headmovement

download headmovement

of 36

Transcript of headmovement

  • 8/12/2019 headmovement

    1/36

    Head Movement and the Minimalist Program

    Ian Roberts

    Downing College, University of Cambridge

    (igr20!am"a!"#$%

    0. Introduction

    In this paper I will begin by recapitulating the essentials of the analysis of head-

    movement as it was largely agreed on in mainstream syntactic theory by the late

    1980s. This approach was in essence unaltered in the earlier versions of minimalism

    (homs!y (199"# 199$# aside from section %.10&&. In 'ection # I consider the reasons

    which lead homs!y (001) "*-8& to suggest e+cluding head-movement from the core

    operations of the narrow synta+. 'ection " reviews the various alternatives to the

    earlier conception of narrow-syntactic head-movement which have been put forward)

    , movement# remnant phrasal-category movement and repro/ective movement#

    focussing on a case study of each alternative. inally# in 'ection %# I will consider the

    conceptual status of head-movement in relation to the general goals of the minimalist

    program.

    1. The GB approach1

    arlier versions of generative grammar often featured head-movement operations2 see

    for e+ample 3ffi+ 4opping in homs!y (19$* 5cawley6s (19*1& Tense-attraction

    rule# monds6 (19*1# 19*7& have/be-raising and his (19*8& verb-movement rule for

    rench# den esten6s (198"& analysis of ermanic verb second# rench sub/ect-clitic

    inversion and nglish sub/ect-au+iliary inversion. ut it was only in the period

    that these ideas were systematised and a series of theoretical postulates were put

    forward that together provided a clear characterisation of head-movement# arising

    primarily from the wor! of :oopman (198% Travis (198%& and a!er (198$# 1988&.

    The central idea in these approaches is (1&)

    (1& 4ead movement is the case of 5ove-; where ; is oberts (001&.

    1

    mailto:[email protected]:[email protected]
  • 8/12/2019 headmovement

    2/36

    In most versions of D is an unlicensed free variable or that there is Fimproper

    movement6 with tultimately bound in the domain of the head of its chain ( ibid&.

    iven the assumptions in homs!y (1987 this would violate ,rinciple of the

    binding theory. If the head moves on to an 3-position (i.e. a position ad/oined to

    another head then# we have improper movement. If it does not# the trace of head-

    movement counts as an unlicensed free variable. This proposal does not# however#

    rule out head-to-head ad/unction# and was not intended to. In fact# the upshot of

    homs!y6s reasoning is that head-movement can only move a head to another head

    position. It was generally assumed that head-movement ad/oined the moved head to

    the host head# forming a structure li!e (&)

    (& CG < G D

  • 8/12/2019 headmovement

    3/36

    (ut see >iAAi H >oberts (1989& for a more elaborate proposal&. :ayne (1991# 199%&

    proposed that head-ad/unction is always left-ad/unction# as depicted in (&.

    oncerning locality# the central condition on head movement was the 4ead

    5ovement onstraint (45 first e+plicitly formulated in Travis (198%&. I give it in

    the following form)

    ("& 4ead movement of < to G cannot s!ip an intervening head .

    (>oberts (000)11"&&

    Intervention is understood in terms of asymmetric c-command in the usual way (

    intervenes between G and < iff G asymmetrically c-commands both < and # while

    asymmetrically c-commands

  • 8/12/2019 headmovement

    4/36

    from sub/ects and ad/uncts was impossible. a!er (1988 in particular# showed in

    detail that the various forms of incorporation he proposed satisfy this condition.

    'econd# downward head-movement is not allowed# since a fundamental

    re?uirement imposed by the , is that an antecedent c-command its trace. This

    implies that the 3ffi+ 4opping# as conceived in homs!y (19$*& and elsewhere#

    could not be an instance of head-movement if this is seen as a core syntactic

    operation. (,olloc! (1989& and homs!y (1991 among others# sought to avoid this

    conse?uence by treating the , as holding of @ representations and allowing

    downward movement in the overt synta+ as long as the effects of this were obliterated

    by the time the , applied&. Third# to the e+tent that# following >iAAi (1990 the

    , featured some form of relativised minimality constraint# the 45 itself can be

    derived from the ,. 4ence the local nature of head-movement follows.

    The conception of head-movement# then# was that this was a core syntactic

    operation raising a head < to an immediately super/acent (governing& head G where omance !ind# to of the ermanic !ind and to clause-initial position of the

    !ind found in N'K languages# nglish sub/ect-au+iliary inversion# rench sub/ect-

    clitic and comple+ inversion# Italian 3u+-to-omp# inversion of inflected infinitives

    in uropean ,ortuguese# a whole range of phenomena involving movement of the

    Boun within L, (including 'emitic construct states# al!an and 'candinavian

    postposed articles and the relative ordering of Bouns in relation to possessors and

    modifiers of various !inds2 see in?ue (199% @ongobardi (199%& clitic-movement#

    and many other phenomena (see >oberts (001& for overview# illustration and further

    references&.

    2. The minimalit pro!ram

    In the early versions of the minimalist program# the conception of head-

    movement was by and large retained. The discussion of N-movement to T and 3gr

    and related issues in homs!y (199")*-"M199$)19$-199& introduces chec!ingtheory# and ma!es it clear that N-movement# li!e other forms of movement# obeys the

    %

  • 8/12/2019 headmovement

    5/36

    core constraints that this theory imposes. The same is true for the notions of chec!ing

    domain# internal domain and complement domain2 moreover# head-movement plays a

    role in giving rise to e?uidistant positions# which is central to capturing the

    generalisation that ob/ects move only when the verb moves (an early version of

    4olmberg6s (1987& generalisation&2 see the discussion in homs!y (199")10-

    19M199$)1*7-187&. In homs!y (199$) %.10 the picture changes somewhat# partly as

    a conse?uence of the abandonment of 3gr as a syntactic category. 4ere# homs!y

    proposes analysing multiple sub/ect constructions (e.g. ermanic transitive

    e+pletive constructions such as There painted a tudent the houe or passives li!e

    There have ome ca"e been ba"ed #or the part$ &in terms of multiple specifiers of T#

    since 'pec3gr, is not available if 3gr is not a functional head. homs!y argues for

    an analysis which features the substring%&pletive 'ub(ect T )If N is in T# this is

    clearly the wrong order# the attested order being %&pletive * 'ub(ect. Ta!ing this

    order to be a direct refle+ of the verb-second property of the languages in ?uestion#

    homs!y suggests (199$)"78& that the N-second property .. may belong to the

    phonological component. If that is the case# the observed Ci.e. N E I>D order is

    formed by phonological operations O and may observe the usual constraints (N --P

    but need not# as far as we !now. 3lthough N-to-T movement is assumed

    (homs!y (199$)"7*& the possibility that N orders are derived by something other

    than syntactic head-movement of T to is at least ?uestioned here.

    ut it was in homs!y (001)"*-8& that a series of arguments of a range of

    types are presented that# together# lead homs!y to conclude that a substantial core

    of head-raising processes# e+cluding incorporation in the sense of a!er (1988 may

    fall within the phonological component ("*&.

    irst# homs!y claims that head-movement never affects interpretation) the

    semantic effects of head-raising in the core inflectional system are slight or non-

    e+istent# as contrasted with

  • 8/12/2019 headmovement

    6/36

    -- in the two classes of languages. 'uch effects are not found# leading to the

    suggestion that head-movement is confined to the , part of the grammar.

    'econd# homs!y raises the ?uestion of the nature of the trigger for head-

    movement. The issue arises when we consider# for e+ample# T in a language such as

    rench# which has consistent L,-movement into 'pecT, and consistent N-movement

    to T (following ,olloc! (1989&&. 4ence T must contain the relevant triggers for these

    movements) (uninterpretableMunvalued& Q-features and an ,, feature to trigger L,-

    movement and# presumably# some form of N-features combined with a movement-

    triggering feature triggering N-movement. 3ll other things being e?ual# the system

    has to have sufficiently rich featural information to be able to correctly distinguish the

    two sets of triggers) an

  • 8/12/2019 headmovement

    7/36

    operations. 4owever# as we have see# head-movement was thought to derive

    structures such as that in ( by ad/oining one head to another. 'uch an operation

    does not involve e+tension of the root# at least in any obvious way without appeal to a

    special notion of root (which is imaginable but has not been proposed2 the assumed

    notion of root is that node < such that there is no node G that irrefle+ively dominates

  • 8/12/2019 headmovement

    8/36

    ad/unction of < to G# forming ( further movement of G to a higher head J forms

    the comple+ head [W [Y X Y ] W ]. In other words# iterated head-movement always

    forms a successively more comple+ head. 'uccessive-cyclic head-movement# on the

    other hand# would involve e+corporation of < from CG < G D# moving < on to form [W

    X W ]. 3s already pointed out# >oberts (1991& observed that nothing prevented this in

    the conception of head-movement# and that it was in fact empirically desirable.

    The general view# however# has remained that this possibility is not found (the

    empirical cases >oberts adduced can be analysed in other ways&. If so# then an

    e+planation is re?uired. homs!y says that if head-movement were seen as a

    morphological operation# then this might be why we do not observe e+corporation

    (iterability is a general property of operations of narrow synta+# and these alone (p.

    "8&&. ut if we treat head-movement as syntactic movement# then we have to e+plain

    why successive-cyclic movement# so clearly available for phrasal movement (both 3

    and 36-movement& is not available to head-movement.

    homs!y6s arguments have given rise to various reactions# as we shall see. In

    general they have been influential# in that many researchers have been led to loo! for

    alternatives to the earlier approach to head-movement# either by eliminating it

    altogether# eliminating it from the core computational system of narrow synta+ or

    radically redefining it. In many cases# new phenomena have been brought to bear on

    the issues# or at least older data has been reconsidered in a new light. Two points

    should# however# be made here. irst# although homs!y6s arguments naturally lead

    to a re-evaluation# at least# of the account of head-movement s!etched in 'ection 1# he

    does not articulate a theoretical principle which would force# either directly or as a

    deductive conse?uence# the elimination of head-movement from narrow synta+. The

    ?uestion that remains open if we accept homs!y6s conclusions is then) why is head-

    movement not part of narrow synta+R The second point is related) to what e+tent do

    these ?uestions bear on the conceptual goals of the minimalist programmeR To put the

    ?uestion# in a sense# the other way around (albeit tendentiously&) could this discussion

    regarding the nature of head-movement have been -internalR I will return to these

    points in 'ection %.

    +. ,lternative to core-$ntactic head-movement

    8

  • 8/12/2019 headmovement

    9/36

    Three main alternatives to the earlier form of syntactic head-movement have been

    proposed since homs!y (001 one of which developed to some e+tent

    independently of homs!y6s remar!s. These are the ,-movement approach# which

    homs!y himself advocated# the remnant-movement approach# which partly stems

    from :ayne (199% and the repro/ective approach. I will loo! at each of these turn#

    focussing on one case study of how an earlier analysis or family of analyses involving

    core-syntactic head-movement is replaced by the alternative mechanism.

    +.1 -movement

    To /udge from homs!y6s (199$)%.102 001)"*-8& comments# the alternative he has in

    mind to syntactic head-movement is a ,Eoperation. This becomes clear when we

    consider that the ,-movement alternative is unproblematic in relation to all the

    arguments homs!y ma!es) clearly we do not e+pect ,-movement to have to obey

    the +tension ondition or the c-command condition (nglish 3ffi+-4opping could

    be a case of , head-movement but cannot be syntactic movement (see 'ection 1&&2

    we e+pect it to be triggered ?uite separately from syntactic

  • 8/12/2019 headmovement

    10/36

    't/epanoviS (001 who propose that pseudogapping# as in ($ provides evidence for

    , verb-movement in nglish)

    ($& 3lthough ohn doesn6t eat piAAa# he does E pasta.

    'tarting from @asni! (199$ e+amples of this !ind have been ta!en as evidence for

    syntactic ob/ect-shift in nglish# combined with remnant N,-deletion after ob/ect shift

    (i.e. deletion of [* eat pata ]in ($&&.7oec!+ H 't/epanoviS (001& observed that

    @asni!6s original account of why the verb moves when there is no pseudogapping (to

    deriveohn eat pata& in the second con/unct in ($& by the combination of ob/ect

    shift and N-movement which relies on , feature strength re?uiring either N-

    movement or N-deletion by ,# cannot be maintained in the 3gree-based theory of

    movement of homs!y (000# 001&. 3ssuming all three operations (ob/ect shift#

    verb-movement and ellipsis& to be intrinsically unordered# the ?uestion then becomes

    why N-movement followed by ellipsis is not possible# giving the ungrammatical (7&)

    (7& U O he eats CN, (eats& pasta D. .

    oec!+ H 't/epanoviS argue that the ?uestion concerns ordering# and point out that

    ob/ect shift must precede both head-movement and ellipsis# while the latter two can

    appear in either order)

    (*& a. Kb/ect 'hift P ellipsis (head-movement bled& pseudogapping)

    ## he does pasta CN,eats (pasta& D.

    b. Kb/ect shift P head-movement P ellipsis)

    .. he eats pasta CN,(eats& (pasta& D.

    c. U4ead-movement P ellipsis (ob/ect shift bled&)

    U .. he eats CN,(eats& pasta D.

    They conclude that the right result can be guaranteed if ob/ect shift is a syntactic

    operation# with both ellipsis and N-movement ta!en to be ,-processes. 4ence

    ellipsis can either precede or follow , N-movement2 in the former case# as in (*a

    7 TracesMcopies of moved elements are in round brac!ets.

    10

  • 8/12/2019 headmovement

    11/36

    pseudogapping results# in the latter# NK order results# as in (*b&. (*c& is impossible

    since ob/ect shift# as a syntactic operation must precede verb-movement.

    4owever# altin (00) 7$$& observes that the same movementMdeletion

    options apply to non-verbal predicates such as#ond in (8& and to phrasal categories as

    in (9&)

    (8& 3lthough he isn6t fond of piAAa# he is (fond& of pasta.

    (9& a. 3lthough he isn6t very fond of piAAa# he is (very fond& of pasta.

    b. 3lthough he didn6t try to persuade 5ary# he did (Mtried to persuade&

    5artha.

    In (9a& the gapped string is ver$ #ond# presumably an 3,# and (9b& it is tr$ to

    peruade. altin further observes that it seems that the o#-,, has undergone ob/ect

    shift in (8& and (9a raising ?uestions about @asni!6s initial conclusion. The

    following e+amples underscore both points)

    (10& a. 3lthough ohn isn6t easier to please than 5ary# he is E than ill.

    b. 3lthough ohn isn6t easier to convince the students to tal! to than

    5ary# he is E than ill.

    4ere# than Billmust have undergone putative ob/ect shift# which is surprising since

    this category is usually ta!en to be either a ,, an elliptical , and the pseudogapped

    constituent is the comple+ 3,# containing a possibly unbounded 36-dependency.

    In fact# it appears that the ob/ect shift operation should really be seen as an

    optional focussing operation# moving an

  • 8/12/2019 headmovement

    12/36

    d. U O he did her angry --.

    e. U ... he did angry --.

    @et us suppose# then# that nglish has an

  • 8/12/2019 headmovement

    13/36

    'ection . 4owever# , head-movement must be entirely without @ effects# and a

    number of arguments showing that some cases of head-movement have @ effects

    have been given# notably by @echner (00$& (see also in?ue (1999) 18%# n. 8

    >oberts (forthcoming# hapter Kne wart (001&& have adduced cases where

    apparent head-movement has @ effects. >oberts (forthcoming# hapter Kne& points

    to the following paradigm (see also 5clos!ey (1997)89 :ayne6s (000)%%&&)

    (1"& a. UJhich one of them does anybody li!eR

    b. Jhich one of them doesn6t anybody li!eR

    c. UThey succeeded in finding out which one of them anybody li!ed.

    d. UThey succeeded in finding out which one of them anybody didn6t

    li!e.

    e. They succeeded in finding out which one of them wasn6t li!ed by

    anybody.

    4ere it appears that the B,I an$bod$in sub/ect position in (1"b& is licensed by the

    au+iliary raised to . This argument depends on the standard assumption that B,Is

    must be c-commanded by their licensers at @. 5ovement of the au+iliary in

    e+amples li!e (1"b& above affects @ by altering c-command relations involving the

    moved item# and as such is head-movement analogue of raising in (1%&)

    (1%& a. 3fter the meeting# nobody seemed to anybody to be satisfied with

    the outcome.

    b. U3fter the meeting# it seemed to anybody that nobody was satisfied

    with the outcome.

    urthermore# 5atushans!y (007)10-%& provides a plausible reason for why it should

    be the case that verb-movement# in particular# often lac!s semantic effects) essentially

    this is because verbs are predicates. To ?uote 5atushans!y whether we assume that

    predicates must reconstruct .. or allow them to be interpreted in their final position#

    the outcome is the same) predicate movement is not reflected at @ (10"&.

    There may well be reasons# then# to thin! that not all head-movement ta!es

    place at ,. This does not imply that no head-movement ta!es place at ,# of course#although unambiguous evidence to this effect is lac!ing (and if the suggestion in Bote

    1"

  • 8/12/2019 headmovement

    14/36

    $ that Internal 5erge cannot ta!e place at , is correct then it may be that , head-

    movement is impossible# after all&.

    +.2 3emnant phraal movement

    To some degree as a direct response to homs!y6s (001& arguments# summarised in

    'ection # and to some e+tent as a conse?uence of the re-evaluation of the status of

    clitic pronouns following on from :ayne (199% a number of authors have proposed

    remnant-movement accounts for some of the phenomena previously handled as head-

    movement# including verb-movement of various !inds (see :oopman H 'Aabolcsi

    (000 Bilsen (00" 5Vller (00% Ji!lund H entAen (00* Ji!lund#

    4rafnb/argarson# entAen H 4rWarsdWttir (00* entAen (00*# to appear# and

    several of the contributions in 5aha/an (00"&&2 see also the recent treatments of

    various forms of inversion in rench in :ayne H ,olloc! (001 ,oletto H ,olloc!

    (00% ,olloc!# ,oletto H 5unaro (00" ,olloc! (007 and several of the papers

    on verb-initial languages in arnie# 4arley H Looley (00$ and# on the synta+ of

    nominals# 'hlons!y (00% in?ue (00$# forthcoming& and the references given

    there&.

    These approaches share the central idea that analyses positing head-movement

    relations of the type schematised in (1$& should be replaced by analyses of the general

    type in (17&)

    (1$& O 4 O C

  • 8/12/2019 headmovement

    15/36

    presumably present in core synta+ but deleted in ,&. 5ovement of iemsdi/! (199%

    ThrXinsson (000&&. This is followed by N,-fronting to the first position in the clause#

    usually thought to be 'pec,# satisfying the N constraint here. This combination of

    operations is entirely licit# and e+plains what would otherwise be an anomalous N

    construction# involving /ust a participle preceding the inflected verb. (18 from

    5Vller (1998 illustrates the interaction of remnant-movement with reeAing and the'trict ycle)

    1$

  • 8/12/2019 headmovement

    16/36

    (18& UJorVber hat CL, ein uch (worVber& D !einer (L,& gelesenR

    Jhat-about has a boo! noone read

    Jhat did noone read a boo! aboutR

    (18& can be derived by moving the L, ein Buch 7or8ber first# and then by

    sube+tracting 7or8ber2 this violates reeAing. 3lternatively# if 7or8ber is first

    moved# and then L,# movement of the L, violates the +tension ondition. Je see#

    then# that remnant-movement and the constraints relevant to it are motivated.

    5Vller (00%& argues that the analysis of verb second (N& constructions

    which postulates two separate movements# one head-movement of the verb and the

    other

  • 8/12/2019 headmovement

    17/36

    the boo! has ritA read

    ritA has read the boo!.

    3s 5Vller points out)

    In this approach# the pre-NM position is occupied by whatever category happens to beat the left edge of v, earlier in the derivation E this will typically be the sub/ect B, oran adverb# but# after scrambling# it may also be an ob/ect B,# a ,,# a ,# or a N,(complete or remnant .. &. (pp. 18-"&

    In addition# there is no reason to postulate head-movement2 in e+amples of the !ind in

    (0 hat is assumed to have merged directly in v. Jhere a main verb appears in

    second position# it has not moved to v# but rather counts as being on the edge of v,owing to the first clause of the definition of dge Lomain# which runs as follows

    (5Vller6s (7 p. 18%&)

    (1& %d!e 9omain:

    3 category ; is in the edge domain of a head < iff (a& or (b& holds)

    a. ; is the highest overt head refle+ively c-commanded by

  • 8/12/2019 headmovement

    18/36

    (& Noi lo vedete.

    Gou(pl& himMit see

    Gou see himMit.

    :ayne adopts three postulates. irst# that morphologically derived forms such as

    vedete are syntactically formed# possibly by syntactically combining the root ved-

    with the theme vowel4e-and the ending Ete. 'econd# that the @3 applies to sub-

    word-level operations# and# third# that the @3 bans multiple head-ad/unction. iven

    these three postulates# the clitic would have to ad/oin to the verb root ved-# followed

    by ad/unction of [ lo ved-] to (the functional head occupied by& 4e- and then

    ad/unction of [[ lo ved-] 4e-] to4te. Jhere the verb bears a prefi+# as in lo prevedete

    (you foresee it the clitic would have to attach to the prefi+.

    :ayne goes on to suggest that a more plausible option is to assume that clitics

    ad/oin to empty functional heads. :ayne further observes enclisis to infinitives and

    imperatives of the type in ("&)

    ("& a. ais-le. (rench&

    Lo it

    b. ,arlargli sarebbe un errore. (Italian&

    To-spea!.to-him would-be a mista!e

    'ince it is very li!ely that the verb moves to in imperatives li!e ("a& (see among

    others >ivero (199%a#b& and that the infinitive is in a high position in ("b&

    (elletti (1990 :ayne (1991& :ayne concludes that in general verb-movement to

    does not carry along clitics. It then follows that# in a rench e+ample li!e (%

    involving sub/ect-clitic inversion with an ob/ect proclitic on the inverted au+iliary#

    the cliticYau+iliary combination has not moved to )

    (%& @6as-tu faitR

    It.have-you doneR

    4ave you done itR

    :ayne follows 'portiche6s (1999& proposal that there may be N-movement to at @#

    hence accounting for the root nature of the construction (he suggests that the cliticmay delete at @2 see his Bote 17&. inally# :ayne observes high-register e+amples#

    18

  • 8/12/2019 headmovement

    19/36

    which show the order litic-3dverb-Infinitive (e.g. .. le bien #aire .. it well to-do&

    support the idea that the clitic and the verb do not have to combine.

    ,oletto H ,olloc! (00% ,olloc!# ,oletto H 5unaro (00"& and ,olloc!

    (007& endorse :ayne6s general conclusion that clitics and verbs cannot and do not

    combine in synta+# but propose that# instead of covert verb-movement into the -

    system in e+amples li!e (% there is overt remnant movement. The derivation of (%&

    would proceed as follows)

    ($& a. Tu Cac!ows!i H Travis (000 and

    many of the papers in arnie# Looley H 4arley (00$&&. In her study of NK' and

    N'K in Biuean# for e+ample# 5assam (000& argues that there is an operation

    fronting a verbal constituent# and that this constituent is fronted to a position within

    T,. 'he then shows that there is a general operation which fronts non-verbal

    predicates which are clearly larger than heads# e.g. relative clauses. Third# 5assam

    shows that what has been called noun-incorporation in Biuean (e.g. by a!er (1988&&

    cannot be movement of B into N (pacea!er& since there are clear cases where a

    8 It is also worth pointing out that although ,oletto H ,olloc! (00%& and ,olloc! (007& sharewith 5Vller (00%& the basic idea that N-type verb movement (full in the latter case# residual in theformer should be reanalysed as remnant movement# they do not suggest that remnant movementshould be a global replacement for head-movement# as they continue to assume both N-to-T movement

    of the type argued for in ,olloc! (1989& and a head-movement analysis of (sub/ect& cliticisation# hencetu in ($c& moves to the head whose specifier the fronted remnant

  • 8/12/2019 headmovement

    20/36

    constituent larger than B undergoes this operation. 'he proposes instead that putative

    noun-incorporation is really the absence of ob/ect-shift to a N,-e+ternal position. In

    that case# the fact that the apparently incorporated noun moves with the verb shows

    that what is moved is N, rather than N. NK' order is thus derived by N,-fronting#

    and N'K by ob/ect-shift to a N,-e+ternal position combined with remnant N,-

    fronting# as shown in (7&)9

    (7& a. CT,CN,N K D T Cv, ' v .. (N,& DD -- NK'

    b. CT,CN,N (K& D T Cv, ' v C3bs, K (N,& DDD -- N'K

    3s (7& shows# the landing-site of N,-fronting is ta!en to be 'pecT,2 5assam argues

    that this is motivated by essentially the same property as that which causes the sub/ect

    to raise to 'pecT, in languages li!e nglish# rench and 5ainland 'candinavian) the

    operations can be seen as two reflections of a single ,, predication feature

    (5assam (000)111&&. This type of analysis# first put forward by 5assam H

    'mallwood (199* and developed by >ac!ows!i H Travis (000& as well as several

    of the papers in arnie# Looley H 4arley (00$ has been applied to a number of

    languages which display both NK' and N'K orders (mainly but not e+clusively

    5acronesian and 5ayan languages2 unli!e rigidly N'K languages such as the eltic

    languages# where it is at the very least much harder to motivate a remnant N,-fronting

    analysis&. 4ere too# though# the ?uestion of the trigger for some of the movements

    arises (see hung (00$&&.

    @et us now briefly evaluate remnant-movement approaches against

    homs!y6s (001& arguments# given in 'ection . irst# it is clear that remnant-

    movement avoids the problems head-movement causes for the +tension ondition

    and the definition of c-command. 'ince it is

  • 8/12/2019 headmovement

    21/36

    is entirely unclear why this particular type of

  • 8/12/2019 headmovement

    22/36

    head-movement (pace homs!y then# remnant-movement approaches may have

    difficulties with c-command effects.

    In conclusion# reanalysing head-movement as remnant-movement avoids a

    number of the problems homs!y pointed out for head-movement# although if

    anything the trigger problem may be e+acerbated. >econsidering head-movement in

    this light has been productive in the cases of B-to-L movement# creation of verb-

    clusters of the Jest ermanicM4ungarian type (see :oopman H 'Aabolsci (000& and

    the papers in :iss and van >iemsdi/! (00%& and has led to new ideas in the case of

    N and certain cases of inversion in >omance. This approach is almost certainly# then#

    an alternative in some cases# but it remains unclear to what e+tent it represents a

    global alternative to head-movement# and# given the trigger problem# it is not clear

    that it is conceptually simpler.

    +.+ =3epro(ective> movement

    'till another strand which has been pursued as part of the general reconsideration of

    the nature of head-movement is represented by a class of analyses which we can

    collectively label repro/ective. This approach has been developed primarily by

    ury (00"# 00* Lonati (007 :oeneman (000& and 'urXnyi (00$# 00*# 008&.

    The basic idea is to ta!e head-movement to be syntactic movement# but to treat it as

    arising from a different set of conditions from

  • 8/12/2019 headmovement

    23/36

    cannot in principle rule out head-movement to a specifier position (this point is also

    made by 5atushans!y (007 >oberts (00$ Toyoshima (000 Nicente (007&&.

    3gain similar to what we have seen here# she suggests that the 4ead 5ovement

    onstraint is irrelevant to the ?uestion of the e+istence of head and phrasal movement#

    in that locality constraints act on the search operation# not on movement itself# hence a

    single set of locality constraints should govern both types of movement. 'he further

    observes that there are empirical doubts about the 45# citing long verb-

    movement in reton (see orsley# >ivero H 'tephens (1997&& and the similar cases

    in 'outh 'lavic and archaic >omance discussed by \avar H Jilder (199% @ema H

    >ivero (1990# 1991 >ivero (1991# 199"a#b# 199%a#b# 199* >ivero H TerAi (199$&&.

    Instead# she adopts the chain uniformity condition of homs!y (199$)$"&& in (8&

    and the minimality condition on 5erge in (9&)

    (8& 3 chain is uniform with regard to phrase structure status (where phrase

    structure status means the (relational& property of ma+imal# minimal or

    neither&.

    (9& 5erge /ust enough material for convergence.

    (9& applies to both internal and e+ternal merge. inally# she assumes that a head#

    when merged either e+ternally or internally# pro/ects2

  • 8/12/2019 headmovement

    24/36

    In ("0a we have the derived structure of a wh-interrogative. Lonati argues that the

    wh-feature cannot move as a head in this case as it would turn the interrogative clause

    into a L,. Thus# re?uirements of @ convergence (the structure must be interpretable

    as an interrogative clause& cause the non-minimal L,-movement option to be ta!en.

    4owever# in a conte+t compatible with L,-selection and showing no phrasal pied-

    piping (Lonati (007)"& the option in ("0b& should be available. This# Lonati

    argues# is what we find in free relatives and comparatives.

    or free relatives# Lonati6s evidence comes from paradigms li!e the following)

    ("1& a. UI will visit C what town D you will visit.

    b. I wonder C what town D you will visit.

    c. I will visit C what D you will visit.

    ("1b& clearly contains an indirect ?uestion# i.e. a ,-complement to 7onder with

    7hat to7nin its 'pecifier. Kn the other hand# viitdoes not ta!e an indirect-?uestion

    complement# or indeed any !ind of ,# but only a L,. The complement in ("1c& is

    thus a L,# a free relative. ,ied-piping to the edge of a free relative is impossible# as

    ("1a& shows (following :ayne (199% Lonati assumes that 7hatever-type relatives E

    as inI 7ill viit 7hatever to7n $ou 7ill viitE are not in fact free relatives&. ("1c&

    thus involves repro/ective movement of a CYwhDL# giving rise to a derived structure

    li!e ("1b&.

    Lonati goes on to argue the same for comparatives# !nown to involve wh-movement

    since homs!y (19**&. The idea that comparatives are comple+ nominals is supported

    by the fact that they e+press a description of a degree# and by the fact that this

    e+pression can enter into scope ambiguities of the type first discussed in >ussell

    (190$&)

    ("& I thought your yacht was bigger than it is.

    If the comparative e+pression (than it i& is outside the scope of thin"# we have a non-

    contradictory reading for ("&2 if it is inside the scope of thin"we have contradictoryreading (cf. I thought your yacht was bigger than itself&. Third# comparatives are

    %

  • 8/12/2019 headmovement

    25/36

    strong islands for e+traction# suggesting they are comple+ L,s (here eat is elided

    while 7hat and&-?uic"l$are wh-moved&)

    (""& UJhat do you eat the soup more ?uic!ly than ,aul does (eat& (what&

    (+-?uic!ly&R

    inally# Lonati gives evidence that in >omanian and ulgarian the same wh-element

    moves overtly as a head in comparatives but as a phrase in interrogatives. In fact# the

    same can be shown with non-standard varieties of nglish which allow 7hatto appear

    in comparative subdeletion and to act as an adnominal wh-determiner)10

    ("%& a. 5ary ate more coo!ies that what she ate C (what& candies D.

    b. U 5ary ate more coo!ies that what candies she ate (what candies &.

    c. Jhat candies did she eat (what candies& R

    d. UJhat did she eat C(what& candiesDR

    Lonati (007)"9& concludes there is no principled reason for wh-movement to be

    restricted to phrases.

    Knce again# let us consider these proposals in relation to homs!y6s (001&

    arguments. >epro/ective movement does not target heads# and so the +tension

    ondition and c-command problems homs!y raises do not apply in this case. The

    triggering problem appears to be dealt with by Lonati by @) if the movement is not

    repro/ective# one !ind of structure and interpretation must result2 if it is# then a

    different one results. The synta+ itself allows either option. In a sense# then# Lonati

    has @ act as a filter on the syntactic derivation. >egarding onward movement#

    presumably the possibilities here are determined by the repro/ection option. 'tic!ing

    to Lonati6s e+ample with CYwhDL# we can note that free relatives are unbounded and

    sub/ect to island constraints# and therefore must involve standard# non-repro/ective

    wh-movement on earlier cycles prior to a last step of repro/ective movement)

    10 The grammaticality of ("%a& implies that even some varieties of nglish allow left-branche+traction# at least in comparative subdeletion cases li!e this. The comments in Lonati (007)"*-8&

    could provide a basis for understanding why the left-branch e+traction is possible in ("%a& but not("%d&. or discussion of whether comparative subdeletion involves movement or unbounded deletion#see resnan (19*7 homs!y (19**&.

    $

  • 8/12/2019 headmovement

    26/36

    ("$& a. I will visit what Tom says ill thin!s 5ary believes you will visit.

    b. RUI will visit what Tom believes the claim you will visit.

    onversely# the entire L, formed by repro/ection can move# and indeed undergo 3-

    movement)

    ("7& a. Jhat you will visit# (Tom says& I will visit.

    b. Jhat you will visit seems to have been visited by many tourists.

    'o it seems that each step of movement can in principle be either repro/ective or not#

    but once repro/ection ta!es place# it cannot be undone. The latter constraint can

    arguably be seen as an instance of the general no-tampering condition# in that once

    L has pro/ected the label cannot be unpro/ected but# conversely# until repro/ection

    ta!es place# it is always possible in principle. The locality properties are directly tied

    to the nature of pro/ected category) L,s# as in ("7b& can undergo 3-movement and the

    wh-phrases can undergo 36-movement# with the moved category obeying standard

    locality conditions in each cases. The @ properties of the structure resulting from

    movement category are crucial# as we have seen2 on the other hand# , appears to

    play no role in this approach.

    3gain# repro/ection appears to be a valid alternative approach which avoids

    the general difficulties discussed by homs!y. It leads to an interesting account of

    free relatives# and# in ury (199" of some cases of verb-movement. 4ow far it can

    be e+tended as a global alternative to head-movement remains to be seen# however

    (see :oenemann (000 iberauer H >oberts (008& for a repro/ective account of N-

    to-T movement&.

    +.@ Aoncluion

    4ere we have loo!ed at the three main alternatives to standard head-movement that

    have been discussed in the literature# in many cases directly responding to homs!y6s

    (001& comments. Bo single version is entirely free of problems# and none appears to

    be a global alternative to traditional head-movement in the sense that it is clear thatall former cases of head-movement can and should be reanalysed in the relevant

    7

  • 8/12/2019 headmovement

    27/36

    terms. This may in fact be a good state of affairs) it is ?uite possible that the

    mechanisms of head-movement were overe+tended in the earlier approach. 3t least in

    the L, and in the areas of verb-clustering it really seems that

  • 8/12/2019 headmovement

    28/36

    from as!ing ?uestions about !nowledge of language# ?uestions whose answers have

    led to the postulation of ]# to as!ing ?uestions about the !ind of world which

    produces a mental ob/ect li!e ] with the properties we observe it to have.

    This# in turn# has led to an emphasis on the third factor determining the

    nature of the adult language faculty.To see this what this means# observe that adult

    competence is the result of the interaction of three factors) (i& e+perience of the

    primary linguistic data (,@L we need this to learn the vocabulary and set the

    parameters of our native language2 (ii& ]niversal rammar# the innate endowment

    which ma!es it all possible# construed as a set of principles with parameters initially

    open2 (iii& ,rinciples not specific to the faculty of language.11 These principles

    constitute the third factor in language design# and include (a& principles of data

    analysis that might be used in language ac?uisition and other domains2 (b& principles

    of structural architecture and developmental constraints that enter into canaliAation#

    organic form# O # including principles of efficient computation .. It is the second of

    these subcategories that should be of particular significance in determining the nature

    of attainable languages (homs!y (00$a) 7&&. The first and second ?uestions may

    answer the epistemological ?uestion# but third-factor postulates seem to be implied in

    answering the metaphysical ?uestion. In a sense# we have to move beyond ] and#

    so# beyond e+planatory ade?uacy (in the homs!y (197%& sense&.

    In pursuing the a+iomatisation of principles# we sub/ect every postulate to

    a minimalist criti?ue) do we really need itR an it be reduced to something elseR

    Je want to get bac! to the first principles of synta+. Je want to reduce the

    theoretical postulates to those which are (virtually& conceptually necessary. 3t the

    same time# we want our e+planatory postulates to relate to the higher level of

    e+planation constituted by the attempt to answer the metaphysical ?uestion by

    invo!ing third-factor considerations. The 'trong 5inimalist Thesis ('5T& e+presses

    one hypothesis which can do this)

    ("*& @anguage is an optimal solution to legibility conditions (homs!y 000) 9*&

    11 This isn6t an entirely new idea) cf. there is surely no reason today for ta!ing seriously aposition that attributes a comple+ human achievement entirely to months (or at most years& of

    e+perience# rather than to millions of years of evolution or to principle o# neural or!aniCation thatma$ be even more deepl$ !rounded in ph$ical la7 (homs!y (197$)$9 emphasis mine&.

    8

  • 8/12/2019 headmovement

    29/36

    The notion of legibility conditions here relates to interface properties. 'o the idea is

    that the core computational system of synta+ provides the optimal way of relating an

    arbitrary set of le+ical items to the interfaces (, and @# for simplicity& in such a

    way as to satisfy whatever conditions the intrinsic properties of the le+ical items and

    the interfaces may impose.

    valuating head-movement# in any of its potential technical guises# against the '5T

    is difficult# since movement in general appears to be an unnecessary complication.

    'urely a system which lac!ed movement operations (of any !ind# 36# 3# head- or

    anything we might imagine& is simpler and more optimal than a system with such

    operations. homs!y (00%)110& provided a compelling negative response to this

    con/ecture) '5T entails that 5erge of ;# [ is unconstrained# therefore either e&ternal

    or internal. ]nder e+ternal 5erge# ; and [ are separate ob/ects2 under internal 5erge#

    one is part of the other 5erge yields the property of Fdisplacement6. To the e+tent

    that movement reduces to Internal 5erge (I5 then# we e+pect to find it in natural

    language.

    oncerning the status of narrow-syntactic head-movement# we might then reason that#

    all other things being e?ual# I5 and 5 are supposed to be e+actly the same

    operation e+cept that I5 ta!es place within a structure in the process of being while

    5 introduces the element to be merged from outside. 'ince 5 ?uite

    uncontroversially applies to heads# i.e. single le+ical items or feature bundles# we need

    a very good reason to treat I5 in a different way (this point is made by both Lonati

    (007& and >oberts (forthcoming&&. If head-movement is absent altogether# or

    restricted to the , interface# there must be an e+planation for this in terms of what

    differentiates I5 and 5.

    Kf course# 5erge is restricted to a search space. 5 can only loo! to the Bumeration2

    I5 is sub/ect to syntactic locality constraints. 'o if we can find a reason in the theory

    of locality for the absence of syntactic head-movement# we would have a principled

    reason to e+clude it from narrow synta+. The 3-over-3 principle is a good candidate.

    onsider the formulation of this condition given in ("& (fromDan!ua!e and ind

    (homs!y (007)%$&&)

    9

  • 8/12/2019 headmovement

    30/36

    ("8& If a transformation applies to a structure of the form

    C'O C3O D O D O

    for any category 3# then it must be interpreted so as to apply to the ma&imal

    phrase of the type 3.

    (4ere ma+imal is not intended in the

  • 8/12/2019 headmovement

    31/36

    add nothing to what we appear to have to assume anyway regarding movementMI5#

    and this seems to be most compatible with the '5T. 4owever# both approaches

    appear to have problems with some @ effects of head-movement. This suggests a

    combined approach) we could e+tend the operation of head-to-specifier movement

    (independently needed for some types of predicate cleft& and combine it with a ,

    rebrac!eting operation along the lines of 5arantA6s (198%# 1988& notion of merger.

    This would allow us retain the idea that I5 can apply to heads# and allow an account

    of the observed @ effects# while at the same time ac!nowledging that these cases of

    head-movement are partially morphological. This !ind of approach is advocated in

    >oberts (00$& and 5atushans!y (007&2 the difficulty with it is that the merger

    operation really has to be part of the head-movement operation in order to avoid the

    difficulties with c-command# the +tension ondition and successive cyclicity

    pointed out by homs!y. This entails something of a departure from pure I5# and

    so again creates a conceptual difficulty. It seems# then# that some alternative notion of

    incorporation may after all be needed.1

    Referen!es

    a!er# 5. (198$&Incorporation: , Theor$ o# Grammatical-unction Ahan!in!. ,hL

    dissertation# 5IT.a!er# 5. (1988& Incorporation: , Theor$ o# Grammatical-unction Ahan!in!.hicago) ]niversity of hicago ,ress.altin# 5. (00& 5ovement to the higher N is remnant movement# Din!uitic

    In?uir$ "") 7$"-9.elletti# 3. (1990& GeneraliCed *erb ovement. Turin) >osenberg H 'ellier.elletti# 3. (00%& 3spects of the @ow I, 3rea. In @. >iAAi (ed& The 'tructure o# Iand A. K+ford) K+ford ]niversity ,ress# pp. 17-$1.entAen# :. (00*& Erder and 'tructure in %mbedded Alaue in Forthern

    For7e!ian" PhD dissertation, University of &roms'"

    entAen# :. (to appear& F'ub/ect positions and their interaction with verb movement.6

    To appear in 'tudia Din!uitica.den esten# 4. (198"& Kn the interaction of root transformations and le+ical deletiverules# in J. 3braham (ed& En the ormal '$nta& o# the Wet!ermania. 3msterdam)ohn en/amins# pp. %*-1"1.den esten# 4. H . Jebelhuth (1990& 'tranding. In . rewendorf H J.'ternefeld (eds& 'cramblin! and Barrier.3msterdam) ohn en/amins# **-9.

    1 3 further issue concerns 3gree) why is this operation not sub/ect to the 3-Kver-3 onditionand what# if any# is its connection to head-movementR Kne might e+pect there to be some connectionsince 3gree is a head-head relation. ,erhaps we could envisage an approach which derives all the

    properties of head-movement from 3gree# adding essentially nothing to our conception of thatoperation. 'uch an approach would be conceptually very appealing# and is developed in >oberts(forthcoming&. ut this is not the place to evaluate that alternative.

    "1

    http://www.hum.uit.no/a/bentzen/K.Bentzen/Bentzen-Subj_VerbMove.pdfhttp://www.hum.uit.no/a/bentzen/K.Bentzen/Bentzen-Subj_VerbMove.pdf
  • 8/12/2019 headmovement

    32/36

    iberauer# T. H I. >oberts (008&. 'ub/ects# Tense and Nerb movement in ermanicand >omance. Aambrid!e Eccaional aper in Din!uitic ") %-%".(3vailable on-line at) http)MMwww.ling.cam.ac.u!MK,I@MNol"(Knline&Mvol".html&oec!+# . H '. 't/epanoviS (001& 4ead-ing towards ,. Din!uitic In?uir$ ")"%$-"$$.

    orsley# >obert L.# 5aria-@uisa >ivero and anig 'tephens (1997& @ong headmovement in reton# in >. orsley H I. >oberts (eds& The '$nta& o# the Aeltic

    Dan!ua!e# ambridge) ambridge ]niversity ,ress# $"-*%.resnan# . (19*7& vidence for a theory of unbounded transformations#Din!uitic

    ,nal$i) "$"-"9".ury# L. (00"& hrae 'tructure and 9erived 6ead. ,hL dissertation# ]niversityollege# @ondon.ury# L. (00*& Nerb movement and N'K-NK' alternations. AD Wor"in! aperin Din!uitic 19.arnie# 3.# 4. 4arley H '. Looley (00$& *erb irt.3msterdam) ohn en/amins.\avar# L. H . Jilder (199%& @ong head movementR Nerb movement and

    cliticiAation in roatian.Din!ua9") 1-$8.homs!y# B. (19$*& '$ntactic 'tructure. The 4ague) 5outon.homs!y# B. (197%& Aurrent Iue in Din!uitic Theor$. The 4ague) 5outon.homs!y# B. (197$&,pect o# the Theor$ o# '$nta&.ambridge# 5ass.) 5IT ,ress.homs!y# B. (19**& Kn Jh-5ovement. In 3. 3!ma/ian# ,. ulicover H T. Jasow(eds&ormal '$nta&. Bew Gor!) 3cademic ,ress# pp. *1-1".homs!y# B. (1987&Barrier.ambridge# 5ass.) 5IT ,ress.homs!y# B. (1991& 'ome Botes on conomy of Lerivation and >epresentation# in>. reidin (ed& rinciple and arameter in Aomparative Grammar. ambridge#5ass.) 5IT ,ress# pp. %1*-%$%.homs!y# B. (199"& 3 5inimalist ,rogram for @inguistic Theory# in :. 4ale H'.. :eyser (eds& The *ie7 #rom Buildin! 20: %a$ in Din!uitic in 6onor o#'$lvain Bromber!er. ambridge# 5ass.) 5IT ,ress# pp. 1-$.homs!y# B. (199$& The inimalit ro!ram. ambridge# 5ass.) 5IT ,ress.homs!y# B. (000& 5inimalist In?uiries) The ramewor!# in >. 5artin# L.5ichaels H . ]riagere!a (eds& 'tep b$ 'tep: %a$ on inimalit '$nta& in 6onoro# 6o7ard Dani". ambridge# 5ass.) 5IT ,ress# pp. 89-1$7.homs!y# B. (001& Lerivation by ,hase# in 5. :enstowicA (ed&Hen 6ale: , Di#ein Dan!ua!e. ambridge# 5ass.) 5IT ,ress# pp. 1-$.homs!y# B. (00& En Fature and Dan!ua!e. ambridge) ambridge ]niversity,ress.

    homs!y# B. (00%& eyond +planatory 3de?uacy# in 3. elletti (ed& 'tructureand Be$ond: The Aarto!raph$ o# '$ntactic 'tructure *olume + K+ford) K+ford]niversity ,ress# pp. 10%-1"1.homs!y# B. (00$a& Three actors in @anguage Lesign# Din!uitic In?uir$"7) 1-.homs!y# B. (00$b& iolinguistics and the 4uman apacity# lecture given at5T3# udapest# 5ay 1*# 00%2 ms. 5IThoms!y# B. (00*& 3pproaching ] from below# inInter#ace J 3ecurion K

    Dan!ua!e< Ahom"$L inimalim and the *ie7 #rom '$nta&-'emantic# ].'auerland and 4.-5. ^rtner. erlin) 5outon de ruyter# pp. 1-9.homs!y# B. (008& Kn ,hases# in >. reidin# .,. Kteroand 5.-@. ubiAarreta

    (eds& oundational Iue in Din!uitic Theor$ %a$ in 6onor o# ean-3o!er*er!naud. ambridge# 5ass.) 5IT ,ress.

    "

    http://mitpress.mit.edu/catalog/author/default.asp?aid=4788http://mitpress.mit.edu/catalog/author/default.asp?aid=35339http://mitpress.mit.edu/catalog/author/default.asp?aid=522http://mitpress.mit.edu/catalog/item/default.asp?ttype=2&tid=11493http://mitpress.mit.edu/catalog/author/default.asp?aid=4788http://mitpress.mit.edu/catalog/author/default.asp?aid=35339http://mitpress.mit.edu/catalog/author/default.asp?aid=522http://mitpress.mit.edu/catalog/item/default.asp?ttype=2&tid=11493
  • 8/12/2019 headmovement

    33/36

    homs!y# B. (007& Dan!ua!e and ind. ("rd edition&. ambridge2 ambridge]niversity ,ress.hung# '. (00$& Jhat rontsR Kn the N, >aising 3ccount of Nerb-Initial Krder_ In*erb-irt: aper #rom the Tucon Wor"hop# ed. 3ndrew arnie# 4eidi 4arley# and'heila 3nn Looley# 9-9. en/amins.

    in?ue# . (199%& Kn the evidence for partial B-movement in the >omance L,. In. in?ue et al (eds& ath To7ard niveral Grammar: 'tudie in 6onor o#

    3ichard '. Ha$ne# eorgetown) eorgetown ]niversity ,ress# 8$-110.in?ue# . (1999&,dverb and unctional ro(ection. K+fordMBew Gor!# K+ford]niversity ,ress.in?ue# . (00$& Leriving reenberg6s ]niversal 0 and its e+ceptions#Din!uitic

    In?uir$"7) "1$-"".in?ue# . (forthcoming& The '$nta& o# ,d(ective: , Aomparative 'tud$. ambridge#5ass.) 5IT ,ress.orver# B. H 4. van >iemsdi/! (199%& 'tudie on 'cramblin! : ovement and Fon-

    ovement ,pproache to ree Word-Erder henomena.

    erlin) 5outon de ruyter.Lonati# . (007& Kn Wh-4ead 5ovement# in @. heng H B. orver (eds& Wh-

    ovement: ovin! En. ambridge# 5ass.) 5IT ,ress# pp. 1-%7.mbic!# L. H >. Boyer (001& 5ovement operations after synta+# Din!uitic

    In?uir$")$$$-$97.monds# . (19*1& 3oot 'tructure-reervin! and Docal Tran#ormation. ,hddissertation# 5IT.monds# . (19*7&, Tran#ormational ,pproach to %n!lih '$nta&: 3oot 'tructure-

    reervin! and Docal Tran#ormation. Bew Gor!) 3cademic ,ress.monds# . (19*8& `The Nerbal omple+ N-N_ in rench#`Din!uitic In?uir$9)1$1-1*$.rewendorf# . H J. 'ternefeld (1990& 'cramblin! and Barrier.3msterdam) ohnen/amins. rewendorf H 'ternefeld.4alpern# 3. (199&,pproachin! 'econd. 'tanford) '@I ,ublications.4olmberg# 3. (1987& Word Erder and '$ntactic eature in 'candinavian Dan!ua!eand %n!lih. ,hL Lissertation# ]niversity of 'toc!holm.ohnson# :. (001& Jhat N,-llipsis an Lo# Jhat it an6t# ut Bot Jhy in 5.altin H . ollins (eds& The 6andboo" o# Aontemporar$ '$ntactic Theor$. K+ford)lac!well# pp. %"9-%80.:ayne# >. (1991& >omance litics# Nerb 5ovement and ,>K#Din!uitic In?uir$) 7%*-787.

    :ayne# >. (199%& The ,nti$mmetr$ o# '$nta ambridge# 5ass.) 5IT ,ress.:ayne# >. (000&arameter and niveral# K+ford) K+ford ]niversity ,ress.:ayne# >. H .-G. ,olloc! (001& Bew Thoughts on 'tylistic Inversion# in 3. 4ul!H .-G. ,olloc! (eds& 'ub(ect Inverion in 3omance and the Theor$ o# niveralGrammar# K+ford) K+ford ]niversity ,ress# pp. 10*-17.:iss# . H 4. van >iemsdi/! (00%& *erb Aluter. 3msterdam) ohn en/amins.:oeneman# K. (000& The "e&ible Fature o# *erb ovement. ]trecht# @KT,ublications.:oopman# 4. (198%& The $nta& o# verb-movement: #rom verb movement rule in the

    Hru lan!ua!e to niveral Grammar. Lordrecht) oris.:oopman# 4. H 3. 'Aabolsci (000& *erbal comple&e# ambridge# 5ass.) 5IT

    ,ress.@andau# I. (007& hain >esolution in 4ebrew N(,&-ronting# '$nta&9) "-77.

    ""

    http://www.bookfinder.com/dir/i/Studies_on_Scrambling-Movement_and_Non-Movement_Approaches_to_Free_Word-Order/3110135728/http://www.bookfinder.com/dir/i/Studies_on_Scrambling-Movement_and_Non-Movement_Approaches_to_Free_Word-Order/3110135728/http://www.bookfinder.com/dir/i/Studies_on_Scrambling-Movement_and_Non-Movement_Approaches_to_Free_Word-Order/3110135728/http://www.bookfinder.com/dir/i/Studies_on_Scrambling-Movement_and_Non-Movement_Approaches_to_Free_Word-Order/3110135728/http://www.bookfinder.com/dir/i/Studies_on_Scrambling-Movement_and_Non-Movement_Approaches_to_Free_Word-Order/3110135728/
  • 8/12/2019 headmovement

    34/36

    @asni!# 4. (199$& 3 Bote on ,seudogapping# In IT Wor"in! aper inDin!uitic 2M: aper in inimalit '$nta pp. 1%"-7". >eprinted with minorcorrections in 4. @asni! (1999&inimalit ,nal$i. K+ford) lac!well# pp. 1$1-1*%.@echner# J. (00$& Interpretive effects of head-movement ms ]niversity ofTVbingen (linguAAM0001*8&.

    @ema# . H 5.-@. >ivero (1990& @ong head-movement) , vs. 45. roceedin!o# F%D'0# """-%*.@ema# . H 5.-@. >ivero. 1991. Types of verbal movement in Kld 'panish) modals#futures and perfects.robu") "*-*8.@ongobardi# . (199%& ,roper names and the theory of B-movement in synta+ andlogical form#Din!uitic In?uir$$) 709-77$.5aha/an# 3. (00"& '$nta& at 'unet +: 6ead ovement and '$ntactic Theor$.]@3 Jor!ing ,apers in @inguistics# 10.5arantA# 3. (198%& En the F ature o# Grammatical 3elation. ambridge# 5ass.)5IT ,ress.5arantA# 3. (1988& litics# 5orphological 5erger and the 5apping to ,honological

    'tructure in 5. 4ammond H 5. Boonan (eds& Theoretical orpholo!$. Bew Gor!)3cademic ,ress# pp. $"-*0.5assam# L. (000& N'K and NK') 3spects of Biuean Jord Krder#arnie# 3. H .uilfoyle (000& The '$nta& o# *erb-Initial Dan!ua!e. K+ford) K+ford ]niversity,ress# pp. 9*-117.5assam# L. H . 'mallwood (199*& ssential eatures of ,redication in Biuean andnglish in :. :usumoto (ed& roceedin! o# F%D' 2M# @'3# ]niversity of5assachusetts# 3mherst. pp. "7-*.5atushans!y# K. (007& 4ead movement in linguistic theory# Din!uitic In?uir$"*)79-110.5cawley# . (19*1& Tense and time reference in nglish# in . illmore H L.@angendoen (eds& 'tudie in Din!uitic 'emantic. Bew Gor!) 4olt# >inehart andJinston# pp. 9$-11". >eprinted in . 5cawley (19*"& Grammar and eanin!# pp.$*-*.5clos!ey# . (1997& `The scope of verb-movement in Irish#`Fatural Dan!ua!e and

    Din!uitic Theor$1%) %*-10%.5Vller#. (1998&Incomplete Aate!or$ rontin!. Lordrecht) :luwer.5Vller# . (00%& Nerb-'econd as v,-irst ournal o# Aomparative Germanic

    Din!uitic*) 1"9-*%.Bilsen# . (00"& %liminatin! oition: '$nta& and 'emantic o# 'entenceodi#ication. ,hL Lissertation# ]niversity of ]trecht.

    ,oletto# . H .-G. ,olloc! (00%& Kn the left periphery of some >omance Jh-uestions# in @. >iAAi (eds& The 'tructure o# I and AN The Aarto!raph$ o#'$ntactic 'tructure *olume 2. K+ford) K+ford ]niversity ,ress# pp. $1-97.,olloc!# .G. (1989& Nerb 5ovement# ]niversal rammar# and the 'tructure of I,#

    Din!uitic In?uir$0# "7$-%%.,olloc!# .-G. (007& 'ub/ect litics and omple+ Inversion# in 5. veraert H 4.van >iemsdi/! (eds& The Blac"7ell Aompanion to '$nta& *olume I*. K+ford)lac!well# pp. 701-7$9.,olloc!# .-G.# ,oletto# . H B. 5unaro (00"& ppur si muove Kn comparingrench# ,ortuguese and ellunese Wh-5ovement# in ,. ,ica (ed& Din!uitic*ariation Yearboo". 3msterdam) ohn en/amins# pp. 1%*-180.

    >ac!ows!i# 3. H B. >ichards (00$& ,hase dge and +traction) 3 Tagalog ase'tudy#Din!uitic In?uir$"7) $7$-$99.

    "%

  • 8/12/2019 headmovement

    35/36

    >ac!ows!i# 3. H @. Travis (000& N-initial @anguages) < or omance#Din!ua89) 11"-1%1.>ivero# 5.-@. (199"b& initeness and 'econd ,osition in @ong 4ead 5ovement@anguages# ms. ]niversity of Kttawa.>ivero# 5.-@. (199%a& Begation# Imperatives and Jac!ernagel ffects#3ivita di

    Din!uitica1) 91-118.>ivero# 5.-@. (199%b& lause structure and N-movement in the languages of theal!ans.Fatural Dan!ua!e and Din!uitic Theor$1# 7"-10.>ivero# 5.-@. (199*& Kn Two @ocations for omplement litic ,ronouns) 'erbo-roatian# ulgarian and Kld 'panish# in 3. van :emenade H B. Nincent (eds&

    arameter o# orpho$ntactic Ahan!e. ambridge) ambridge ]niversity ,ress# pp.

    1*0-08.>ivero# 5.-@. H 3. TerAi (199$& Imperatives# N-5ovement and @ogical 5ood#

    ournal o# Din!uitic "1) "01-"".>iAAi# @. (1990&3elativiCed inimalit$. ambridge# 5ass.) 5IT ,ress.>iAAi# @. H I. >oberts (1989& omple+ Inversion in rench#robu1# 1-"0.>oberts# I. (1991& `+corporation and 5inimality#`Din!uiticIn?uir$# # 09-18.>oberts# I. (001& 4ead 5ovement# in 5ar! altin and hris ollins (eds&

    6andboo" o# '$ntactic Theor$. K+ford) lac!wells# pp. 11"-1%*.>oberts# I. (forthcoming& 4ead 5ovement) litics# Incorporation and Nerb-5ovement# ms. ]niversity of ambridge.>oss# . (197*& Aontraint on *ariable in '$nta 5IT ,hL Lissertation.>ussell# . (190$& Kn Lenoting#ind1%) %*9-%9".'iegel# 5. (198%& apping and Interpretation#Din!uitic In?uir$ $"-$"0.'hlons!y# ]. (00%& The form of 'emitic Boun ,hrases#Din!ua 11%) 1%7$-1$7.'portiche# L. 1999. 'ub/ect clitics in rench and >omance# comple+ inversion andclitic doubling. In :. ohnson and I. >oberts (eds.&. Be$ond rinciple and

    arameter: ea$ in memor$ o# Evaldo ae!!li. Lordrecht) :luwer# 189-.'tar!e# 5. (199"& %n deu&iOme poition en %urope Aentrale. 3 Lissertation#]niversity of eneva.'urXnyi# . (00$& 4ead movement and repro/ection. In) ,nnale niveritati'cientiarum Budapetineni de 3olando %PtvP Fominatae. 'ectio Din!uitica.

    Tomus

  • 8/12/2019 headmovement

    36/36

    Nicente# @. 007. The $nta& o# head and phrae: a tud$ o# verb phrae #rontin!.,h.L. dissertation) @eiden.Ni!ner# '. (199$&*erb ovement and %&pletive 'ub(ect in the Germanic Dan!ua!e#K+ford ]niversity ,ress# K+ford.Je+ler# :. H ,. ulicover (1980& ormal rinciple o# Dan!ua!e ,c?uiition.

    ambridge# 5ass.) 5IT ,ress.Ji!lund# 3. H :. entAen (00*& >ethin!ing 'candinavian Nerb 5ovement#

    ournal o# Aomparative Germanic '$nta&10) 0"-"".Ji!lund# 3.# . 4rafnb/argarson# :. entAen H T. 4rWarsdWttir (00*& The3fterglow of Nerb 5ovement# ms. ]niversity of Troms.wart# .-J. (001& 'yntactic and ,honological Nerb 5ovement# '$nta&%) "%-7.