LINGUIST List 12.2972

Wed Nov 28 2001

Sum: Movement Paradoxes

Editor for this issue: Marie Klopfenstein <>


  1. Carsten Breul, movement paradoxes (12-2826)

Message 1: movement paradoxes (12-2826)

Date: Sat, 24 Nov 2001 11:31:24 +0100
From: Carsten Breul <>
Subject: movement paradoxes (12-2826)

Dear all,

what follows is the summary of the responses to my query on 
movement paradoxes (see issue 12-2826, I will not repeat 
the query here because of the already considerable length 
of the present message.

Every reply I have got was helpful -- from reminding me of 
something which I had forgotten, over confronting me with 
further interesting observations to providing 
bibliographical leads. Below I pick out some of the points 
made, and give the bibliographical references (some of 
which mentioned in more than one message) which are 
directly concerned with movement paradoxes in P&P and MP 

The most extensive reply that I have got is from Norio 
Nasu. His message provides a very informative bird's eye 
view on movement paradoxes, giving much theoretical 
background and many more references (some of which also 
mentioned in replies by other respondents). Norio Nasu 
allowed me to quote extensively from his reply, which I am 
happy to do further below. 

The examples from Ward 1986/1988 (repeated at the end of 
this message as (3)) have been found dubious by And Rosta 
(speaker of BrE), who tends to think of such cases as 
performance errors: "[...] by the time the speaker gets to 
the auxiliary before the gap, they have forgotten the 
morphosyntactic form of the fronted V, so the paradoxes 
would be due to shortterm memory failures. I'd still 
conjecture that if US English [see also below] is different 
it arose, diachronically, from a grammaticization of such 
performance errors." 

For one AmE informant (3a) is ungrammatical and needs _did_ 
instead of _have_. For another respondent (BrE), _did_ is 
needed instead of _have_ in (3d) and preferable in (3c) as 
well; other similar examples, such as 

(4) They said we should stand firm, and stand firm we have 

are judged ok, though. For a third respondent (AmE), Ward's 
judgments are ok, and 

(5) We had to stand firm, and stood firm we have 

i.e. the version of (3d) where the perfective participle is 
used instead of the infinitive, is odd. Actually, Ward 
(ib.) also observes that in some cases a preposed 
infinitive is preferred over a preposed perfective 
participle. For still another speaker of AmE the versions 
where the verb in front position has the perfective 
participle form are not only odd, but "consistently 
It seems that we have quite significant idiolectal and/or 
dialectal differences here.

It was suggested that a Spanish example like 

(6) Ver no lo vi 
 ver: see-INF; no: not; lo: him; vi: saw
(I'm not sure what the best English translation of this 
construction is) 

may involve something similar as the movement-paradox 
examples. My impression is that (6) is rather a type of 
left dislocation/hanging topic construction. But this is, 
first, an uninformed guess of mine and, second, not to 
imply that the English constructions discussed above and 
this Spanish one have nothing in common. 

The Bresnan-judgments of (1a,b, 2a,b) (repeated at the end 
of this message) including the note that sentences like 
(2a) are accepted by some speakers, are found plausible by 
two of the respondents who commented on them. Carson 
Schutze suspects "that people who think [(1a, 2a)] sound OK 
are actually imagining that 'the fact' occurs before the in 
situ that-clause." For one speaker (1b, 2b) need a pause 
after the _that_-clause in front position. It is 
interesting to note in this context that for one speaker 
who comments on the Bresnan-example (Ib) mentioned by Norio 
Nasu (see below), "the star is at most a question mark." As 
the type of Bresnan-examples instantiated by Norio Nasu's 
(Ia,b) involves what is commonly considered to be A-
movement in contrast to A'-movement as in (1, 2) which show 
fronting/topicalisation/preposing, this judgment casts 
doubt on the existence of a movement-paradox in this type 
of A-movement context. 

Kleanthes Grohmann proposes that the binding- and island-
behaviour of the _that_-clause in front position be tested 
in order to find out whether it can be assumed to have 
moved there in the first place (for applications of such 
tests see e.g. Grohmann, K. 2000: Prolific peripheries, 
Diss. U of Maryland; 2000: "Copy left dislocation" in: 
Proceedings of the 19th West Coast Conference On Formal 
Linguistics; 2000: "A movement approach to contrastive left 
dislocation" in: Rivista di Grammatica Generativa 25). For 
example, finding a consistent pattern of judgments with 
respect to (7) under the distributive reading of the 
possessive may tell us if (7a) patterns with (7b), 
suggesting movement, or with (7c), not suggesting movement. 

a. [That his-i father was wrong] every-i boy talked about 
for days.
b. His-i father, every-i boy talked about a lot. 
c. His-i father, every-i boy talks about him. 

Annabel Cormack and Niina Zhang point out that 'movement 
paradoxes' may be problematic for the copy theory of 
movement, but not necessarily for alternative approaches 
(see e.g. Epstein, S. E. & Groat, R. & Kawashima, H. 
Kitahara, 1998: A Derivational Approach to Syntactic 
Relations, Oxford University Press; Cormack, Annabel & 
Smith, Neil, 2000: "Fronting: The syntax and pragmatics of 
'focus' and 'topic'": UCL Working Papers in Linguistics 

Carson Schutze proposes the following solution of the 
movement paradox: 

"I think there's a pretty simple story to be told here, 
which I however haven't seen in the literature: 
prepositions are unpronounced when they precede a clause. 
(Cf. the fact that in some languages, case markers are 
unpronounced on a DP adjacent to a verb, but obligatory if 
the DP is moved; or the fact that in English, infinitival 
complements to verbs like 'want' must be introduced by 
'for' except when that complement is adjacent to the verb.) 
I.e., this is a fairly 'superficial' PF fact about English, 
consistent with the fact that in Icelandic, for example, 
prepositions take CP complements with no problem. (Thus, 
for example, a story along the lines of Stowell's Case 
Resistance Principle cannot be right, unless one can find 
evidence that the Icelandic 'CPs' are really of a different 
syntactic category from English CPs.)"

I suggest to take also the following (perhaps even simpler) 
story into account: It can be assumed that the sources of 
(1b, 2b) are not those in which _of_ and _about_ take a 
_that_-clause as complement, but a [the fact that ...]-DP:

a. he didn't think of [the fact that he might be wrong]
 > [the fact that he might be wrong] he didn't think of 
b. we talked about [the fact that he was sick for days]
 > [the fact that he was sick] we talked about _t_ for 

The string _the fact_ can then be assumed to be deleted in 
the phonological component of the grammar or, perhaps 
preferably, during phonetic processing in performance.

The bibliographical hints I have got for P&P or MP analyses 
directly concerned with the paradoxes are these: 

Boskovic, Z. 1994: "Case properties of clauses and the 
Greed principle": Studia Linguistica 49: 32-53. 

Oku, Satoshi. 1998: A theory of selection and 
reconstruction in the minimalist perspective. Diss U of 

Oku, Satoshi. 1996: "Perfective participle paradox in 
English VP-fronting. In: Green, Antony Dubach & 
Motapanyane, Virginia (eds.). 1996: Proceedings of the 
thirteenth Eastern States conference on linguistics '96. 
Ithaca: Cornell U. 282-293. 

Webelhuth, Gert. 1992: Principles and parameters of 
syntactic saturation. New York & Oxford: OUP. 


Extensive quotes from Norio Nasu's reply: 

You can find an analysis of the patterns like your (1) and 
(2) in Boskovic (1994). 

In addition to the data of topicalisation/preposing, 
similar paradoxes are found in A-movement as well. For 
instance,Bresnan (2001:17) gives the following examples: 

a. That languages are learnable is captured by this 
b. *This theory captures that languages are learnable.
cf. This theory captures the fact that languages are 
learnable. (Bresnan 2001:17) 

Additionally, movement paradoxes can be seen as a matter of 
reconstruction. That is, sentences like (Ia, b) exhibit the 
circumstance where a moved category cannot be put back to 
its original position. From this viewpoint, the following 
sentences also exhibit a similar paradox. 

(II) a. The claim that John-1 was asleep seems to him-1 to 
be correct. (Chomsky 1995:204) b. *It seems to him-1 that 
John-1 was asleep. 

(IIb) violates Condition C of the binding theory. The 
question is why (IIa) doesn't violate this condition. 
Considering that the subject 'the claim that John was 
sleep' is moved from the infinitival complement, 
reconstruction of the subject into the embedded clause 
creates a structure where 'John' is bound by 'him.' 

In face of these paradoxes, there seem to be at least three 
approaches: (i) There is no movement. A seemingly moved 
category is in fact base-generated where it appears. (ii) 
Movement exists, but a moved category doesn't leave 
anything in its original position or leaves an element non- 
identical with it. (iii) Movement exists and a moved 
category leaves its copy. Movement paradoxes are 
attributable to some unknown factor(s). 

As far as A-movement is concerned, (i) is a position taken 
by constraint-based grammars like LFG and HPSG. There are, 
however, several works in the principles-and-parameters 
framework, which adopt (i). The most recent paper is 
Manzini and Roussou (2000). Williams (1994) also advocates 
the abolition of movement. Jacobson (1992) presents 
arguments for (i) in the framework of Categorial Grammar. 
Cormack and Smith (1997) may also be helpful. 

Nevertheless, movement paradoxes do not seem to lead 
immediately to the rejection of movement operations. The 
insight behind these phenomena is that a moved category and 
what it leaves behind are not always identical. 

This insight is reflected in the approach (ii) and the 
trace theory employed in the GB framework (cf. Fiengo 1977) 
might also belong to this group on the assumption that a 
trace is not identical with a moved category. 

Movement paradoxes may pose a more serious question to the 
copy theory of movement advocated by Chomsky (1995). In 
this respect, Lasnik (1999) for example considers that an 
argument does not leave any copy at all. 

In contrast to the data (seemingly) against movement (such 
as (I) above), it seems that the opposite possibility needs 
to be taken into consideration. That is, is there any 
phenomenon which indicates that a base position has to be 
postulated? What occurs to me in this connection is the 
phenomena called radical reconstruction or total 
reconstruction, whereby a moved category is forced to be 
interpreted in its base position. Of course radical/total 
reconstruction can also be accounted for without recourse 
to base positions of moved categories but it is equally 
possible to use the relevant data to argue for the 
existence of such positions. I have a chapter in my 
forthcoming PhD thesis (which will/has to be ready in two 
weeks time), where I explored this issue. 

Another point which seems to be of some relevance is the 
fact that there are certain differences between A- and A'- 
operations with respect to reconstruction. A common view 
is that an category undergoing A'-movement leaves its copy, 
whereas it remains controversial whether this is true for A-
movement. On the other hand, movement paradoxes are found 
not only in A-movement but also in A'- movement and head-
movement (see Bresnan 2001:18, 19). If movement paradoxes 
are regarded purely as a matter of reconstruction, it 
remains unclear how the A vs A' distinction in question 
should be handled. So, it seems to be desirable to examine 
whether movement paradoxes are really a matter of 


Boskovic, Z. 1994. Case properties of clauses and the Greed 
principle. _Studia Linguistica_ 49, 32-53. 

Chomsky, N. 1995. _The Minimalist Program_. MIT Press. 

Cormack, A. and N. Smith. 1997. Checking features and split 
signs. _UCL Working Papers in Linguistics_ 9. 223-252. 

Jacobson, P. 1992.Raising without movement. In R. K. Larson 
et al. (eds.) _Control and Grammar_. Dordrecht: Kluwer. 149-

Lasnik, H. 1999. Chains of arguments. In S. Epstein and N. 
Hornstein (eds.) _Working Minimalism_.Cambridge, Mass.: MIT 
Press. 189-215. 

Manzini, M. R. and A. Roussou. 2000. A minimalist theory of 
A-movement and control. _Lingua_ 110. 409-447. 

Williams, Edwin. 1994. _Thematic Structure in Syntax_. 
Cambridge, Mass.: MIT Press. 


a.	*He didn't think of that he might be wrong
b.	That he might be wrong he didn't think of 

a.	*We talked about that he was sick for days
b.	That he was sick we talked about for days 

a.	We had both been thrown into the water to sink or swim, 
and swim we had -- we had swum from very far apart 
b.	They told him that he had to be there all day long and 
be there all day long he has!
c.	They provided us with enough beer to drink all day long 
and drink all day we have!
d.	We had to stand firm, and stand firm we have! 

Dr. Carsten Breul
Universit�t Dortmund
Institut f�r Anglistik und Amerikanistik
Emil-Figge-Str. 50
44221 Dortmund
phone: (0231) 755-2898
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue