| EDITORS: Gisella Ferraresi & Maria Goldbach
TITLE: Principles of Syntactic Reconstruction
SERIES: Current Issues in Linguistic Theory 302
PUBLISHER: John Benjamins
Chiara Gianollo, University of Konstanz
This volume collects eight contributions that discuss the possibility of
syntactic reconstruction, that is, tracing back by comparative analysis features
of the syntax of an undocumented ancestor of extant daughter languages.
Syntactic reconstruction raises numerous important methodological issues which
directly impinge on the way language change is described and understood. The
debate concerning the feasibility of this line of inquiry has been revived,
particularly in the past few years, with more diachronic research from formal
syntactic and typological perspectives.
Ferraresi and Goldbach have brought together scholars approaching syntactic
reconstruction from different theoretical perspectives, and drawing data from
various language families. The resulting collection mirrors the challenges and
uncertainties in the ongoing debate. Essays range from extremely skeptical to
moderately optimistic about the possibility of syntactic reconstruction, and
consensus is reached on a few methodological points. The volume, with such
heterogeneity, marks a milestone and will be of interest to historical linguists
and those interested in models of linguistic variation.
The volume opens with a foreword by Giuseppe Longobardi, stressing
epistemological motivations for discussing syntactic reconstruction. He
acknowledges that syntactic reconstruction would help emancipate historical
syntax from its somehow ancillary status with respect to traditional historical
linguistics and theoretical syntax. He describes the conceptual turns triggered
by the development of the biolinguistic framework and the Principles &
Parameters model of variation. Parameters, offering a universal finite set of
discrete entities, are argued to represent suitable entities for comparison and
reconstruction, differently from patterns or constructions, less abstract
notions which may result from different parameter settings in different
languages. If parametric values could be shown to be significantly persistent in
time, they could represent suitable entities not only for comparison, but also
for establishing genealogical kinship. Importantly, according to Longobardi
retrieving phylogenetic relations and reconstructing protolanguages are not
necessarily connected: it is possible to define probabilistically a
statistically sufficient number of similarities which may hint to language
kinship, without having to decide which similarities must be considered features
of the protolanguage.
Like other contributors, Longobardi considers a restrictive theory of possible
changes a prerequisite for syntactic reconstruction. He describes Inertia, first
proposed by Keenan (1994), as 'the potentially most restrictive framework for a
theory of grammatical change' (p. XV). Inertia is compared to the Neogrammarian
principle of 'Ausnahmslosigkeit' (absence of exceptions) in sound change in
being superficially defied by many disturbing factors, which can however be
reconciled with the original, restrictive hypothesis by supplementary
considerations, like borrowing, analogy, and competing 'laws' in phonology.
Ferraresi and Goldbach's 'Syntactic reconstruction: Methods and new insights'
briefly reviews each chapter. It also surveys the main directions in syntactic
reconstruction and tries to set priorities for future research. The editors
argue that historical-comparative linguistics and (especially generative)
diachronic syntax are still often two separate lines of inquiry, the latter not
sharing some classic concerns of the former, such as the identification of
genealogical relations and the reconstruction of protolanguages. Reconciling
diachronic syntax with the Comparative Method is given great weight and is
argued to motivate the discussion in the volume.
Ferraresi and Goldbach present the founding principles of the Comparative Method
and the conditions for its successful application to internal and comparative
reconstruction. They discuss some of its general shortcomings, such as the fact
that it can prove the relatedness but not the non-relatedness of languages, and
the difficulty of transposing the method, which has achieved outstanding results
in reconstructing the phonology and to a lesser extent the morphology of
protolanguages, to syntax. The most significant obstacle in applying the
Comparative Method to syntax is the nature of the entities to be compared
(comparanda), i.e. the difficulty of establishing precise correspondence sets
such as the arbitrary sound-meaning pairings on which the Comparative Method is
based in phonological reconstruction.
A further key difference with respect to the Comparative Method's traditional
domain of inquiry is the fact that syntactic features alone have never
represented a proof of kinship. That is, if syntactic reconstruction is to be
undertaken, the relatedness of languages under observation must be proven
independently, i.e. by applying the Comparative Method to the phonological
Ferraresi and Goldbach stress that any sound proposal for syntactic
reconstruction must be couched within a coherent theoretical model of syntax and
of syntactic change. In this respect, they understand the Principles &
Parameters model as central to understanding linguistic variation and
investigating linguistic change, thanks in particular to the potential of
parameters to subsume numerous structural differences under single, discrete
points of variation. A clear understanding of the correlation among different
superficial properties would allow for reconstruction, e.g. in cases where only
a subset of supposedly connected properties were represented in historical
In 'How much syntactic reconstruction is possible?', Acrisio Pires and Sarah G.
Thomason reach the conclusion that reconstruction in syntax, while worth
pursuing, is bound to yield less successful and reliable results than in
phonology, especially given our current understanding of syntactic change.
Pires and Thomason claim that reconstruction should be explicitly concerned with
the retrieval of parametric values (formal features of lexical items) and the
restitution of models of mental grammars, not with descriptive generalizations
about surface patterns. Nonetheless, they allow for reconstruction of surface
forms in the parent language using surface patterns as comparanda (as e.g. with
basic word order in the Uralic family), with the proviso that more than one
grammar might be compatible with this kind of result.
Constraints on syntactic variation and change are discussed and argued to
substantially overlap, with no specific restrictions applying to syntactic
change (such as the Inertia principle). Contrary to Inertial theories, they also
assume syntactic features to be directly influenced by processes of change. They
discuss the notion of regularity in phonological and syntactic change: the
fundamental difference between the two would be the blindness to all
considerations of meaning and the minimal and easily detectable role of
analogical processes in phonological change.
Pires and Thomason argue for moderate skepticism, showing how different
interpretations are possible for the same scenario (e.g. animacy in Slavic, the
Romance future) depending on certain basically arbitrary decisions, and they
evaluate the usefulness of notions such as markedness and naturalness. They
conclude that, in spite of limitations, some interesting and useful hypotheses
of reconstruction are possible, especially for phenomena where syntactic
variation overtly correlates with morphophonological distinctions.
In 'Reconstruction in syntax. Reconstruction of patterns', Alice Harris offers a
survey of the advantages of the notion of 'pattern', which she has been
defending most notably since Harris and Campbell (1995), in the study of
syntactic change and syntactic reconstruction. A pattern is here defined as 'a
repeated form that is paired with a consistent function or distribution' (p.
86). According to Harris, only patterns can be used in successful syntactic
reconstruction, because patterns are the only entities actually represented in
Patterns can be compared using sentences where all the lexical material can be
proved to be cognate, but also when none of the lexical material is shared.
Thus, for instance, it is possible to compare yes/no questions in a group of
languages already independently known to be related (Kartvelian in Harris'
example). Harris argues that in this case a common construction can be
reconstructed for Proto-Kartvelian, and that, with the help of phonological
comparison, even a specific functional element, the question particle *-a, can
be attributed to the ancestor.
Other examples are case marking patterns (core uses of ergative, absolutive,
dative) in the Lezgian languages, which would allow for both comparative and
internal reconstruction; issues in case distribution (exceptional marking for
the subjects of certain unergative verbs) in Kartvelian, which show the
importance of studying relics -- such as lexically conditioned exceptions -- and
dialectal data; gender-class agreement in Lezgian languages, which, by the
combination of crosslinguistic comparison and the analysis of fossilized
relics, suggests the reconstruction of a morphosyntactic pattern: the languages
in the family which do not productively use gender-class agreement markers on
the verb do show morphological relics of this system (the invariant marker -b-),
which allow for its reconstruction in the protolanguage.
Harris' perspective on the feasibility of syntactic reconstruction is,
therefore, very positive. She attributes the strong skepticism encountered in
the field to two main factors: the fact that most scholars have been working on
Indo-European, encountering special problems due to the time depth of separation
within the family, and the almost exclusive focus on word-order phenomena, which
might be more difficult to reconstruct due to the high number of factors
involved. She concludes by singling out, as particularly favorable environments
for reconstruction, complex constructions where syntax is mirrored by morphology
and phonological correspondence strengthens the comparison.
Ferdinand von Mengden's 'Reconstructing complex structures. A typological
perspective' insightfully discusses the import for syntactic reconstruction of
two lines of research, the study of implicational universals and
grammaticalization theory. Both share the goal of discovering directional
constraints to syntactic change by capitalizing on crosslinguistic regularities.
Von Mengden sets apart genetic classification and reconstruction as two
different tasks. He discusses the limits of the Comparative Method when applied
to syntax, which he attributes to two reasons: the nature of the compared
entities -- complex structures resulting from the recursive combination of
segments, i.e. syntagmatic relations, unlike the paradigmatic relations
investigated in the reconstruction of phonemic inventories --, and their
sensitivity to contact-induced change.
The diachronic application of implicational universals, on the other hand, would
allow one to inferentially establish historical connections between syntactic
features. Von Mengden criticizes past attempts in this direction for focusing
too much on the supposed 'strive towards consistency' in language history,
falsified by the widespread typological inconsistencies witnessed by all
attested languages. He favors an approach that follows Hawkins' (1983)
discussion of syntactic reconstruction, where typological consistency,
reformulated as 'Cross-Categorial Harmony', is not considered sufficient for
predicting change. Hawkins uses implicational universals to deduce diachronic
statements on the typological plausibility of a reconstructed grammatical system
and the relative chronology of innovations. Thus, for example, implicational
word-order universals would predict that in a language characterized by
noun-possessive and noun-adjective orders, like 'Late Common Germanic' (Hawkins
1983: 265), among the co-existing noun-genitive and genitive-noun orders, the
latter must represent an older relic.
Studies on grammaticalization can help reconstruct grammatical systems, by
finding in the attested morphology traces of previous complex analytic
structures. Von Mengden discusses the risk of an extreme interpretation of
grammaticalization theory, according to which every attested morpheme would be
traced back to an analytic construction in the protolanguage, running in so
doing against the commonly assumed Uniformitarian Principle. He also claims that
a wider empirical coverage is needed to reach serious conclusions on
crosslinguistic pathways. He concludes that, although any attempt to
reconstruction will plausibly remain a hypothesis not subject to falsification,
investigations on protolanguages may contribute to a better understanding of
Rosemarie Luehr presents, in 'Competitive Indo-European syntax', her conclusions
on the feasibility of syntactic reconstruction, drawn from a wide comparative
study conducted by combining the classic Comparative Method with the insights of
formal syntax. In Luehr's opinion, syntactic reconstruction cannot be more than
a 'Wahrscheinlichkeitsschaetzung' (Dressler 1971), a probability estimation
which depends on a number of factors. It can help, nonetheless, to better
understand language history, by determining the least common denominator among
functionally equivalent alternatives in the daughter languages.
Luehr compares the different strategies for expressing sentential object clauses
in Hittite, Vedic, Iranian, Greek, Latin, and Germanic (Old Saxon). She starts
from the observation that the frequency of subordinate clauses introduced by a
complementizer parallel to English 'that' in modern Indo-European languages
clashes with the remarkable rarity of such a construction in older Indo-European
varieties. In the languages under analysis, other functionally correspondent
constructions ('that-clause competitors') are employed, arranged according to
their degree of inflection (which I interpret as the number of functional
categories involved): abstract deverbal nouns, infinitival constructions,
structures with participles or predicative adjectives, and quotative
constructions. Of these competitors, the only pattern common to all the
languages investigated, and thus a potential candidate for reconstruction in a
protolanguage, is the construction involving a participle or a predicative
adjective, analyzed as a bare verb clause. Another construction common to all
the languages examined is the explicative clause, i.e. an adjoined attributive
clause introduced by a relative element and linked to a (nominal or pronominal)
reference in the main clause. Luehr traces back to this construction the origin
of the more modern 'that-clause', by means of a reanalysis process triggered in
particular semantic contexts and followed by a pragmatic change responsible for
the dropping of the reference element.
In 'Principles of syntactic reconstruction and morphology as paleosyntax' Irene
Balles shows, by means of a case study from Indo-European, the pros and cons of
the application to syntax of the Comparative Method and grammaticalization theory.
Starting from the observed unidirectionality from analytic to synthetic involved
in grammaticalization, and capitalizing on Givon's (1971) proposal to treat
'today's morphology' as 'yesterday's syntax', she analyzes some secondary
Indo-European synthetic verbal formations (Germanic dental preterite, Latin and
Slavic imperfect) to see whether it is possible to trace them back to an
analytic origin, as constructions formed from a nominal predicative element +
copula, similar to what would be represented, according to her analysis, by the
Old Indic cvi-construction. She concludes her analysis by casting serious doubts
on the possibility of drawing such a parallel and, therefore, performing such a
reconstruction, but she remains convinced of the fact that a substantially
improved theory of change and an enriched empirical database could make this
sort of enterprise more successful.
Claire Bowern's 'Syntactic change and syntactic borrowing in generative grammar'
deals with the role of language contact and syntactic borrowing and discusses
strategies for recognizing borrowed constructions. The author criticizes the
generative approach for being exceedingly 'synchronic', treating diachronic
syntax simply as a branch of comparative syntax and accomplishing a phenetic
rather than phylogenetic comparison. She attributes this lack of attention to
the temporal dimension to the I-language approach, which would entail the
assumption that 'there is no real continuity in grammar' (p. 190). Neglecting
the E (external)-language dimension also leads to disregarding the role of
language contact, which is instead argued to have pervasive effects on syntactic
systems, and which represents an important factor in phonological reconstruction.
Bowern discusses the desiderata of a theory of historical syntax, among which
the understanding of the dynamics of change within communities and a typology of
changes, explaining also differing change rates and relative stability of
systems, are particularly relevant. She then discusses the role of borrowing in
Inertial theories of syntax (borrowing with or without lexical transfer, the
relative order of categorial reanalysis and structural shift), the nature of
syntactic traits involved in borrowing, the conditions for the occurrence of
borrowings in syntax, and the methods to single them out in research. The
criteria are shown to be parallel to those traditionally adopted in phonological
comparison and reconstruction: irregular correspondences (language-internal
exceptions), 'exotic' constructions representing counterexamples to typological
principles, features associated with borrowed lexical items, areal convergence,
Let us now turn to some general issues, which are repeatedly dealt with from
different theoretical perspectives across various essays.
A. I-LANGUAGE AND CHANGE
It has been often remarked -- and criticism on this line is one important point
in Bowern's paper -- that an I-language perspective on language change rendes
questions about linguistic genealogy and long-range relations among languages
devoid of significance. If the only meaningful relation is the one established
between the grammar(s) providing the primary data for acquisition and the new
resulting grammar, then how can we find in this framework a place for notions
such as the development of French from Latin?
Longobardi defends the ability of an I-language approach to capture historical
relations, by capitalizing on the recursiveness of derivation, which renders the
vague concept of E-language dispensable: besides the relation of immediate
derivation (diachronic contiguity) holding between two I-languages, also
relations between non-immediately contiguous I-languages are historically
significant, when interpreted as recursive diachronic contiguity. The shift to
I-language in generative historical syntax (Lightfoot 1979) had two main
motivations: one was to comply with the insights of generative theories of
syntax, which stressed the vagueness and the consequent scientific
inapplicability of the traditional notion of 'language' (E-language), and the
necessity to focus on a better definable object, the individual linguistic
competence; the other, tightly connected reason was the criticism of long-range,
teleological theories of language change, particularly influential in studies on
syntactic reconstruction in the 1970s. Working on I-languages from a historical
perspective requires a greater degree of idealization with respect to synchronic
work. In fact, I can compare the language of Cicero and the language of
Augustine for historical purposes without the actual possibility of
demonstrating the chain of direct derivations taking from one I-grammar to the
other. Any description and explanation of syntactic change using I-language
requires additional evaluation of the wider sociolinguistic dimension.
Nonetheless, it seems that the methodological advantages of focusing on a
clearly defined object and task (describing the structure and transmission of
mental grammars) are worthwhile.
B. COMPARANDA IN SYNTAX
The Comparative Method requires comparanda to share precise correspondences in
form and meaning. How are these correspondences to be sought in syntactic data?
Contributions discuss two approaches: one relies on parameters, and the other on
patterns. The main criticism addressed to patterns is that they are shallow and
often misleading, since e.g. corresponding linear orders in different languages
might well result from completely different grammars (in this sense, the
parallel Harris, p. 86, draws between her notion of pattern and Jackendoff's
1994 use of the term seems misplaced). Thus, in von Mengden's phrasing (p. 104),
the use of patterns for syntactic reconstruction 'entails the risk of comparing
analogues, whereas the Comparative Method requires the comparison of homologues'.
On the other hand, controversy arises also on the status of parameters: here,
the possible advantage of abstractness in explaining co-variation of surface
constructions is undermined by the paucity of such parameters thoroughly and
convincingly argued for in the literature and, more in general, by the still
insufficient amount of empirical evidence to substantiate the parametric
research program (cf. Newmeyer 2005).
The parametric approach's validity will have to be further defended by empirical
studies and research on the format of parametric variation (e.g. Baker 2001,
Roberts and Roussou 2003, Gianollo, Guardiano and Longobardi 2008), including
the investigation of possible schemata and the extent of implicational
embedding, whose importance for diachronic investigation is particularly
stressed by von Mengden. Clearly, despite such shortcomings, the application of
parametric models to diachronic research is very promising for the field of
syntactic reconstruction, for it potentially combines a principled and
fine-grained analysis of syntactic phenomena with an ambition to typological
C. KINSHIP AND RECONSTRUCTION
The classic Comparative Method, to date the only successful method for
(phonological and morphological) reconstruction, is based on two fundamental
characteristics (cf. Rankin 2003): the safety of comparanda and the regularity
of sound change. With respect to the first point, the possibility of comparison
itself, and of genealogical grouping -- eventually and optionally followed by
reconstruction --, is guaranteed by the arbitrariness of the sound-meaning
pairing, by virtue of which exact correspondences in form and meaning represent
agreements beyond chance.
Since the same kind of arbitrariness is not found in syntactic data, the status
of syntactic features as meaningful comparanda for genealogical purposes is a
matter for discussion. (These papers disagree with respect to an optimistic
attempt by Gianollo, Guardiano, and Longobardi 2004 -- now superseded by
Longobardi and Guardiano 2009.) Several contributions point out the difficulty
in distinguishing shared inheritance from areal diffusion or from typological
correspondence with no genealogical import in syntactic data. But the general
view represented by those positive about the feasibility of syntactic
reconstruction is that the task is approachable even if syntactic data cannot
disclose genealogical relations. That is, syntactic reconstruction would be
possible on languages whose kinship is proved by means of other methods
(basically, the Comparative Method applied to phonology). The following question
arises: if -- as commonly assumed -- success in genealogical grouping does not
necessarily entail reconstruction of a protolanguage, can we say, on the other
hand, that reconstruction can be performed on data which are assumed not to
carry a signal of genealogical relatedness? The answer proposed by several
papers is only indirect: syntactic reconstruction is deemed to be more feasible
when there is a close correspondence between syntactic features and their
morphological exponents, i.e. elements which can be captured with the
Comparative Method. This can also be connected to Hale's (1998: 15f.)
observation that much of the difference in progress between our understanding of
phonological and syntactic change has to do with the fact that functional heads
are frequently phonologically null.
D. THEORIES OF SYNTACTIC CHANGE
Every paper in the collection points out that a theory of syntactic change is
necessary to frame an analysis of reconstruction. In fact, regular sound change
is the foundation of the Comparative Method (together with the safety of
comparanda, as discussed). In the same way, the discovery of constraints on
possible syntactic changes is the prerequisite for any successful attempt to
Ferraresi and Goldbach and von Mengden discuss implicational universals, which
have been controversially interpreted by Lehmann (1974) and Vennemann (1974) as
directing the evolution of languages towards typological consistency, but which
can also be seen, as in Hawkins (1983), as a way to assess the
logical/chronological order of changes. Another domain considered promising in
most of the papers is grammaticalization, interpreted as a unidirectional and
grammar-simplifying process in cases of ambiguity of analysis, as in Roberts and
Pires and Thomason forcefully argue that no constraints should be posited for
syntactic change apart from those imposed on variation by the language faculty.
If it is true that no constraint or hypothesis on change should be formulated
that does not conform with the attested typological range of variation, it is
also clear that we have to look for independent restrictions on the way grammars
are transmitted, in order to explain why we do observe that change (even in
syntax) is not random, and the amount of syntactic change witnessed in the
history of languages is incomparable, as far as we can judge, to the range of a
priori permitted variation. The notion of naturalness, which Pires and Thomason
discuss as a tentative hypothesis to be independently motivated, could be better
understood by further pursuing, within a restrictive framework, research on
markedness, cross-categorial harmony, drift, and least effort principles (see
Roberts 2007 for an insightful perspective on these topics). A general research
program addressing constraints on syntactic change has been proposed by Inertial
theories, which I shortly comment below.
E. INERTIAL MODELS OF SYNTAX AND LANGUAGE CONTACT
Ideally, according to the original formulation of Inertia in Keenan (1994), no
change should result apart from change caused by decay (essentially,
phonological erosion) or by outside forces, i.e. interference (intrinsic to the
heterogeneity of primary corpora). Now, it is immediately clear, in view of the
actual historical data, that this formulation is only valid as a petition of
principle, to set the direction for a restrictive theory of change. Even if
absolute Inertia cannot be maintained, as Longobardi acknowledges here and in
his (2001) paper, still the core insight of the hypothesis is that primitive
changes from the inside are to be kept to a minimum in diachronic explanation.
Thus, Longobardi (2001) relativizes Inertia to narrow syntax, and assumes an
entirely deterministic mechanism of acquisition, missing convergence with the
source I-language only in case of interference or of changes in other modules,
i.e. at the interfaces with phonology and semantics.
Inertial hypotheses do not necessarily entail, as in the critical interpretation
of Pires and Thomason (p. 47), that 'syntactic change (i.e., innovation)
resulting from interference should in principle have significantly different
structural properties from syntactic change that does not result from
interference'. The dichotomy between change due to interference and change due
to language-internal interface pressures only concerns the primitive cause of
change, the 'actuation' problem; change is implicitly assumed to be parametric
in both cases (since this is the only format of variation assumed by the
theory). It is certainly matter for future research to investigate whether
change by interference can in fact be shown to display peculiar characteristics
with respect to the nature of the categories involved or its transmission within
communities. What is sure is that, to date, we have no comprehensive evaluation
of the role of language contact on syntax. Proper syntactic 'borrowing', in this
respect, seems to offer quite a limited perspective: indirect influence on
frequency of native constructions through language contact may be even more
Importantly, scholars embracing some form of the Inertial hypothesis do not
exclude language contact from their theory of change, as some criticism by
Bowern (p. 202) may lead one to think; on the contrary, interference is
considered the most obvious primitive cause for change. What Inertia says, in
fact, is that there is only one other source of syntactic change, namely change
at the interfaces. It is true that research inspired by the Inertial hypothesis
shifts its focus from the study of interference to the study of change at the
interface, in order to provide empirical support to the main claim that narrow
syntax is in itself inert. The result is that language contact may figure as a
marginal topic in some generative diachronic literature. However, there is a
body of generative research couched in a variationist framework (see Kroch 2001
and Pintzuk 2003 for surveys) that, starting from similar basic assumptions
about (absence of) endogenous change, has been successful in explaining
syntactic change by focusing on sociolinguistic dynamics within heterogeneous
The volume comes out almost ten years after the workshop on syntactic
reconstruction (held in Konstanz in 1999 on the occasion of the annual meeting
of the Deutsche Gesellschaft fuer Sprachwissenschaft), which gave rise to a
lively debate, hosted by the Journal of Linguistics, between David Lightfoot
(Lightfoot 2002a and b) and Alice Harris and Lyle Campbell (Campbell and Harris
2002). In the meanwhile, the explicit confrontation and dialogue between the
various mainstreams in diachronic research has improved, also thanks to
initiatives like the presented one coordinated by Ferraresi and Goldbach.
Hopefully, future research will substantially widen the empirical base for the
discussion of theoretical and methodological stances that are still irreconcilable.
Baker, M. 2001. 'The Atoms of Language'. New York: Basic Books.
Campbell, L. & A. C. Harris. 2002. Syntactic reconstruction and demythologizing
'Myths and the prehistory of grammars'. 'Journal of Linguistics' 38.3, 599-618.
Dressler, W. 1971. Ueber die Rekonstruktion der Indogermanischen Syntax.
'Zeitschrift fuer vergleichende Sprachforschung' 83, 1-25.
Gianollo, C., C. Guardiano, & G. Longobardi. 2004. Historical Implications of a
Formal Theory of Syntactic Variation. Paper presented at DIGS VIII, Yale University.
Gianollo, C., C. Guardiano & G. Longobardi. 2008. Three fundamental issues in
parametric linguistics. In T. Biberauer (ed.), 'The Limits of Syntactic
Variation'. Amsterdam: Benjamins, 109-142.
Givón, T. 1971. Historical syntax and synchronic morphology: An archaeologist's
field trip. 'Proceedings of the Chicago Linguistic Society' 7, 394-415.
Hale, M. 1998. Diachronic syntax. 'Syntax' 1, 1-18.
Harris, A. C. & L. Campbell. 1995. 'Historical syntax in crosslinguistic
perspective'. Cambridge: Cambridge University Press.
Hawkins, J. 1983. 'Word order universals'. London: Academic Press.
Jackendoff, R. 1994. 'Patterns in the mind'. New York: Basic Books.
Keenan, E. 1994. Creating anaphors. An historical study of the English reflexive
pronouns. Los Angeles: UCLA. Ms.
Kroch, A. 2001. Syntactic change. In M. Baltin & C. Collins (eds), 'The Handbook
of contemporary syntactic theory'. Oxford: Blackwell, 699-729.
Lehmann, W.P. 1974. 'Proto-Indo-European Syntax'. Austin: University of Texas Press.
Lightfoot, D. 1979. 'Principles of Diachronic Syntax'. Cambridge: Cambridge
Lightfoot, D. 2002a. Myths and the prehistory of grammars. 'Journal of
Linguistics' 38.1, 113-136.
Lightfoot, D. 2002b. More myths. 'Journal of Linguistics' 38.3, 619-626.
Longobardi, G. 2001. Formal syntax, diachronic Minimalism, and etymology: the
history of French 'chez'. 'Linguistic Inquiry' 32.2, 275-302.
Longobardi, G. & C. Guardiano. 2009. Evidence for Syntax as a Signal of
Historical Relatedness. 'Lingua', doi:10.1016/j.
Newmeyer, F. 2005. 'Possible and probable languages. A generative perspective on
linguistic typology'. Oxford: Oxford University Press.
Pintzuk, S. 2003. Variationist approaches to syntactic change. In Joseph, B. &
R. Janda (eds), 'The Handbook of historical linguistics'. Oxford: Blackwell,
Rankin, R. 2003. The Comparative Method. In Joseph, B. & R. Janda (eds), 'The
Handbook of historical linguistics'. Oxford: Blackwell, 183-212.
Roberts, I. 2007. 'Diachronic Syntax'. Oxford: Oxford University Press.
Roberts, I. & A. Roussou. 2003. 'Syntactic Change: a Minimalist Approach to
Grammaticalization'. Cambridge: Cambridge University Press.
Vennemann, T. 1974. Topics, subjects and word order. From SXV to SVX via TVX. In
Anderson J. & C. Jones (eds.), 'Historical Linguistics. Proceedings of the First
International Congress of Historical Linguistics'. Vol II. Amsterdam:
ABOUT THE REVIEWER