Wallace Chafe's "Putting grammaticalization
in its place" (pp. 395-412), which contains a model
for human language description based on the existence
of a "grammatical structuring" level on a par with
the "semantic structuring" level. Chafe argues that
it is the existence of this level that is responsible
for the grammaticalization where the meaning is built
from certain "quasi-meanings" (the idea that brings
grammaticalization together with idiomaticization).

The last paper is "Grammaticalization as an analogue
of hypothetico-deductive thinking" (pp. 413-422) by
Esa Itkonen. In this paper, the author tries to give
a general picture of grammaticalization as a process
based on the abduction (instantiated in reanalysis)
and rationality (reflected in extension).


The collection of papers presented in "New
Reflections on Grammaticalization" is certainly
outstanding in that it incorporates much theoretical
and empirical work, both providing a large amount of
non-trivial data and outlining directions for future
research. Since this volume is a result of a
symposium, its authors should not be in agreement nor
should their work be closely related. This allows a
reader to find a variety of observations which in
some sense, complement each other not only in
supporting one or other idea, but also in forming a
notion of the wide range of phenomena covered under
the label 'grammaticalization' and their possible
influence on theories. In connection with this, I
should commend the editor's achievements especially
in that how all these quite different papers are
ordered making the impression of a continuum of

This diversity seems to have a seamy side, however.
Perhaps, one possible reason of why so many different
facts are accounted under the rubric of
grammaticalization is that this very notion is
somewhat amorphous. Grammaticalization is usually
understood in two ways. The first is the unifying of
various processes on the basis of some of their
observable characteristics such as phonological
reduction, increase in frequency etc. The second
approach (close to what is advocated, for example, by
Bybee as well as by most theorists participated in
the volume under discussion) postulates the existence
of grammatical meanings (and/or functions) with the
origin and evolution of their expressions being the
main topic of the grammaticalization theory. Although
these two understandings are certainly related, there
is no perfect motivation for equating them, since (i)
there are examples of phonological reduction and
increase in frequency, which cannot be considered
instances of grammaticalization (e.g., the use of the
formant 'soc-' < 'socialisticheskij' 'socialistic' in
the Soviet Russian speech), (ii) in many cases what
is thought to be a grammatical meaning can be used
lexically (cp. 'most' and 'majority' in English; see
von Fintel 1995 for a discussion).

Evidently, in light of that we cannot identify the
two approaches, it may well be that no serious
theoretical account of grammaticalization is possible
until some definite criteria of what belongs (or is
likely to belong) to the domain of grammar are
obtained. Thus, for example, why should we consider
the development of evaluative intensifiers into
scalar ones (Lorenz's paper) or the rise of discourse
markers to be (or not to be) instances of
grammaticalization if we are not sure that 'very' or
'though' are grammatical formants? Barth-Weingarten
and Couper-Kuhlen suggest that we can think of the
development of 'though' as about non-prototypical
grammaticalization - but does this mean that 'though'

Maybe, the problem is in that not only
grammaticalization may (and must) affect linguistic
theories but also theoretical conceptions and models
may affect our understanding of the grammatical
change. This is an apparent motivation of problems in
defining grammaticalization, but the same goes also
with the interpretation of data. Consider, for
example, the former masculine singular nominative
adjective inflection -e in Swedish. As Norde puts it,
"it would be incorrect to say that Modern Swedish -e
is the same morpheme as Old Swedish -e, since it no
longer expresses nominative case, yet it continues to
denote masculine gender" (p. 52), hence this change
is classified as "functional reduction". Note,
however, that if we accept the view according to
which the nominative is the UNMARKED case, the data
may be interpreted as 'nothing has happened', since
although the Swedish case system has been lost, the
unmarked masculine form remained the unmarked
masculine form (and indeed, it would be quite
surprising to see, say, a descendant of the oblique
form instead of -e).

With this interdependence of theories and facts in
mind, let me turn to some more concrete problems
touched in the volume.

The idea of DEGRAMMATICALIZATION (concerned in the
papers by Lehmann, van der Auwera, Klausenburger,
Norde, and Doyle) as a process reversal to
grammaticalization brings us to the same problems as
those discussed above. Nevertheless, even if the
existence of degrammaticalization is taken for
granted (as van der Auwera does), it is almost
certainly a marked process, so that the
grammaticalization theorist should take into account
that since it is NOT NATURAL for anything (?) to
degrammaticalize, special factors are required for
stimulating such a grammatical change. This makes me
doubt that degrammaticalization can be viewed as
something uniform (even if grammaticalization is
such; cf. Traugott 2001).

That what happens usually is GRAMMATICALIZATION OF A
CONSTRUCTION rather than that of a single
word/morpheme has become an important point in recent
studies (e.g., Dahl 1998; Maisak 2002a; see also
Maisak 2002b). This is also reflected in a number of
papers of the volume reviewed here (see especially
the contributions of Lehmann and Salminen). The idea
is, of course, productive, and it is a common place
of many papers that some cases of grammaticalization
can be described only in this way.
But Lehmann makes a stronger claim that "the
grammaticalization of a construction does not entail
the grammaticalization of any of its component
elements" (p. 7). And further: "If there is an
element that mediates the relation between the
constituents of a construction, then
grammaticalization of the construction will involve
grammaticalization of this element. But if there is
no such element present, grammaticalization may
proceed, anyway" (ibid.)
I believe that THIS is true, and indeed for a long
time I thought that grammaticalization of a lexeme is
simply restoration of compositionality. Thus, for
example, in Modern Russian it is possible to say (1a)
as well as (1b), although compositionally (1a) should
be interpreted as a partitive 'three of these
soldiers' (see Lander 2001 for substantiation of this
point on the data of Austronesian languages):

(1) a. dvoe etix soldat
two these-Genitive soldiers-Genitive
b. eti dvoe soldat
these two soldiers-Genitive
'these two soldiers'

What has happened here, I believe, is the
grammaticalization of a CONSTRUCTION with numerals,
which (i) made this construction conventionalized and
uncompositional (in the sense of the Construction
Grammar), (ii) put some constraints on the
participants of the construction (e.g., that the head
must be a numeral; therefore separating the category
of numerals); (iii) while fixing the construction,
weakened constraints on the position of
demonstratives. What is important here is that the
grammaticalization of a construction leads to the
formation of grammatically relevant categories, and
in some cases this may be a closed class. Now, when
the compositionality recovers, this small class is
recategorized and grammaticalized. Thus, we can have
grammaticalization of a construction with or without
(subsequent) grammaticalization of some of its
components, and this goes in line with Lehmann's
But... there ARE cases where items are
grammaticalized without linguistic constructions.
Thus, I cannot imagine one concrete construction
where, say, Indonesian 'sahaya' (meaning 'servant')
has grammaticalized into 'saya' (meaning 'I'). So we
also have grammaticalization of morphemes/words with
or without grammaticalization of a construction. So
neither the role of a construction nor the role of a
word/morpheme should be overacted. But I think that
this is in fact the beginning of the long way towards
the well-established TYPOLOGY of grammaticalization.

considerable attention in the volume (see papers by
Heine, Diewald, Sohn, Laitinen among some others).
The basic idea is that some LINGUISTIC contexts may
"trigger" grammaticalization of certain items, which
are then, being already grammaticalized, can expand
to other contexts, more prototypical for their
original meaning (see Company's contribution). Now,
in connection with the previous discussion, one may
ask what are the relations between "constructions"
and "contexts", and it is surprising that this
question has not been asked in the volume. I believe,
however, that these are simply different views (so
for instance, a construction may be grammaticalized
in a context or as a context).
But even then, this cannot be the whole story, since
as we have seen in the case of the grammaticalization
of pronouns, there can be NO linguistic (vs.
extralinguistic) contexts for grammaticalization, so
again we see that the theories proposed work only for
a part of data.

Finally, one very interesting topic, which was,
however, only briefly touched upon in a few papers,
is PRECONDITIONS for a concrete grammatical change.
Note that there is a more or less established idea
that grammaticalization paths are universally
accessible (as is reflected in the "cognitive" status
of diachronic semantic maps; see van der Auwera &
Plungian 1988; Haspelmath (forthc.) for discussion).
A number of papers in the volume (including the
contributions of Diewald, Laitinen, and some others)
suggest, however, that a given development of a
linguistic item may depend on the very structure of a
language: thus, for example, the grammaticalization
of the morphemes meaning 'hand' to causative markers
described by Mithun would be hardly possible if North
American languages were dependent-marking.

As the very fact of having had the conference "New
Reflections on Grammaticalization 2" at the
University of Amsterdam (April 2002) shows, the
grammaticalization theory (or theories?) is not
futureless. And applying the principle of retention
to the linguistic theory, obviously this volume
already contain preconditions for future changes.


Dahl, O. (1998) Grammaticalization and the life-
cycles of constructions. Keynote lecture at the 17th
Scandinavian Conference of Linguistics, August 20-22,
1998. [Available at:

Giacalone Ramat, A. & P.J. Hopper (eds.) (1998a)
Introduction. In Giacalone Ramat & Hopper 1998b, 1-

Giacalone Ramat, A. & P.J. Hopper (eds.) (1998b) The
Limits of Grammaticalization. Amsterdam/Philadelphia:
John Benjamins.

Haspelmath, M. (forthc.) The geometry of grammatical
meaning: Semantic maps and cross-linguistic
comparison. To appear in M. Tomasello (ed.), The New
Psychology of Language, vol. 2, Mahwah: Lawrence
Erlbaum. [Available at:]

Lander, Yu.A. (2001) Austronesian partitives. In V.B.
Kasevich (ed.), 6th International Conference on the
Languages of Far East, South-East Asia and West
Africa: Proceedings and Abstracts of Papers,
St.Petersburg: SPbGU, 263-272.

Maisak, T.A. (2002a) Tipologija grammatikalizacii
konstrukcij s glagolami dvizhenija i glagolami
pozicii. [Grammaticalization paths of motion and
posture verbs.] Ph.D Dissertation, Moscow State

Maisak, T.A. (2002b) Review of Heine, B. & T. Kuteva,
World Lexicon of Grammaticalization. LINGUIST List
13.2166. [

Pagliuca, W. (ed.) (1994) Perspectives on
Grammaticalization. Amsterdam/Philadelphia: John

Traugott, E.C. (2001) Legitimate counterexamples to
unidirectionality. Paper presented at Freiburg
University, October 12, 2001. [Available at

Traugott, E.C. & B. Heine (eds.) (1991) Approaches to
Grammaticalization. 2 vols. Amsterdam/Philadelphia:
John Benjamins.

van der Auwera, J. & V. Plungian (1998) Modality's
semantic map. Linguistic Typology 2.1: 79-124.

von Fintel, K. (1995) The formal semantics of
grammaticalization. In J.N. Beckman (ed.),
Proceedings of NELS 25, vol. 2, Amherst: GLSA, 175-
189. [Available at:]

ABOUT THE REVIEWER Yury A. Lander is a research fellow in the Institute of Oriental Studies, Moscow. His current interests include the typology of noun phrases, Austronesian, Australian and Slavic linguistics. i