Boskovic, Zeljko (2001) On the Nature of the Syntax-Phonology Interface:
Cliticization and Related Phenomena. North-Holland/Elsevier,
North-Holland Linguistic Series: Linguistic Variations Volume 60,
ix + 328pp, hardback ISBN 0-8-043935-7, $91.00/NLG185.00/EUR83.95.
Kleanthes K. Grohmann,
Johann Wolfgang Goethe-Universit�t Frankfurt am Main &
Zentrum f�r allgemeine Sprachwissenschaft, Typologie und
Universalienforschung Berlin (Germany)
The syntax-phonology interface has been subject to much discussion,
especially in the minimalist program (e.g., Chomsky 1995). In many forms of
this research program, the theory of a locally evaluated, economy-driven
computational system plays a major role, and so does the aim to posit as few
additional conditions as possible (beyond "Bare Output Conditions," i.e.
those that apply to the LF- and PF-interfaces). The relevance to the
syntax-phonology interface is that it has been argued that movement does not
only take place in the syntactic (pre-Spell Out) and the interpretational
components (post-Spell Out, on LF), but also in the phonological component
(PF). Overt and covert movement are arguably essential assumptions of any
generative theory. PF-movement proper, on the other hand, would be an
unwelcome conceptual result, as one would have to posit parallel interfaces
of the syntax, interpretation and phonology, something that is hard to
integrate into the traditional T-model. Even a ("standard") minimalist
version of the T-model does not make available a bi-directional connection
between the syntax and the phonology, but assumes that syntax feeds
phonology (which, in turn, has no effect on either syntax or LF).
A major empirical domain of potential PF-effects (possibly on the syntax,
certainly on the computation) is the placement of clitics, especially in the
(South) Slavic languages. Boskovic sets out to investigate the properties of
South Slavic (clausal) clitic placement and by doing so reconsiders the
empirical arguments for PF-operations. He also proposes a detailed
theoretical account to capture the relevant facts, arguing for a mixed
syntax-phonology approach to the facts at hand focussing around a
modification of which links of non-trivial chains may be deleted and which
ones may be pronounced. Boskovic then shows that this approach is not an ad
hoc mechanism, but one of wider use, which he then applies to a number of
The languages mainly considered are Serbo-Croatian, Slovenian, Bulgarian and
Macedonian. In addition, Boskovic draws from Polish data to compare how his
theory of clitic placement works outside the South Slavic languages. He also
relates the theoretical proposal to at first glance unrelated phenomena,
namely multiple Wh-fronting (Slavic and Romanian), Verb Second (Germanic),
Object Shift and Stylistic Fronting (Scandinavian), and Negation (Romance).
The book consists of three main chapters (with transparent titles and
headings in the table of contents), couched between an "Introduction"
(chapter 1, pp. 1-6) and a "Conclusion" (chapter 5, 283-284). It also
contains a meticulously generated index of 20 pages, which is always
welcome. These three chapters are "Serbo-Croatian Second Position
Cliticization: Syntax and/or Phonology?" (chapter 2, 7-96), "More on Second
Position Clitics: Pronunciation of Non-Trivial Chains" (chapter 3, 97-178),
and "Bulgarian and Macedonian Clitics" (chapter 4, 179-282).
Boskovic briefly introduces what is at stake for the study of the
syntax-phonology interface in chapter 1, and why he ventures out to employ
South Slavic clausal clitics (especially pronominal, auxiliary and
complementizer clitics) to investigate the nature of the syntax-phonology
Chapter 2 lays out in considerable detail the issues relating to
cliticization in Serbo-Croatian, concentrating on the "second position
effect," i.e. the fact that clitics obligatorily appear in the second
position of the clause. This second position is generally literal, unlike
Verb Second: the latter is a requirement that the finite verb appears after
the first full phrase of a clause, while second position clitics typically
follow the first word of a clause -- this can be a complementizer, but also
the first word of a complex (e.g., noun) phrase. Moreover, all clitics
cluster; locating any clitic in non-second position or breaking up this
clitic cluster is not possible.
The main body of this chapter contains an elaborate discussion in which
Boskovic contrasts four approaches to the phenomenon: (i) the "strong syntax
approach," (ii) the "strong phonology approach," (iii) the "weak syntax
approach," and (iv) the "weak phonology approach." After going over
advantages and disadvantages of each of these approaches, Boskovic concludes
that the second position effect has to lie in the phonology, though there is
no need to postulate movement proper on the phonological side. The
descriptive generalization that surfaces is that "[Serbo-Croatian] clitics
must be located in the second position of the I-phrase in which the syntax
places them, which indicates that the second position effect is phonological
in nature (I-phrases are phonological units) but that clitics undergo
movement in the syntax" (p. 94).
Modifying proposals by Klavans (1985) and Radanovic-Kocic (1988), Boskovic
proposes a mixed syntax-phonology account: after the syntactic placement of
clitic(s) and host(s), phonological requirements on clitics (basically, a
filtering effect of the phonology on the syntax) induce the second position
effect. He takes this to be one of the ways in which PF affects word order
without application of the operation Move, a way he elaborates in the
subsequent part of the book.
In this chapter, Boskovic also shows that contrary to what has been argued
in the literature, (i) Serbo-Croatian clitics do not all sit on the same
head (which is the prevailing, if not sole, assumption) and (ii) second
position cliticization in Serbo-Croatian does not support non-standard
claims about the nature of the syntax-phonology interface (often crucially
based on exactly the phenomena under discussion). In particular, there is no
empirical support for the possibility of movement in the phonology, there is
no need for a "look-ahead implementation" from the syntax to the phonology
(in a derivational model in which syntax feeds phonology; see "Overview"
above), and there is no evidence for the necessity of a co-presence,
bi-directional model in which the phonology can feed information to the
syntax (as put forward in Zec & Inkelas 1990).
The empirical content of this chapter is very strong. I believe Boskovic
succeeds in presenting a quite complex state of affairs in manageable ways.
At no point is the reader overwhelmed with data and discussion, but data and
discussion are plentiful throughout this chapter (and the entire book, in
fact). What helps is that the relevant facts of the issue are presented
piecemeal, couched in the critical evaluation of the four approaches
(i)-(iv) mentioned above, spread over two main sections, the syntax- and the
phonology-based approaches. At the end of the chapter, the reader and the
author are on the same page, set up perfectly for the theoretical discussion
In chapter 3 Boskovic presents the theoretical mechanism that allows him to
capture phonological effects on word order without assuming the operation
Move to apply. Assuming the Copy Theory of movement throughout (see Chomsky
1993, 1995), Boskovic offers a way of determining which copy of a
non-trivial chain (created through overt, syntactic movement) will be left
active at the PF interface. In essence, PF may allow pronunciation of lower
copies of non-trivial chains, rather than categorically force pronunciation
of the highest copy. The proposal heavily builds on, modifies and extends
recent work by a large number of scholars, but it concentrates on the
specifics of Franks 1998 (see also Hiramatsu 1997, Pesetsky 1997 and many
Boskovic then shows how this mechanism offers an explanation for a number of
(otherwise) puzzling properties of clitic placement in Serbo-Croatian, such
as the contrast in the placement of the third person singular past tense
auxiliary and other auxiliary clitics within the clitic cluster.
He also goes beyond Serbo-Croatian in some detail and discusses clitic
placement in Slovenian and Polish. The upshot is that all differences in
clitic placement among the languages in question (i.e. South Slavic
languages as well as Polish) are the result of a few simple, independently
motivated differences in the phonological properties of clitics in these
languages. Boskovic argues that the syntax of clitics (and elements relevant
to clitic placement) is the same in all these languages.
The chapter also contains a first extension of the account of the second
position clitic effect (from chapter 2) to other, (supposedly) unrelated
phenomena, such as the Verb Second effect in Germanic languages, which is
argued to be phonological in nature, on a par with the clitic second effect.
He also touches on Object Shift in Scandinavian (revisited in Appendix A of
chapter 4) and multiple Wh-fronting in Slavic languages (as well as the role
of clitic placement in such questions). The discussion of multiple
Wh-fronting also considers Romanian, a non-Slavic language (but part of the
Balkan 'Sprachbund') that employs obligatory fronting of all Wh-phrases.
Once the rich descriptive/empirical work from chapter 2 (including an
evaluation of popular approaches) has been absorbed, one is ready for
theory. Holding fast to a unidirectional connection between syntax and
phonology, it becomes apparent that something else needs to be done than
postulating PF-movement. The Copy Theory offers a straightforward way, which
Boskovic exploits gracefully (granted, he can build on a rich body of
relevant approaches in the literature). To set the reader up for the
specifics on (second position) clitic effects, Boskovic first presents Copy
Theory and then applies the idea that (for non-trivial chains) it must be
the case links other than the highest may sometimes be kept for PF-purposes.
The discussion of multiple Wh-fronting in Slavic languages suggests this
strongly. Drawing from data from the acquisition of English double-auxiliary
constructions and Scandinavian Object Shift facilitate the proposal and make
it look plausible, if not even natural. Boskovic thus gets away easily with
a rather basic implementation of the proposal to the clitic phenomenon in
Serbo-Croatian. After chapter 2, it just makes sense. I find it a welcome
strategy to look for further consequences of this approach instead, which he
does. One major plus of the idea that lower copies may be sent to PF is
that, as Boskovic puts it ""[s]ome optional movements become obligatory"
(the heading of section 3.3.1. on p. 132). (Recall that "optional" movement
has been a thorn in the minimalist eye from Day One; naturally, this
approach does not generalize over all of such "optional" phenomena, such as
the classic case of Scrambling, but the discussion of Object Shift heads in
the right direction).
On the other hand, the discussion of Verb Second is rather sketchy and
centers around on particular case from a North Norwegian dialect. It
basically takes for granted that Verb Second is phonological in nature,
without discussing alternatives or elaborating on the story. However, given
the title of the book, this may be forgiven. (This is a more general point
of slight criticism: while it is welcome to consider "related phenomena"
[part of the sub-title of the book] beyond cliticization, Boskovic seems to
stretch the notion of "relation" at times in which he is not always as
successful as he is when he discusses cliticization proper.)
After the mostly Serbo-Croatian dominated discussion (plus the treatment of
Polish and Slovenian clitics in chapter 3), chapter 4 examines cliticization
in Bulgarian and Macedonian, which has given rise to some of the strongest
arguments for PF-movement in the literature. A major part of the chapter
concerns the complementizer clitic 'li', which can cliticize to elements
that are immobile in the syntax, a fact that has been used as an argument
that PF movement can provide a host for 'li' (see, for example, Rudin et al.
1999). Boskovic shows how the mechanism of pronunciation of lower copies can
be employed to account for all the relevant facts concerning cliticization
in Bulgarian and Macedonian without appealing to PF-movement.
He also shows that the order of clitics within the clitic cluster in these
two languages can be dealt with without assuming rightward head-adjunction
(banned by Kayne 1994, for example), as is standardly done in the
literature. This discussion leads Boskovic to specific conclusions
concerning the structural representation of clitics, which, he argues, hold
cross-linguistically. (It also allows him to delve into the phenomenon of
Stylistic Fronting in Icelandic.)
The chapter ends with two appendices. Appendix A examines the contexts in
which Macedonian clitics function as second position clitics and makes a
proposal how to capture the interaction between verbal procliticization and
second position encliticization in this language. Appendix B gives several
arguments for Multiple Spell Out (an idea originally due to Uriagereka 1999
in the minimalist framework, but in a more recent approaches developed also
by Chomsky 2000 and subsequent work for a phase-based system). These are
based on cliticization in Bulgarian (see also Franks & Boskovic 2001),
Object Shift in Scandinavian languages (lending support to Holmberg�s 2000
attempt to integrate Participle(P)-fronting into "Holmberg�s
Generalization"), and Negation in Romance languages.
The discussion of Bulgarian and Macedonian clitics serves as a nice
compendium to the rich literature (see, for example, papers in collections
by Halpern & Zwicky 1996, van Riemsdijk 1999 or Beukema & den Dikken 2000).
The extensions to other, non-clitic-related phenomena, is, again, rather
thin, but can be excused by being part of an appendix. What surprised me is
the strong tie between Multiple Spell Out and "phases" (of Chomsky's recent
work) that Boskovic entertains towards the end. The surprising part is that
Multiple Spell Out proper was meant to be a natural enrichment of "classic"
minimalism (or so I understand Uriagereka's work). (For pre-minimalist
notions of phonology's multiple access to the syntax, see Bresnan 1971.) The
'phase'-model seems like a much more radical departure from this classic
framework. Moreover, at no other place in the book does Boskovic discuss
phases, and it is not clear that he would want to. The idea of multiple
access of the interfaces to the derivation has been picked up by various
authors since Uriagereka's original 1996 manuscript, some of whom are
mentioned in Appendix B (another one is my own dissertation; see Grohmann
Chapter 5 concludes this study.
Beukema, F. & M. den Dikken, eds.. 2000. Clitic Phenomena in European
Languages. Amsterdam: John Benjamins.
Bresnan, J. 1971. Contraction and the Transformational Cycle. Manuscript,
Massachusetts Institute of Technology, Cambridge.
Chomsky, N. 1993. A Minimalist Program for Linguistic Theory. In K. Hale &
S.J. Keyser, eds. The View from Building 20. Essays in Linguistics in Honor
of Sylvain Bromberger. Cambridge, MA: MIT Press, 1-52.
Chomsky, N. 1995. The Minimalist Program. Cambridge, MA: MIT Press.
Chomsky, N. 2000. Minimalist Inquiries. In R. Martin, D. Michaels & J.
Uriagereka, eds. Step by Step. Essays on Minimalism in Honor of Howard
Lasnik. Cambridge, MA: MIT Press, 89-155.
Franks, S. 1998. Clitics in Slavic. Paper presented at the Comparative
Slavic Morphosyntax Workshop, Indiana University, Bloomington, June 1998.
[Downloadable at http://www.indiana.edu/~slavconf/linguistics/index.html.]
Franks, S. & Z. Boskovic. 2001. An Argument for Multiple Spell-Out.
Linguistic Inquiry 32, 174-183.
Grohmann, K. K. 2000. Prolific Peripheries: A Radical View from the Left.
Doctoral dissertation, University of Maryland, College Park. [Downloadable
Halpern, A.L. & A.M. Zwicky, eds. 1996. Approaching Second: Second Position
Clitics and Related Phenomena. Stanford, CA: CSLI.
Hiramatsu, K. 1997. A Production/Judgement Asymmetry in Children's negative
Questions. Manuscript, University of Connecticut, Storrs.
Holmberg, A. 2000. Scandinavian Stylistic Fronting. Linguistic Inquiry 31,
Kayne, R. S. 1994. The Antisymmetry of Syntax. Cambridge, MA: MIT Press.
Klavans, J. 1985. The Independence of Syntax and Phonology in Cliticization.
Language 61, 95-120.
Pesetsky, D. 1997. Some Optimality Principles of Sentence Pronunciation. In
P. Barbosa, D. Fox, P. Hagstrom, M. McGinnis & D. Pesetsky, eds. Is the Best
Good Enough? Cambridge, MA: MIT Press and MITWPL, 337-383.
Radanovic-Kacic, V. 1988. The Grammar of Serbo-Croatian Clitics: A
Synchronic and Diachronic Perspective. Doctoral dissertation, University of
van Riemsdijk, H., ed. 1999. Clitics in the Languages of Europe. Berlin:
Mouton de Gruyter.
Rudin, C., C. Kramer, L. Billings & M. Baerman. 1999. Macedonian and
Bulgarian 'li' Questions: Beyond Syntax. Natural Language and Linguistic
Theory 17, 541-586.
Uriagereka, J. 1999. Multiple Spell Out. In S.D. Epstein & N. Hornstein,
eds. Working Minimalism. Cambridge, MA: MIT Press, 251-282.
Zec, D. & S. Inkelas. 1990. Prosodically Constrained Syntax. In D. Zec & S.
Inkelas, eds. The Phonology-Syntax Connection. Chicago: University of
Chicago Press, 365-378.
After my doctoral dissertation (Grohmann 2000), I spent some time as a
researcher at the Zentrum f�r allgemeine Sprachwissenschaft, Typologie
und Universalienforschung (ZAS) in Berlin. I am currently the syntax- and
typology-postdoc at the newly formed Graduiertenkolleg"Satzarten:
Variation und Interpretation" at the Johann Wolfgang Goethe-Universit�t
Frankfurt am Main, a mini-graduate school focussing on sentence types.
Among my research interests are the syntax of clitics, Wh-question
formation, left-peripheral phenomena, clause structure and interface issues
(especially syntax-discourse). I am also working on a revision of my
dissertation to be published by John Benjamins under the title "Prolific
Domains: On the Anti-Locality of Movement Dependencies," a minimalist
textbook with Norbert Hornstein and Jairo Nunes (with the working title
"Understanding Minimalism. An Introduction to Minimalist Syntax," contracted
for Cambridge University Press), and a volume on multiple Wh-fronting
strategies, co-edited with Cedric Boeckx (under contract with John Benjamins
with the working title "Multiple Wh-Fronting"). You can find out more about
me on my rarely updated homepage at http://www.punksinscience.org/kleanthes.
I'm also a co-founder of Punks in Science, dedicated to all punks within the
scientific community. Unfortunately, our web-site, which you can find at
http://www.punksinscience.org, is updated even more rarely. As we're always
looking for new friends to join, be creative and provide content-related
input and administrative help, you can make yourselves available by
contacting me at firstname.lastname@example.org or the "other guy," Jeffrey
Parrott, a third-year graduate student in the Department of Linguistics at
Georgetown University (email@example.com).
I want to publicly curse the person(s) who removed my laptop from my office
without notifying me. Apart from a painful loss, this incident also led to
the late appearance of this review.