Review of Grammatical Change
|EDITORS: Dianne Jonas, John Whitman, Andrew Garrett
TITLE: Grammatical Change
SUBTITLE: Origins, Nature, Outcomes
PUBLISHER: Oxford University Press
George Walkden, Department of Linguistics & English Language, University of
This volume has its origins in the Eighth Diachronic Generative Syntax (DiGS)
conference, at Yale in 2004. The seventeen papers within showcase a variety of
languages and perspectives.
The first four chapters all deal with the vexed question of the relationship
between one of the founding assumptions of diachronic generative syntax - the
denial of independent diachronic principles - and apparent directionalities of
change in the historical record. Paul Kiparsky’s opener is characteristically
ambitious. The paper seeks to restore a version of the definition of
“grammaticalization” proposed by its coiner, Meillet (1912), in place of the
definitions familiar from modern work on the topic; Kiparsky also defends a
version of unidirectionality. The notions of grammaticalization and analogy are
put to work to explain a wide variety of changes relating to Finno-Ugric case
systems. Kiparsky is sceptical about the explanatory force of reanalysis - a
theme picked up by Andrew Garrett’s chapter, which is a blistering assault on
the perceived reanalocentricity of modern historical syntax. Garrett takes issue
with the textbook account of the emergence of _for_ NP _to_ VP infinitivals,
according to which this pattern arose through reanalysis of the object of the
preposition as an infinitival subject; instead he suggests that the source
construction was the simple NP to VP infinitival pattern, with _for_ being
borrowed from the purposive _for to_ VP construction by analogy.
The following chapter, by Batllori & Roca, aims to characterize the synchronic
and diachronic variation across Ibero-Romance in the alternation between ser and
estar. This is done by means of a careful formal analysis of the syntax and
semantics of the relevant constructions, supported by much data, and a proposed
grammaticalization trajectory consistent with the principles proposed by Roberts
& Roussou (2003) and van Gelderen (2004). David Willis’s chapter on Jespersen’s
Cycle in Welsh rounds off this section. Notable here is the substantial
discussion of stage 2 of the cycle, arguing that the new negator in Welsh, ddim,
passes through a stage of being a negative polarity item before being reanalysed
as a pure SpecNegP element. The chapter also contains valuable discussion of the
lexical approach to syntactic variation in relation to acquisition and change.
Section 2 is devoted to the nominal domain, an area which has not been as
extensively explored in diachronic generative syntax as the clausal/verbal
domain but which is well-represented here. Bergeton & Pancheva address the
development of English reflexives, arguing against existing analyses in favour
of one in which a new phonologically null reflexive emerged to take over from
the personal pronouns. Pronoun + _self_, in the meantime, emerged not as a
reflexive but as an intensifier, and only begins to be grammaticalized as a
reflexive in the Modern English period. Gertjan Postma’s chapter continues the
reflexive theme, with a focus on the innovation of the reflexive zich in eastern
Dutch dialects. His thesis is that internally-driven change created a gap in the
linguistic system which needed to be filled, with dialects of German serving
merely as the source from which zich was co-opted. Quantitative data is drawn
from a new diachronic corpus of 15th-century eastern Dutch.
With the next two papers we move from reflexives to Balkan definite articles.
Dimitrova-Vulchanova & Vulchanov discuss the rise of the article in
Bulgarian.They present a range of diagnostics for article status which they
argue show that in Old Bulgarian the demonstrative had already been partially
grammaticalized as a definite article. Cristina Guardiano’s chapter deals with
changes in the history of articles in Greek, focusing mainly on the apparent
optionally of occurrence of the definite article in Ancient Greek. Her
conclusion is that a null expletive is possible in Ancient Greek, and that this
is linked to the fact that +/-count has not yet been grammaticalized at this
stage. Rounding off the section on nominals is Paola Crisma’s chapter on
genitives in the history of English. Crisma shows that the distribution of
genitives in Old English, though on the surface exhibiting wide variation, can
be reduced to a few core patterns with semantic and structural differences
between them, and explores some of the consequences of the loss of postnominal
The following five papers are on the clausal domain. Eric Haeberli & Susan
Pintzuk’s chapter is a detailed quantitative study of verb clusters in Old
English, showing that this stage of the language conforms to generalizations
made about verb clusters on the basis of modern West Germanic, and that the
relevant structural properties do not change significantly over the Old English
period. The chapter by van Kemenade & Milićev addresses subject positions in
early English; they argue that assuming the adverbs tha and thonne to occupy
fixed positions in subordinate clauses, serving as information-structural
partitions, sheds light on the distribution of subjects. To finish up the
English-fest we have Brady Clarkfs paper, which also addresses the question of
subject positions in the history of English but from a radically different
theoretical perspective. Clark explores the application of Stochastic Optimality
Theory to diachronic syntax, providing an interesting challenge to standard
categorical approaches to variation and change.
Ana Maria Martins examines the percolation of Portuguese inflected infinitives
into ECM environments, suggesting that independent clauses containing an
inflected infinitive could be reanalysed by acquirers as gapped embedded clauses
when conjoined to and following a clause containing an ECM verb. This
configuration also paved the way for embedded negation in non-finite structures.
John Sundquist’s chapter, meanwhile, looks at negative movement in Norwegian, a
relic of a once more widely productive OV pattern. He analyses its retention in
terms of Sobin’s (1997) Virus Theory, as the result of a process operating
outside the core grammar for sociolinguistic reasons.
With the final section we ‘zoom out’ from exclusively European languages to
Uto-Aztecan and Austronesian, to consider issues of morphosyntactic typology.
The thesis of Jason Haugen’s chapter is that a decomposition of Baker’s (2001)
Polysynthesis Parameter is needed. He supports this with evidence from a variety
of languages; however, the main focus of the chapter is on the history of
Nahuatl. Haugen outlines a scenario in which object polysynthesis was
grammaticalized first, followed by subject polysynthesis, with noun
incorporation a separate parameter. Last but not least is Edith Aldridge’s paper
on alignment change in Austronesian languages. She puts forth an analysis in
which absolutive case is valued by T in intransitives and v in transitives, and
outlines a theory of the change from ergative to accusative in which the
decisive step is the reanalysis of the antipassive as transitive.
There is no doubt that this long-awaited volume will be an invaluable resource
for historical syntacticians. Beyond that, however, the first part of the volume
in particular serves as a great introduction to, and manifesto for, the “DiGS
approach” to historical syntax. The introduction lays out three principles - an
emphasis on rigorous synchronic description, an emphasis on reliable and
well-understood data, and scepticism towards independent diachronic processes -
that are fundamental to this approach. Historical linguists dissatisfied with
the reification of diachronic grammars and explanatory handwaving still
occasionally found elsewhere in the literature would do well to turn to this book.
Another great strength of this volume, and indicative of the depth that the
field is now reaching, is the incorporation and formalization of semantic,
pragmatic and information-structural factors in some accounts (e.g. Batllori &
Roca, Bergeton & Pancheva, van Kemenade & Miliæev, Sundquist): sophisticated
modelling of continuity and change in these areas is essential if we want to get
the bigger picture, especially given the now-widespread view that syntactic
change itself is rare and occurs only when it has to (Keenan 1994, Longobardi
2001), or that syntactic change per se may not even exist (Hale 1998: 14). In
general, though, the level of formal analysis across the papers is extremely
variable: while some contain detailed analyses (e.g. Willis, Clark, Aldridge),
others merely hint at what the formalization of their approach would be (e.g.
Garrett, Haugen). The latter approach is not necessarily problematic, assuming
with Chomsky (1990: 145-146) that formalization is a means to an end rather than
an end in itself. However, the extra rigour provided by description within a
well-articulated formal framework is one of the cornerstones of the DiGS
literature, and I hope that this does not decline with future volumes.
The book is well laid out and typeset, and typos and other errors (e.g.
“Lighfoot”, p4; “peeks” for “peaks”, p148; “6 case in where”, p242) are rare. As
regards the content of the volume, I have two main criticisms, relating to the
use of quantitative data and to discussion of diachrony more generally, which I
will discuss in turn.
In the introduction it is stressed, rightly, that diachronic syntacticians have
been instrumental in the development of annotated corpora for historical
linguistic research. This is with good reason: since students of
historically-attested languages have no access to speaker judgements, all
historical syntax is either corpus linguistics or bad corpus linguistics. With
few exceptions, however, the papers in this volume are unforthcoming with regard
to their evidential basis. In particular, few papers present quantitative
information, and where such information is presented the authors often play fast
and loose. Postma, for instance, provides a number of graphs showing percentages
(pp142-144) but does not give the percentages in numeric form or the raw
numbers; Dimitrova-Vulchanova & Vulchanov provide percentages, but without raw
numbers, and use the phrase “statistically significant” (p165) without reference
to any statistical test; Crisma, while presenting more substantial data, refers
to a ratio of “about 1:100” (p203). The only authors to make use of inferential
statistics are Haeberli & Pintzuk and Clark. Of course, the provision of
quantitative data, like formalization, is only a means to an end; the numbers
only tell us what we might want to consider explaining, and do not provide
explanations in and of themselves. Nevertheless, given that another of the DiGS
cornerstones as expressed by the introduction is the use of reliable data, it is
to be hoped that future volumes of this kind will contain higher levels of
quantitative sophistication, as this enables the reader to have greater
confidence in the evidential basis of the claims being made.
The third DiGS cornerstone set out by the introduction is the rejection of
independent principles of change as diachronically causal. While this
constitutes an important methodological advance, it is notable in this volume
that many of the papers do not focus on explaining syntactic change at all (e.g.
Dimitrova-Vulchanova & Vulchanov, Guardiano, Haeberli & Pintzuk, van Kemenade &
Miliæev, Haugen). In those papers in which more than one synchronic stage is
discussed, the link between the two stages is often given very little
spaceGuardiano, for instance, briefly mentions that a generalization under
discussion is “the consequence of a parameter resetting” (p192), and Haugen
discusses a number of stages but only attempts to link them together on p330
with a brief discussion of possible reanalyses. It goes without saying that the
juxtaposition of two accounts of the synchronic syntax of different stages of a
language, however elegant and well-motivated the analyses might be, does not
constitute an explanation for a change, even though such analysis is a
prerequisite for successful discussion of diachrony. The rejection of
independent principles of change is all well and good, but should not lead to
avoidance of addressing the causes of change.
The papers by Kiparsky and Garrett do focus on the mechanisms of change, arguing
that reanalysis occupies too dominant a position in historical syntactic
explanation; see also de Smet (2009). However, like de Smet, both propose to
solve the perceived problem with recourse to analogy, and I have some
reservations about this. Kiparsky’s notion of analogy is very powerful: it may
be proportional or non-proportional (p48), exemplar-based or non-exemplar-based
(p49), and paradigmatic or non-paradigmatic (p41); and it subsumes
degrammaticalization. Whether grammaticalization is also subsumed under the
heading of analogy is rather unclear from the paper: Kiparsky demonstrates,
convincingly, that the interaction of grammaticalization and analogy is complex
(pp23-37), argues that grammaticalization (like degrammaticalization) is a
special case of analogical change (p49), then concludes that analogical change
and grammaticalization are “mechanisms” of change that can interact (p51). He
criticizes reanalysis for being a ‘looser’ theory of change (p50), but by
reducing essentially all grammatical change to analogy it seems to me that
Kiparsky has reinstated analogy as a ‘dustbin category’, of no theoretical
consequence and making no empirical predictions in itself. Garrett, meanwhile,
while attacking the ‘blinkered’ attachment of many historical syntacticians to
reanalysis, offers no particular definition of analogy beyond the traditional
vague intuition (p52). While I accept that reanalysis can be abused in the way
that Kiparsky and Garrett are cautioning against, then, I see no reason to
relegate it to a back seat - particularly in view of the elegant
reanalysis-based accounts of change elsewhere in this volume (Willis, Martins,
In sum, though, this volume shows that approaching historical syntax from a
generative perspective can be fruitful. It shows that proponents of this
approach are not afraid to question old orthodoxies, test out the implications
of new theories, or explore areas of grammar beyond the syntax itself, and it
shows that they are committed, by and large, to a rigorous methodology for
description. All in all, this volume shows that the field is thriving.
Historical syntax is by its nature a backward-looking discipline, but looking to
the future, with a volume such as this in hand, one can’t fail to be excited.
Baker, Mark C. 2001. The atoms of language: the mind’s hidden rules of grammar.
New York: Basic Books.
Chomsky, Noam. 1990. On formalization and formal linguistics. Natural Language &
Linguistic Theory 8, 143-147.
De Smet, Hendrik. 2009. Analysing reanalysis. Lingua 119, 1728-1755.
Gelderen, Elly van. 2004. Grammaticalization as economy. Amsterdam: John Benjamins.
Hale, Mark. 1998. Diachronic syntax. Syntax 1, 1-18.
Keenan, Edward. 1994. Creating anaphors: an historical study of the English
reflexive pronouns. Ms., University of California at Los Angeles.
Longobardi, Giuseppe. 2001. Formal syntax, diachronic Minimalism, and etymology:
the history of French chez. Linguistic Inquiry 32, 275-302.
Meillet, Antoine. 1912. L’évolution des formes grammaticales. Scientia 12, 384-400.
Roberts, Ian, & Anna Roussou. 2003. Syntactic change: a Minimalist approach to
grammaticalization. Cambridge: Cambridge University Press.
Sobin, Nicholas. 1997. Agreement, default rules, and grammatical viruses.
Linguistic Inquiry 28, 318-343.
ABOUT THE REVIEWER
| ABOUT THE REVIEWER:
George Walkden is a Lecturer in the Department of Linguistics & English
Language, University of Manchester. His PhD dissertation, on aspects of
early Germanic clause structure and the methodology of syntactic
reconstruction, is near completion. He is also the editor of the
recently-founded Journal of Historical Syntax (http://historicalsyntax.org).