Review of : Narrog, Heiko & Heine, Bernd, The Oxford handbook of grammaticalization. Oxford, 2011

Abstract

This is a review of:
Narrog, Heiko & Heine, Bernd (eds.) 2011, The Oxford handbook of grammaticalization. Oxford: Oxford University Press (Oxford Handbooks in Linguistics).

Full text

CLIPP
Christiani Lehmanni inedita, publicanda, publicata
titulus
[Review of : Narrog, Heiko & Heine, Bernd, The
Oxford handbook on grammaticalization. Oxford,
2011]
huius textus situs retis mundialis
https://www.christianlehmann.eu/publ/rev_narrog_heine.pdf
dies manuscripti postremum modificati
04.02.2021
occasio orationis habitae
-
volumen publicationem continens
Beiträge zur Geschichte der Deutschen Sprache und
Literatur 135/3
annus publicationis
2013
paginae
442-456

Review of:
Narrog, Heiko & Heine, Bernd (eds.) 2011, The Oxford handbook of grammaticalization.
Oxford: Oxford University Press (Oxford Handbooks in Linguistics). ISBN 978–0–19–
958678–3. xxxiv + 877 pp.
For a couple of decades, grammaticalization has been firmly established as a major research
area in linguistics. It therefore appears timely that a volume of the Oxford Handbooks in
Linguistics should be devoted to it. Of the two editors, the second in particular has made
conspicuous contributions to that area, co-authoring, among many other things, a lexicon of
grammaticalization (Heine & Kuteva 2002). Moreover, the two editors already published a
handbook of linguistic analysis in the same series (Heine & Narrog (eds.) 2009), to which
they contributed a chapter on “Grammaticalization and linguistic analysis”. The contributors
to the present volume show a nice international distribution; many of them are renowned
linguists to whom the study of grammaticalization owes important advances. The editors have
aimed at broad coverage of the field and of approaches to it. In their introduction, they opt for
a liberal presentation of the issues and their possible solutions rather than for a theoretically
coherent account. Such accounts are provided in some of the other chapters, esp. by Harder &
Boy, Langacker and DeLancey. Moreover, the volume sounds out the intersection of
grammaticalization research with a couple of other strands of linguistics that had hitherto kept
rather aloof. Instead of discussing all the contributions to the volume, which space limits
forbid, this review will concentrate on the first three parts, which deal with theoretical and
methodological issues, and only mention in passing the last two parts, dedicated to specific
empirical domains and languages.
Critical survey
Part I of the volume, “Grammaticalization and linguistic theory”, discusses grammatical-
ization from different theoretical and disciplinary points of view. While this corresponds to a
systematic top-down arrangement of subject matter, the level of discussion here is highly
abstract, and illustration is largely confined to mentioning rather than analyzing examples.
This arrangement is, thus, not too helpful for an uninitiated user who would start reading on p.
1. If one looks for a chapter apt to provide introductory reading, a good candidate might be
DeLanceys chapter 29.
In their introduction, Heine and Narrog outline the structure of the volume, anticipating
some of the recurrent themes. On p. 11, they claim: “Personal pronouns, and more generally
person markers, belong to the most conservative parts of grammar.” This claim, if converted
into a falsifiable empirical generalization, is probably false, witness the kind of evidence that
the authors themselves recognize (renewal of person pronouns in Romance languages).
The chapter by Traugott on “Grammaticalization and mechanisms of change” is a rather
inconclusive set of high-level meta-comments on positions found in the literature. As a note
of self-defense, it might be mentioned that while the text (p. 25) attributes to Lehmann
2004:161 the claim that pure grammaticalization without analogy constitutes the norm as
against analogically-oriented grammaticalization, there is actually no such claim l.c., and in
fact, if ‘norm’ refers to the majority of cases, it is probably the other way around. On p. 22,

Lehmann, Review Narrog & Heine (eds.) 2011 2
the author claims that the pseudo-cleft allowing predicate focus in English originates about
1660 by reanalysis of a purposive construction. That cannot be so for a variety of reasons:
First, the predicate focus construction is much older. Second, the two readings of the example
provided are widely different, so general presuppositions for a reanalysis are not met. Third,
the purpose construction might have any verb in the relative clause, while the focus
construction requires do.
Olga Fischer deals with “Grammaticalization as analogically driven change?” First she
gets rid of reanalysis (p. 40): “Speakers do not reanalyse, they substitute one pattern
holistically for another.” Probably a mitigated version of this black-and-white scenario comes
closer to the truth: Unless there is an analogical model, reanalysis does not take place. Fischer
then downgrades grammaticalization to a particular kind of analogical change: “what
ultimately decides whether a linguistic sign becomes part of a user’s grammatical system is
whether it resembles in some way (semantically, formally or both) an already existing
category. Grammaticalization does not lead to new grammatical structures in any general
sense” (p. 42). In claiming this, she ignores both A. Meillet’s (1912:148) pioneering
hypothesis “il y a vraiment création d'outils grammaticaux nouveaux, et non pas
transformation” and the examples of pure grammaticalization (i.e. those not involving
analogy) provided in Lehmann 2004, §3.2. On p. 32, the author purportedly asks: “What
empirical evidence do we have for grammaticalization …?” However, she somehow manages
to exchange this question by the question of the underlying mechanism in the
grammaticalizing speaker-hearer (p. 33f), obfuscating, thus, the issue.
In the chapter on “Grammaticalization and generative grammar: a difficult liaison”, Elly
van Gelderen reviews the problems that generative grammar has had with grammaticalization,
the gradually increasing toleration of this field in the model, and finally argues (p. 45) “that
the child’s innate principles are in fact responsible for grammaticalization, and that generative
grammar can therefore gain much insight from grammaticalization processes.”
Peter Harder and Kasper Boye deal with “Grammaticalization and functional linguistics”.
For them (p. 58), “Grammaticalization has a central position in such a [usage-based] theory,
because it is the process whereby grammar emerges out of usage.” The authors propose a
coherent theory of grammaticalization. Grammatical expressions are defined (p. 60) as ones
that “are coded as discursively secondary in relation to one or more syntagmatically related
expressions.” In line with this, (p. 63) “Grammaticalization is the diachronic change which
gives rise to linguistic expressions which are coded as discursively secondary.” The concept
of discursive secondariness is rendered concrete by the restriction that grammatical items
“cannot under normal circumstances be used to convey the main point of a linguistic
message”, which notion is, in turn, operationalized by the criterion of stressability. While this
may be an important observation, it requires us to deny grammatical status to stressable
formatives like interrogative, demonstrative and (tonic) personal pronouns, modal verbs,
negators and many others. Moreover, the alleged structural counterpart of this semantic
aspect, “a dependency relation”, remains obscure.
In her chapter on “Usage-based theory and grammaticalization”, Joan Bybee starts
explaining the concept of chunking (known to most of the other authors as phraseology and
univerbation) by the example of the complex preposition in spite of. She correctly observes

Lehmann, Review Narrog & Heine (eds.) 2011 3
the decategorialization of the erstwhile noun spite in the process. However, by reducing
grammaticalization to chunking, she misses the distinction between grammaticalization and
lexicalization: this loss of analysability has nothing to do with grammaticalization and instead
is a definitional feature of lexicalization (s. the chapter by Lightfoot below). The criteria used
in order to measure the extent to which the phrase has been univerbated have been applied
systematically in German studies of complex prepositions since Beneš 1974. Moreover, while
it is true that both grammaticalization and lexicalization are reductive processes, there is an
essential difference with regard to syntactic structure: lexicalization destroys it by deleting
boundaries. Grammaticalization does not delete boundaries, it just shifts a construction to a
lower complexity level.
In his chapter on “Grammaticalization and cognitive grammar”, Ronald Langacker
reformulates familiar grammaticalization processes in terms of cognitive grammar. He offers
a semantic foundation for grammaticalization in the form of “conceptual compression”,
illustrating in detail with the grammaticalization of the be going to future. By subjectivity,
Langacker essentially means aspects of the speaker’s ongoing linguistic activity as opposed to
aspects of the referential world. Metaphorical extensions in grammaticalization often lead
from the latter to the former and would then come under the concept of subjectification.
Langacker’s descriptions of grammatical meaning are unrivalled, but he has little to say on
the formal side of grammar and grammaticalization. For instance, on p. 84f, he describes the
semantic bleaching leading to the grammatical functions of be and have, keeping silent over
their complementary distribution both in English and cross-linguistically. Nor are we told
about the role of be in the progressive form be going to.
In their chapter on “Construction grammar and grammaticalization”, Nicolas Gisborne
and Amanda Patten report on two cases of constructional change and try to show that they are
cases of grammaticalization. In construction grammar, grammaticalization is modelled as
schematization of constructions. The authors claim (p. 102): “The construction grammar
framework allows us to model the changes predicted in grammaticalization theory”. That
seems a bit saucy. In particular, the nature of grammaticalization as a reductive process has
not been accounted for.
In the chapter on “Grammaticalization and linguistic typology”, Walter Bisang
contributes to “a typology of manifestations of grammaticalization” (p. 106). In several South
East Asian languages, words change into grammatical formatives which, on the one hand,
show few symptoms of phonological erosion and, on the other hand, allow diverse
interpretations on the basis of inferences. These languages, thus, do not display co-evolution
of form and meaning. Moreover, such formatives do not form paradigms; grammaticalization
manifests itself more in syntagmatic rigidity.
Writing on “Grammaticalization and sociolinguistics”, Terttu Nevalainen and Minna
Palander-Collin report on the time that the second phase of Jespersen’s cycle (discontinuous
negation becomes simple negation) takes in different Germanic and Romance languages, on
the social evaluation of English adverbs provided by -ly and on the complication of the spread
of parenthetical I think by pragmatic factors. All in all, no differences between the spread of
grammaticalization and other kinds of grammatical change have been observed.

Lehmann, Review Narrog & Heine (eds.) 2011 4
Holger Diessel, in “Grammaticalization and language acquisition”, compares the fate of
grammatical elements in the two processes and finds parallels in the semantic development,
but no analogue in language acquisition to other factors of grammaticalization. It is, therefore,
not clear that semantic changes applied by children have much to do with grammaticalization.
In “Grammaticalization and language evolution”, Andrew D.M. Smith elaborates on the
cognitive mechanisms underlying metaphor and reanalysis and claims these to have been
operative in the evolution of “ostensive-inferential communication”. It seems that the theory
of language evolution can draw richer insights than that from grammaticalization theory.
In “Grammaticalization and linguistic complexity”, Östen Dahl proposes a Neo-
Humboldtian theory of evolutive typology (developed in earlier work), where long-running
grammaticalization finally leads to a “mature”, complex state, displayed by the flexive type.
In their chapter on “Grammaticalization and directionality”, Kersti Börjars and Nigel
Vincent offer a circumspect discussion of the problem of degrammaticalization, all too often
dealt with cursorily and without sufficient theoretical and empirical foundation. Their
conclusion is that there are very few genuine examples of degrammaticalization; but they
make no choice among the alternate explanations for unidirectionality. Their treatment could
have benefited from acquaintance with Lehmann 2004 (referred to by other contributors to the
volume), which puts a number of common misunderstandings of the issue straight. In
particular, German zig ‘umpteen’ is an invalid example of degrammaticalization not because
it is a “citation form” (which it is not), but for the reasons mentioned o.c. §4.1.
In Marianne Mithun’s chapter on “Grammaticalization and explanation”, the objective is
not to explain grammaticalization, but to show how an account in terms of grammaticalization
can explain (better than alternative approaches) certain synchronic facts, and in particular
Navajo verbal morphology, used here as a specimen. In principle, the approach is sound and
illuminating. It is perhaps necessary to add that affixes and affixal paradigms do not
necessarily accrue one by one to a stem. The following scenario is also found: At one stage,
there is a periphrastic construction of an auxiliary and a non-finite verb, where some material
(e.g. pronouns) agglutinates to the left and the right of the auxiliary, while other material (e.g.
adverbs) accrues to the full verb. At a later stage, the complex auxiliary form univerbates with
the full verb, leading thus to a layered rather than linear morphological structure. Such a
development is known from Swahili. All in all, this chapter is more a (rather detailed)
(diachronic-)descriptive account of Navajo verb morphology than a chapter on
grammaticalization and explanation.
In “Grammaticalization: a general critique”, Brian D. Joseph is concerned that research in
grammaticalization amounts to an undue “privileging of one cluster of changes over others”
(p. 198), reminding us of the existence of many other types of linguistic and, in particular, of
grammatical change. The chapter contains a large set of rather misplaced criticisms, e.g. (p.
200): “This principle of unidirectionality is generally viewed as foundational for
grammaticalization” in the face of Lehmann 2004, §4.2 putting this straight; and there are,
once again, spurious examples of degrammaticalization, e.g. hiccoughing being reanalysed as
hicking up. However, some persuasive examples are adduced of changes both from
derivational to inflectional morphology and vice versa. If one considers the former of these

Lehmann, Review Narrog & Heine (eds.) 2011 5
changes as grammaticalization – as does Wischer on p. 361 –, then the latter automatically
becomes a kind of degrammaticalization (as discussed by Norde, p. 482).
Part II “Methodological issues” shows convincingly how sound scientific method serves
the case of grammaticalization theory. In particular, quantitative analyses of the distribution
of grammaticalized items and constructions over centuries, as they are possible in well-
documented languages, demonstrate beyond reasonable doubt that grammaticalization is a
process which may proceed in a series of small steps, all in the same direction. No attempt to
reduce grammaticalization to other kinds of grammatical change, in particular analogy and
reanalysis, can model this oriented character of grammaticalization.
In “Grammaticalization and linguistic variation”, Shana Poplack reminds us of synchronic
variation as the counterpart of change and, consequently, synchronic layering of forms of
different degrees of grammaticality as the counterpart of diachronic grammaticalization. She
offers variationist methodology as a tool to operationalize and test hypotheses on
grammaticalization. Illustrating from Brazilian Portuguese, she shows how the periphrastic
future with ir ‘go’ first competes with other, partly synonymous means of coding future
reference and gradually ousts these from their specific domains, ending up as virtually the
only future expression in the colloquial register. Comparison with data from French and
Spanish yields the surprising result that the contextual grammatical factors favouring the
spread of the synthetic future differ in the three languages, rendering cross-linguistic
generalizations on the conditions of the grammaticalization of ‘go’ to a future marker
hazardous.
Rena Torres Cacoullos and James A. Walker write on “Collocations in
grammaticalization and variation”. They show that periphrastic constructions and collocations
get condensed over time, and provide the criteria and measures for it. This has relatively little
to do with grammaticalization.
In “Grammaticalization and corpus linguistics”, Christian Mair discusses the importance
of electronic corpora and of frequency studies in research on grammaticalization and, by
querying English corpora, shows that the phrasal conjunction on (the) basis (that) is currently
at an incipient stage of grammaticalization.
“Grammaticalization and language change in the individual” is another contribution from
the Helsinki variationist group, this one by Helena Raumolin-Brunberg and Arja Nurmi. The
analysis of variation in the linguistic biographies of three Early Modern Englishmen
concerning the use of auxiliaries and modals does not show that processes of
grammaticalization may be observed “in individuals’ linguistic practices over their lifetime”
(p. 262) or that the diffusion of grammaticalization processes differs from the diffusion of
other kinds of linguistic change.
In “Grammaticalization in non-standard varieties of English”, Bernd Kortmann and
Agnes Schneider report on 22 changes in the tense/aspect domain and 9 changes concerning
pronouns which are observable in contact-based and traditional varieties of English
worldwide. Some of them are very wide-spread in the Anglophone world, others are
apparently due to particular substratum influence. One conclusion worth quoting is that no
difference in kind between internal and contact-induced grammaticalization phenomena can
be discerned. However (p. 277), “the only clear cases of innovations in the

Lehmann, Review Narrog & Heine (eds.) 2011 6
grammaticalization history of English (or, more generally, Germanic) are those which are
contact-induced.”
This topical area is taken up in the next chapter on “Grammaticalization and language
contact” by Yaron Matras. The author attempts to uncover the mechanisms underlying
‘contact-driven grammaticalization’, conceiving it as a case of language convergence in the
bilingual speaker’s mind.
In “The areal dimension of grammaticalization”, Bernd Heine and Tania Kuteva focus on
definiteness and future tense marking to show how the grammar of one language may be
replicated by another. Sharing the contact-linguistic perspective on grammaticalization with
the previous chapter, their emphasis is more on the development of ‘grammaticalization
areas’.
Béatrice Lamiroy and Walter de Mulder compare “Degrees of grammaticalization across
languages”, putting Spanish, Italian and French on a cline of increasing grammaticalization
according to a large set of criteria. The point here is that a relative degree of
grammaticalization can be assigned to a language as a whole and that languages may differ in
their rate of grammaticalization.
In “Grammaticalization and semantic maps”, Heiko Narrog and Johan van der Auwera
show how the results of grammaticalization research may be visualized by semantic maps.
Part III “Domains of grammaticalization” is devoted to the role played by
grammaticalization in the components and subcomponents of the language system and
discourse. In “Grammaticalization and prosody”, Anne Wichmann deals with the reduction of
prosodic prominence accompanying grammaticalization and causing segmental erosion. The
argument is largely based on “discourse markers”. Now, cliticization has often been seen as a
typical ingredient of increase of bondedness (also in the subsequent chapter, p. 346f).
However, it suffices to look at Romance conjugation (e.g. the Spanish future cantaré ‘I will
sing’ and conditional cantaría ‘I would sing’) in order to see that affixes may be stressed. The
relationship between grammaticalization and prosody is not so straightforward.
In the chapter on “The gradual coalescence into ‘words’ in grammaticalization”, Martin
Haspelmath reviews the various criteria of bondedness and wordhood and concludes that,
since they are partly independent of each other, the transition of a syntactic construction into a
word form is a complex and gradual process rather than an instance of (categorical)
reanalysis.
The complex relationships between “Grammaticalization and word formation” are tackled
by Ilse Wischer. Derivation is seen to resemble grammaticalization, phrasal compounding
instead comes under lexicalization, and hypostasis (independentization of affixes) under
conversion.
Scott DeLancey’s chapter on “Grammaticalization and syntax: a functional view” is based
on an all-encompassing conception (p. 366): “Grammaticalization is not simply a mechanism
by which morphological structure develops; it is the constant, universal tendency of language
out of which all structure arises.” Furthermore, he contends (p. 368) “that what is constant
across languages is recurrent grammaticalization pathways, not universal categories. … Thus
the limits and possibilities of syntax are defined by the range of possible motivated
constructions and the phenomena of grammaticalization …, not by constraints on the possible

Lehmann, Review Narrog & Heine (eds.) 2011 7
outputs of the process”. The major specific point is gradient categoriality as the outcome of
grammaticalization.
“Grammaticalization and word order change” is the topic of the chapter by Chaofen Sun
and Elizabeth C. Traugott. On the basis of English and Chinese examples, they show that a
change in word order may either be a consequence of a grammaticalization process or may
trigger further grammaticalization processes. A subset of the examples merely shows that
under grammaticalization, an item may join a different distribution class and consequently
become subject to the order constraints on that class; that does not actually affect the word
order rules of the language.
The chapter on “Grammaticalization and semantic change” by Regine Eckhardt tries to
give a formal account of the semantic changes in grammaticalization. Illustrating with modal
verbs and particles, with auxiliaries and negation from the history of German and English, she
provides detailed semantic analyses. The formalization, however, does not in general go
beyond the prose descriptions found in the literature and occasionally contributes nothing, as
when the “formal” paraphrase offered for epistemic could again contains could (p. 394).
Basing his chapter on “Pragmatic aspects of grammaticalization” on relevance theory,
Steve Nicolle distinguishes between conceptual and procedural meaning and argues (p. 407)
“that grammaticalization begins with the addition of procedural information to the meaning of
a construction”. What may then be lost, by bleaching, is conceptual meaning. Subjectification,
on the other hand, is taken in Langacker’s sense and therefore not grounded in pragmatics.
Richard Waltereit “Grammaticalization and discourse” deals with the grammaticalization
and pragmaticalization of items that were first (p. 423) “used in discourse with a rhetorical
purpose.”
In “Grammaticalization and conversation”, Elizabeth Couper-Kuhlen shows how bipartite
constructions may be thought to evolve through integration from conversational exchanges.
Thus, left dislocation may be conceived as a ‘collapsed form’ (p. 429), in one speaker’s
utterance, of a ‘recognition search sequence’ involving both interlocutors. While the idea is
intriguing, the abstract process posited here has no direct bearing on any grammaticalization
scale, since patterns of connected discourse thus created precede, so to speak, the left pole of
those scales.
The chapter on “Grammaticalization and lexicalization” by Douglas Lightfoot differs
from most of the others by presenting essentially a research report – perhaps the most
appropriate thing to do in a situation where scholars still mean widely different things by the
terms in question. His suggestion of singling out ‘narrow lexicalization’ from among all the
processes by which something may enter the inventory might well carry over to
grammaticalization; s. below.
In “Grammaticalization and pragmaticalization”, Gabriele Diewald argues that while
most previous studies have subsumed under pragmaticalization a set of processes which work
like grammaticalization except that they do not lead into grammar, pragmaticalization is,
instead, (p. 451) “an integral part of grammaticalization.” This argument presupposes that we
regard deixis as a component of pragmatics (rather than semantics) and, at the same time, the
relational character of grammatical meaning as ultimately grounded in deixis.

Lehmann, Review Narrog & Heine (eds.) 2011 8
In “Iconicity vs. grammaticalization: a case study”, John Haiman deals with
grammaticalization as transition from transparency to compactness, illustrating with the
formation of a ‘why p’ interrogative sentence by condensation of a bi-clausal construction of
the type ‘how does it come about that p’. No generalizations about the role of iconicity in a
pregrammatical stage of language evolution and the increasing arbitrariness brought in by
grammaticalization emerge from this largely descriptive account.
Muriel Norde defines “Degrammaticalization” as gain in autonomy or substance. For a
couple of examples, she shows that they underwent changes which must be described by a
reversal of Lehmann’s (2002[T]) parameters of grammaticalization. A relevant example is
Old Bulgarian nĕčĭto ‘something’ becoming the Modern Bulgarian noun nešto ‘thing’. All in
all, she sides with Börjars & Vincent in regarding degrammaticalization as an exceptional
process.
Part IV of the volume is devoted to “Grammaticalization of form classes and categories”.
Elly van Gelderen, Paolo Ramat, Christa König, Walter de Mulder, Anne Carlier, Björn
Wiemer, Manfred Krug, Laurel J. Brinton, Maj-Britt Mosegaard Hansen, Kees Hengeveld,
Debra Ziegeler, Aleksandra Y. Aikhenvald, Noriko O. Onodera, Zygmunt Frajzyngier, Toshio
Ohori, Guy Deutscher, Anna Giacalone Ramat, Caterina Mauri, Sandra A. Thompson and
Ryoko Suzuki deal with agreement, adverbials, case relators, definite articles, passives,
auxiliaries, complex predicates, negatives, tense, aspect, modality, evidentiality, discourse
markers, reference systems, subordination, quotatives, coordinators and final particles.
While English figures prominently among the object languages of the first four parts of
the handbook, followed by Romance languages, but with all other languages keeping a very
low profile, part V “The different faces of grammaticalization across languages” finally does
some justice to grammaticalization research on other languages. It contains chapters by
Roland Pfau, Markus Steinbach, Bernd Heine, Martin Hilpert, Adam Ledgeway, Mario
Martelotta, Maura Cezário, Björn Wiemer, Lars Johanson, Seongha Rhee, Heiko Narrog,
Toshio Ohori, Hilary Chapell and Alain Peyraube on sign languages, African, Germanic and
Latin-Romance languages, Brazilian Portuguese, Slavic and Turkic languages, Korean,
Japanese and Sinitic languages. There is a single bibliography for the set of chapters,
comprising an impressive 82 pages.
Discussion
As is commonly the case when a concept gains in importance in a scientific field, it is applied
to additional phenomena, with the consequence of semantic bleaching. For instance, Dahl (p.
153) defines grammaticalization as “the processes by which grammatical structures and
grammatical markings arise and evolve, including but not restricted to the development from
lexical to grammatical morphemes.” Several other contributors endorse a concept of
grammaticalization which embodies essentially the making (the “emergence”) of grammar.
However, as Fischer (p. 32) puts it: “The widening of the field involved in the phenomenon
called grammaticalization has led to a weakening of the power of grammaticalization as a
clearly circumscribed process of change.” When Meillet (1912) coined the term
‘grammaticalization’, he not only restricted its application to single meaningful units, but also
opposed it to other kinds of grammatical change, especially to analogical change. The first

Lehmann, Review Narrog & Heine (eds.) 2011 9
restriction may be regarded as a typical symptom of Meillet’s time, and most researchers have
dropped it by now, viewing grammaticalization in the context of the paradigmatic and
syntagmatic relations of the meaningful unit affected. The second restriction, however, is
systematic in nature. The mechanisms of analogy are independent of those of
grammaticalization; in a concrete historical instance, one may have one or the other or both or
neither. This narrow concept of grammaticalization is a valid and necessary one. One feels
tempted to take up Lightfoot’s proposal (s. above) concerning the term ‘lexicalization’ and
dub ‘narrow grammaticalization’ the process that subjects meaningful units more to structural
constraints of the individual language system.
Several authors (e.g. Diewald) have adopted the position that was originally put forward
by critics of grammaticalization research, viz. that grammaticalization is not a unified
phenomenon but rather a fortuitous agglomeration of independent elementary processes
which a theory of linguistic change needs, anyway. If that is the case, then the
methodologically appropriate reaction would seem to be an attempt to straighten the concept
of grammaticalization and delineate it cleanly against neighboring concepts. For instance, it is
true that phonological erosion and semantic bleaching occur outside grammaticalization. That,
however, is well captured by subsuming grammaticalization, together with lexicalization,
under the generic concept of reductive change. Similarly, metaphor is such an independent
operation which may contribute to grammaticalization. It is, however, not constitutive for
grammaticalization, as we shall see below. If we strive for a narrow concept of
grammaticalization, the consequence will be that a couple of phenomena that are dealt with in
this volume – and which probably fall under a broad concept of grammaticalization – cannot
be subsumed under it. The relevant properties of grammaticalization must be structural, since
grammar is structure; it is neither a kind of meaning nor a kind of phonological form. The
definitional criterion of getting increasingly subject to structural constraints of the individual
language system should be applied to all those cases of pragmaticalization and
subjectification, to the formation of modal particles, discourse markers, conversational
phraseologisms etc. The goal must be to elaborate a diversified conceptual system in which
each of these finds its place, rather than throwing them all in the basket of
grammaticalization.
The semantic side of grammaticalization needs to be studied more thoroughly. A more
than sufficient number of competing proposals on the general nature of grammatical meaning
and the processes generating it are on the table. It seems clear that grammatical meaning
cannot be defined in substantive terms. While one can enumerate semantic categories and
features that tend to recur in the grammars of languages, these do not provide a criterion to
tell a grammatical from a lexical meaning. A formal approach to grammatical meaning would
take into account its defeasibility. Consider as a background fact that generally, the
combination of lexical items of opposite meaning, like female stallion, leads to a
contradiction. Now if one of the meanings thus combined is a grammatical meaning, then that
gives way to the other one. For instance, gender in human nouns tends to imply sex; but it
suffices to put a noun in a lexical context which forces the opposite sex, and the feature based
on the gender disappears (meine Hilfskraft ist Vater geworden ‘my assistant has become a
father’). The same goes for tenses: unless there are indications to the contrary, they may carry

Lehmann, Review Narrog & Heine (eds.) 2011 10
a specific temporal meaning. But again, it suffices to use a past tense form in a context that
forces future reference, and the grammatical meaning is annulled. As a last example, consider
the question implication of an interrogative construction; it disappears if the construction is
combined with a lexical item signaling a different illocutionary force, e.g. please. This is a
formal manifestation of what Harder & Boy mean by the “discursive secondariness” of
grammatical meaning.
As for the processes generating grammatical meaning, metaphor and metonymy have
been in favor with grammaticalizationists. Metaphor has long been identified as the chief
mechanism of semantic extension. If active in grammaticalization, it leads to semantic
bleaching and is therefore easily reconcilable with the general notion of grammaticalization as
a reductive process. The role of its systematic counterpart, metonymy, is much less clear.
Bybee (p. 77) simply assumes that “adding certain inferences” is a feature of
grammaticalization, passing in silence over the fact that it runs counter to the “losing features
of meaning” which she also regards as constitutive. Likewise, Bisang (p. 105-108) sees
metaphor and metonymy involved in grammaticalization. We must face the fact that we can
either maintain that grammaticalization is a reductive process, therefore involving
desemanticization, or that semantic components inferred from the context may be
incorporated into the meaning of the grammaticalized item. Some have sought the solution to
the apory in serializing the two processes. Thus, Langacker (p. 83) quotes with approval
Traugott’s proposal that grammaticalization first involves semantic enrichment, then loss.
Similarly, Nicolle as reported above suggests that the first phase of grammaticalization
involves accrual of “procedural” components, while the second phase involves loss of
“conceptual” meaning. Since these are independent and, in an essential respect, opposite
processes, they cannot found grammaticalization as a unified process. This again makes us
look for a primarily structural foundation of grammaticalization, of which semantic and
phonological processes are but by-products.
The relationship of grammaticalization with pragmatics, although a recurrent topic in the
volume, remains far from clear, mainly because the concept of pragmatics is controversial. If
one takes it as narrowly as Diewald (s. above), then pragmaticalization may even be a natural
ingredient of grammaticalization. However, such a concept of pragmatics is not needed.
Consider the understanding emerging in a speech situation, the joint creation of sense by the
interlocutors. Part of it is contributed by the meaning of the signs they use according to the
rules of grammar. In other words, it is contributed by the language system. It is there that
semantics resides. The sense of the utterance, however, is something that belongs to the
speech situation. It involves the combination of the interlocutors’ world knowledge,
experience and assessment of the speech situation with their knowledge of the language
system. Pragmatics is concerned with that portion of it all that does not belong to the language
system and is, therefore, not coded in the utterance, but essentially inferred. If that is so and if
grammaticalization is a process enriching the structure of the language system, it follows that
pragmatics has no direct relationship with grammaticalization. It is quite possible that
grammaticalization involves the coding of “procedural” information; but to the extent it is
coded, it is part of the language system, thus, not of pragmatics.

Lehmann, Review Narrog & Heine (eds.) 2011 11
Several of the contributions to the volume endorse Bybee’s (2003) idea that
grammaticalization leads to greater autonomy of the expressions involved, by which she
means the emancipation of coalescing material from their lexical sources and from the
paradigmatic and syntagmatic relations of the latter. It should, however, be clear that that is
the defining feature of lexicalization. When a complex preposition like Span. a pesar de ‘in
spite of’ (ch. 18) is first formed, it is a syntactic construction consisting of three items of the
inventory. The freezing of the collocation creates, first of all, a new entry in the inventory,
which is precisely lexicalization. Apart from that, the new unit may be grammaticalized in its
construction (s. Lehmann 2002[N]).
Bybee (ch. 6) and Torres Cacoullos & Walker (ch. 18) find the driving force of
grammaticalization in frequency. However, frequency is itself only an effect of some cause. It
is, in fact, a consequence of the increasing subjection of the grammaticalized item under
constraints of the language system. Increased frequency, in turn, leads to automatization. And
automatization is finally the essential cause for the virtual irreversibility of
grammaticalization.
Balance
Despite the comprehensive look of the table of contents, the volume leaves us with a few
desiderata. A chapter on the history of research into grammaticalization is conspicuously
missing from part I (Haspelmath, in his contribution to part III, does devote two pages to it).
As a consequence, recognition of the pioneering work by A.W. Schlegel (1818) reduces to
one reference (p. 720) where it is deemed “simplistic”. In his chapter 57, Heine has the
research history on grammaticalization in African languages start with Givón 1971,
disregarding Brockelmann 1908-13 (esp. 359-362) and Meinhof 1936. Second, given the
particular perspective chosen in the chapter entitled “Grammaticalization and linguistic
typology”, the volume does not sufficiently reflect the fact that both at its early beginnings
with Humboldt and Schlegel and in its renaissance in the 1970s, grammaticalization research
has been intimately linked to linguistic typology, especially to its functional variant. Thirdly,
while sociolinguistics in its various facets is well represented in the volume, psycholinguistics
and neurolinguistics are missing. According to a well-established theory of the functional
correlate for grammaticalization (Givón 1989), it is essentially a kind of automatization. That
notion is certainly amenable to psycho- and neurolinguistic methods.
The volume comes closest to the ideal of a handbook in parts IV and V. It is there that a
certain aspect of the entire field is subdivided according to a simple criterion, the chapters
display a certain parallelism and jointly achieve a rather equable coverage of the field. Apart
from that, the merits of the volume do not lie in its function as a handbook. One should not
expect it to codify what may be considered established knowledge in the field. For one thing,
the research strand is young, unripe and heterogeneous. For another, several of the
contributors do not take on the task of explaining in a systematic, encyclopaedic manner what
is known, and rather aim at proposing new, sometimes controversial theses. Central concepts
of grammaticalization theory such as bleaching, renewal, univerbation, etc., while figuring in
the subject index, are nowhere defined. The subject index is incomplete. For instance, in the
lemma ‘univerbation’, the chapter that most uses the term (ch. 35) is not referenced. Some

Lehmann, Review Narrog & Heine (eds.) 2011 12
contributions simply miss the handbook genre, witness such formulations as “but I will argue
that … I shall show how …” (p. 331). The bibliography contains no less than 29 entries
referring to (as yet) unpublished manuscripts. Cross-references to other chapters have
obviously been inserted by the editors, and equally obviously, authors ignore each other’s
chapters. For instance, Joseph’s, Matras’s and Norde’s chapters propose cases of
degrammaticalization that are refuted in Börjars & Vincent’s chapter, with no cross-
references between them. In many respects, the collection is more like a thematically coherent
conference volume than a handbook. Its merits lie in the achievement of bringing together
specialists from numerous fields of linguistics and with widely different methodological
approaches to focus on a common topic. It offers both interesting data and challenging,
sometimes original theoretical claims. It certainly constitutes a milestone in
grammaticalization research.
References
Beneš, Eduard 1974, "Präpositionswertige Präpositionalfügungen." Engel, Ulrich & Grebe, Paul
(eds.), Sprachsystem und Sprachgebrauch. Festschrift Hugo Moser zum 65. Geburtstag, Teil 1.
Düsseldorf: Schwann (Sprache der Gegenwart, 33); 33-52.
Brockelmann, Carl 1908-13, Grundriß der vergleichenden Grammatik der semitischen Sprachen. 2
Bde. Berlin: Reuther & Reichard (Nachdr. Hildesheim: Olms, 1966).
Bybee, Joan 2003, ‘Mechanisms of change in grammaticization: the role of frequency’. Joseph, Brian
D. & Janda, Richard D. (eds), The handbook of historical linguistics. Oxford: Blackwell; 602–
623.
Givón, Talmy 1989, Mind, code and context. Essays in pragmatics. Hillsdale, NJ: L. Erlbaum.
Heine, Bernd & Kuteva, Tania 2002, World lexicon of grammaticalization. Cambridge: Cambridge
University Press.
Heine, Bernd & Narrog, Heiko (eds.) 2009, The Oxford handbook of linguistic analysis. Oxford etc.:
Oxford University Press (Oxford Handbooks in Linguistics).
Lehmann, Christian 2002, Thoughts on grammaticalization. Second, revised edition. Erfurt: Seminar
für Sprachwissenschaft der Universität (ASSidUE, 9).
Lehmann, Christian 2002, “New reflections on grammaticalization and lexicalization”. Wischer, Ilse
& Diewald, Gabriele (eds.), New reflections on grammaticalization. Amsterdam & Philadelphia:
J. Benjamins (TSL, 49); 1-18.
Lehmann, Christian 2004, "Theory and method in grammaticalization." Zeitschrift für germanistische
Linguistik 32(2 [2005]):152-187.
Meillet, Antoine 1912, "L'évolution des formes grammaticales." Scientia 12(26):6ff. Reimpr.: Meillet,
Antoine 1921, Linguistique historique et linguistique générale. Tome II. Genève. Slatkine &
Paris: Champion; 130-148.
Meinhof, Carl 1936, Die Entstehung flektierender Sprachen. Berlin: D. Reimer.
Schlegel, August Wilhelm von 1818, Observations sur la langue et la littérature provençales. Paris:
Librairie grecque-latine-allemande.