[FREE PDF sample] Minimalist Syntax Randall Hendrick ebooks

Visit https://ebookfinal.com to download the full version and
explore more ebooks
Minimalist Syntax Randall Hendrick
_____ Click the link below to download _____
https://ebookfinal.com/download/minimalist-syntax-
randall-hendrick/
Explore and download more ebooks at ebookfinal.com

Here are some suggested products you might be interested in.
Click the link to download
A Minimalist Theory of Simplest Merge 1st Edition Epstein
https://ebookfinal.com/download/a-minimalist-theory-of-simplest-
merge-1st-edition-epstein/
Hemophilia 1st Edition Randall Mcpartland
https://ebookfinal.com/download/hemophilia-1st-edition-randall-
mcpartland/
Syntax An Introduction Talmy Givon
https://ebookfinal.com/download/syntax-an-introduction-talmy-givon/
Analysing English sentences A Minimalist Approach 1st
Edition Andrew Radford
https://ebookfinal.com/download/analysing-english-sentences-a-
minimalist-approach-1st-edition-andrew-radford/

The Ashgate Research Companion to Minimalist and
Postminimalist Music Potter
https://ebookfinal.com/download/the-ashgate-research-companion-to-
minimalist-and-postminimalist-music-potter/
Understanding Syntax 3rd Edition Maggie Tallerman
https://ebookfinal.com/download/understanding-syntax-3rd-edition-
maggie-tallerman/
Understanding Syntax 4th Edition Maggie Tallerman
https://ebookfinal.com/download/understanding-syntax-4th-edition-
maggie-tallerman/
Shore Fishes of Hawaii John E. Randall
https://ebookfinal.com/download/shore-fishes-of-hawaii-john-e-randall/
Perry s chemical Engineer s handbook Section 4 8th edition
Edition Hendrick C. Van Ness
https://ebookfinal.com/download/perry-s-chemical-engineer-s-handbook-
section-4-8th-edition-edition-hendrick-c-van-ness/

Minimalist Syntax Randall Hendrick Digital Instant
Download
Author(s): Randall Hendrick
ISBN(s): 9780631219415, 0631219412
Edition: illustrated edition
File Details: PDF, 1.17 MB
Year: 2003
Language: english

Generative Syntax
General Editor: David Lightfoot
Recent work in generative syntax has viewed the language faculty as a system
of principles and parameters, which permit children to acquire productive
grammars triggered by normal childhood experiences. The books in this series
serve as an introduction to particular aspects or modules of this theory. They
presuppose some minimal background in generative syntax, but meet the
tutorial needs of intermediate and advanced students. Written by leading
figures in the field, the books also contain sufficient fresh material to appeal to
the highest level.
1 Government and Binding Theory and the Minimalist Program
Edited by Gert Webelhuth
2 Logical Form
Norbert Hornstein
3 Minimalist Analysis
Howard Lasnik
4 Phrase Structure: From GB to Minimalism
Robert A. Chametzky
5 Move! A Minimalist Theory of Construal
Norbert Hornstein
6 Derivation and Explanation in the Minimalist Program
Edited by Samuel David Epstein and T. Daniel Seely
7 Minimalist Syntax
Edited by Randall Hendrick

Minimalist Syntax
Edited by
Randall Hendrick

Contents
Acknowledgments vi
List of Contributors vii
Introduction Randall Hendrick 1
1 On Control 6
Norbert Hornstein
2 On Logical Form 82
Danny Fox
3 Steps toward a Minimal Theory of Anaphora 124
Howard Lasnik and Randall Hendrick
4 Syntactic Variation, Historical Development, and Minimalism 152
Höskuldur Thráinsson
5 Phrase Structure 192
Robert A. Chametzky
Index 226

Acknowledgments
This work was supported in part by NSF grant no. 0112231 and by an award
from the Arts and Sciences Foundation of the University of North Carolina at
Chapel Hill.

List of Contributors
Robert A. Chametzky, 736 Wheaton Road, Iowa City, IA 52246, USA. Contact
email address: ui0049@blue.weeg.uiowa.edu
Danny Fox, Department of Linguistics and Philosophy, E39–245, Massachu-
setts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139,
USA. Contact email address: fox@mit.edu
Randall Hendrick, Department of Linguistics, University of North Carolina,
Chapel Hill, NC 27599-3155, USA. Contact email address: hendrick@unc.edu
Norbert Hornstein, Department of Linguistics, 1401 Marie Mount Hall, Univer-
sity of Maryland, College Park, MD 20742-7505, USA. Contact email address:
nh10@umail.umd.edu
Howard Lasnik, Department of Linguistics, 1401 Marie Mount Hall, Univer-
sity of Maryland, College Park, MD 20742-7505, USA. Contact email address:
lasnik@wam.umd.edu
Höskuldur Thráinsson, Department of Icelandic, University of Iceland,
Arnagardi v. Sudurgoetu, 101 Reykjavik, Iceland. Contact email address:
hoski@rhi.hi.is

Introduction
Randall Hendrick
Syntactic theories change over time because they, like all scientific theories,
confront the twin selectional pressures of conceptual simplicity and empirical
adequacy. These forces shape the historical trajectory of theories, and no theory
that we currently know will survive these challenges unchanged. And in this
context we should not be surprised that there is not just one theory with such
an historical arc. Although the existence of competing theories suggests to
some outside observers the presence of confusion in a field of inquiry, competi-
tion is in fact useful to practicing researchers for the simple reason that it
highlights the selectional pressures that give theories life. It is in this spirit of
theoretical evolution that syntacticians should greet the prospect of a minimalist
syntax with enthusiasm.
There are admirable histories of modern phonological theory (e.g. Anderson,
1985). But no one has, as yet, produced a similar companion to syntactic theory,
perhaps because of the epic proportions such a tale has. If it is ever written, I
believe it will depict syntacticians caught in a struggle between the same two
titans that have shaped the phonological world: the colossus of rules and the
colossus of representations.
In the early 1960s researchers explored with initial confidence the hypothe-
sis that grammars were collections of rules that expressed systematic pat-
terns of a language. Work was done, of course, on how these rules should be
formulated (e.g. whether recursion was a property of phrase structure rules),
but much more work was done on issues of representations, for example
whether the history of application of phrase structure rules should be encoded
in a representation (the familiar trees), whether NPs that did not branch were
to be “pruned,” whether negatives were verbs, whether the level of deep
structure existed, and the like. There were interesting questions about rule
application, most notably the theory of the transformational cycle, but most of
the work done in the extended standard theory and in its wake has been
concerned with motivating representations. To see this point one needs only
to look at the vast literature on whether the site of a movement carried some
representational “trace” of the moved constituent. The conceptual gambit that
was replayed in various contexts was to isolate distinct representational sys-
tems, say the lexicon and the transformational component, and to place limits
on their representations that had some important empirical consequences. Such

2 Introduction
a strategy is at work in Chomsky’s (1965) elimination of generalized trans-
formations and in his (1970) analysis of derived and gerundive nominals. It is
also at play in Bresnan’s (2001) distinction between lexical functions respons-
ible for relating active and passive verbs in terms of grammatical relations,
and other operations that build constituent structures. The Government and
Binding Theory of Chomsky (1981) is firmly situated in this representational
tradition. It is a theory with a rich inventory of principles that perform a
filtering function on highly articulated representations. Only the subjacency
principle had a non-representation flavor, and there were strong attempts to
recast it in representational terms.
A completely different strategy of theory construction would be to appeal to
rules rather than representations to carry a dominant explanatory burden.
Some strands of generative semantics explored this route, notably Postal (1974),
but the general strategy encountered resistance based in part on doubts about
the cognitive plausibility of such rules. The negative history of the deriva-
tional theory of complexity summarized in Tannenhaus (1988) and the more
specific appraisal of the transformational rule of passivization in Bresnan (1978)
illustrate well the contours of this debate. It remains possible that, while there
may be no rules in the sense of Chomsky (1965), where transformations paired
structural descriptions and structural changes, properties of derivations pro-
duced by very general, universal operations such as Move α might still be the
locus of explanation for an important class of syntactic phenomena. It is in this
line of research that the Minimalist Program situates itself.
Some of the basic conjectures of Minimalism may already be familiar to the
reader. Among them is included the idea that intermediate levels of syntactic
representation (such as d-structure or s-structure) are to be eliminated if at all
possible for reasons of conceptual parsimony because only levels that interface
with other cognitive systems, such as Phonological Form and Logical Form
(LF), are necessarily required. Movements are unified with the composition of
phrasal structure as one general computational operation (Merge), making it
possible to intersperse movements and the insertion of lexical material. Move-
ments apply obligatorily in the presence of a triggering condition, which is
hypothesized to be a formal feature that cannot be interpreted at an interface
level. Such movements are motivated to remove this uninterpretable feature
and are subject to substantive conditions of economy. These conditions have
at least the appearance of a least effort quality: they opt wherever possible
for covert movements at LF rather than overt movements because LF move-
ments affect single features rather than the complex feature sets that make up
phrasal constituents. These and other tenets of Minimalism set out in Chomsky
(1995) are neatly catalogued in Epstein and Hornstein (1999). Here it is worth
trying to appreciate why such claims come to have the status of theoretical
imperatives.
Minimalism is an attempt to rethink many syntactic phenomena in the
hopes of extracting greater insight by questioning the explanatory role of
representations. One approach to this project is that it is a therapeutic labor
with the goal of eliminating unnecessary representational mechanics. On this
view, Minimalism insists on a maximally elegant theory with the minimum of

Introduction 3
theoretical primitives and statements consistent with familiar goals of descrip-
tion and explanation. Thus, Minimalism is motivated by a priori commit-
ments to simplicity, elegance, and theoretical parsimony. This is the corrosive
applied to the intermediate levels of d-structure and s-structure. It is possible,
though, to embrace Minimalism for substantive reasons. From this perspective
it is an attempt to formalize a substantive concept of economy in syntax that
will resolve a variety of empirical problems that have proven recalcitrant
to previous theorizing in representations. This concept of economy itself is
not purely formal but stands in a special relation to the interface with other
cognitive (sub)systems such as the sound system and the semantic system,
or with some quite general principle of limiting cognitive effort. Every bit
of representation is to be cached out by interpretation at the interface of one
of these (sub)systems, and some types of operations are identified as entail-
ing greater cognitive effort. From this perspective Minimalism instantiates a
kind of functionalism, though it has different commitments and caveats than
other styles of functionalism, as has been pointed out in Pesetsky and Torrego
(2000).
As he observed the inception of the Minimalist Program, Marantz (1995)
announced the death of syntax, echoing Nietzsche’s epigram about the deity.
From that perspective, the very title of this book is something of an oxymoron.
But Marantz’s obituary was provocative and, in the eyes of some syntactic
theorists, premature (see, for example, Lappin, Levine, and Johnson (2000),
and the replies and rejoinders it engendered). Clearly, some syntactic gods
struggle on. There are researchers who have advocated the inevitability of
rules dependent on constructions, and still others who advocate decomposing
syntax into a number of distinct representations. In contrast, as I noted above,
the minimalist project limits the autonomy of syntax and places a significant
explanatory burden on how syntax facilitates, and perhaps even optimizes,
the interpretation of interfacing cognitive systems. It is the severity of this
stricture that was the foundation of Marantz’s assessment of the eclipse of
syntax. Nevertheless, Lasnik (1999, p. 6) has observed that “we are very far
from being able to confidently proclaim any Minimalist details of syntactic
theory.” Such an assessment is sobering, coming as it does from one of the
major proponents of Minimalism. This book takes the challenge of Lasnik’s
cautionary note seriously. It brings together a group of studies that tries to pin
down the specific details of a Minimalist explanation for such important syn-
tactic phenomena as control, anaphora, scope of quantification, verb raising,
and phrase structure. Its search for a substantive theory of syntactic economy
and its re-evaluation of the advantages of derivational explanations lead to
new perspectives and, in many cases, specific results which I hope can answer
Lasnik’s challenge.
This volume collects some reasons to think that a minimized syntax offers
more explanatory value for those interested in understanding why natural
languages have robust recurrent properties. It does this by contributing to
our understanding of the specific effects of the substantive theory of syn-
tactic economy that is at the heart of Minimalism. Throughout, it emphasizes

4 Introduction
empirical problems that can be given an insightful explanation when the correct
substantive theory of syntactic economy is adopted. In Chapter 1, Hornstein
takes on the complex and stipulative theory of control outlined by Govern-
ment and Binding Theory and develops a competing theory that rethinks the
phenomenon of obligatory control as another instance of movement. The idea
that obligatory control is a species of raising is built on questioning the premise
in early minimalist work that there is a fundamental difference between them-
atic roles and morphological features. Only morphological features such as
Case can be the motive for movement, not thematic roles. If Hornstein is right,
movement can be equally driven by thematic roles. It is also standardly assumed
in minimalist syntax that principles of syntactic economy favor covert move-
ment at the expense of overt movement.
Fox reviews, in Chapter 2, the syntactic treatment of quantified expressions
that were raised after s-structure by a covert movement, Quantifier Raising
(QR), to derive the level of Logical Form in Government and Binding Theory.
He questions whether covert application of QR is needed given the copy theory
of movement in a minimalist syntax where one of the copies typically is forced
to delete before a structure is interpreted at Phonological Form. Fox makes
a fascinating case that Quantifier Raising is an overt movement rather than
a covert one as has been traditionally thought. His argument is based on a
careful examination of the interaction of QR with the theory of anaphora and
gives new importance to extraposition phenomena, which are assimilated to
(overt) QR.
The interaction of covert movements and the theory of anaphora is central
to Lasnik’s and my own Chapter 3. We assemble a range of evidence to con-
clude that covert movement does not change anaphoric relations. The evi-
dence also suggests that scope in raising constructions is determined by the
surface position of a DP while other LF licensing restrictions on anaphora,
bound pronouns, idioms, and negative polarity items are free to apply to any
member of a movement chain. This means rather surprisingly that, contrary to
what has been widely believed, there is no scope ambiguity in raising con-
structions and that there is a kind of “A reconstruction.”
Thráinsson takes on the variation of ordering verbs and adverbs in Western
Germanic languages that has been used to motivate verb raising in Chapter 4.
Thráinsson argues that verb raising is subject to two economy considerations.
First, raising never applies optionally and only applies to eliminate an un-
interpretable agreement feature(s). Second, languages are forced to have such
a feature(s) if there is overt evidence of an agreement element independent of
tense. The four chapters I have just mentioned are principally concerned with
minimalism as a substantive claim about syntactic economy.
The facet of minimalism that requires economy of conceptual apparatus
predominates in the fifth chapter. There Chametzky applies this conceptual
corrosive to the familiar concept of phrase structure. Chametzky deploys
several conceptual arguments that lead to the conclusion that Minimalism
is incompatible with phrase structure. In the process of exploring clausal
organization he identifies adjuncts as potentially forming a limit to minimalist
analysis.

Introduction 5
References
Anderson, S. (1985). Phonology in the twentieth century: theories of rules and theories of
representations. Chicago: Chicago University Press.
Bresnan, J. (1978). A realistic transformational grammar. In M. Halle, J. Bresnan, &
G. A. Miller (eds.), Linguistic theory and psychological reality (pp. 1–59). Cambridge,
MA: MIT Press.
Bresnan, J. (2001). Lexical functional grammar. Oxford & Malden, MA: Blackwell.
Chomsky, N. (1965). Aspects of the theory of syntax. Cambridge, MA: MIT Press.
Chomsky, N. (1970). Remarks on Nominalizations. In R. Jacobs & P. Rosenbaum (eds.),
Readings in English Transformational Grammar (pp. 184–221). Waltham, MA: Ginn.
Chomsky, N. (1981). Lectures on government and binding. Dordrecht: Foris.
Chomsky, N. (1995). The minimalist program. Cambridge, MA: MIT Press.
Epstein, S. D., & Hornstein, N. (1999). Introduction. In S. D. Epstein, & N. Hornstein
(eds.), Working minimalism (pp. ix–xviii). Cambridge, MA: MIT Press.
Lappin, S., Levine, R. D., & Johnson, D. E. (2000). Topic . . . Comment. Natural language
and linguistic theory, 18, 665–71.
Lasnik, H. (1999). Minimalist analysis. Oxford & Malden, MA: Blackwell.
Marantz, A. (1995). The minimalist program. In G. Webelhuth (ed.), Government and
binding theory and the minimalist program: principles and parameters in syntactic theory.
Oxford & Malden, MA: Blackwell.
Pesetsky, D., & Torrego, E. (2000). T to C movement. Causes and consequences. In
M. Kenstowicz (ed.), Ken Hale: a life in language (pp. 355–426). Cambridge, MA:
MIT Press.
Postal, P. (1974). The best theory. In S. Peters (ed.), Goals of linguistic theory (pp. 13–70).
Englewood Cliffs, NJ: Prentice-Hall.
Tannenhaus, M. (1988). Psycholinguistics: an overview. In F. Newmeyer (ed.), Language:
psychological and biological aspects. Linguistics, the Cambridge survey, volume 3 (pp. 1–
37). Cambridge: Cambridge University Press.

6 Norbert Hornstein
Chapter one
On Control
Norbert Hornstein
1.1 Introduction
Recently, control has become a hot topic, largely for Minimalist reasons.1
In
particular, the current passion for reexamining the conceptual foundations of
Universal Grammar (UG) has prompted grammarians to reconsider whether,
and how much, the grammatical processes underlying control configurations
differ from the movement operations that lie behind raising constructions. The
idea that the grammatical processes responsible for these two types of struc-
tures are less different than generally believed is not novel.2
What is different
is the conceptual setting afforded by Minimalist commitments, most import-
antly the premium now placed on “simpler” theories that eschew both theory
internal levels (like d-structure) or formatives (like PRO) and multiple ways of
establishing grammatical dependencies (by either movement or construal). The
venerable dualism between control and raising offers a tempting target for
those impressed with the razor sharp Ockhamism of the Minimalist program.
Before reviewing the issues in detail, it is worth pausing to slightly expand
on the cursory comments above as they provide some of the methodological
motivations for reconsidering the distinction between raising and control. The
main motivation comes from two sources.
First, the observation that control brings with it considerable grammatical
baggage, for example a whole additional module of the grammar (the Control
module whose function it is to determine the controllers of PRO and the
interpretation that a particular control structure carries), a theory internal form-
ative – PRO – with its own idiosyncratic distributional requirements (it occurs
in the subject positions of non-finite clauses and this prompts otherwise con-
ceptually and empirically problematic technology, for example null case, to
track this fact), and a set of grammatical processes (construal rules) added to the
movement processes already assumed to be available, whose function it is to estab-
lish dependencies quite similar to those that movement already affords.
Second, the observation that the whole distinction conceptually rests on
assuming properties of UG that are currently considered problematic from
an MP perspective. Two such problematic conditions are d-structure and that
part of the theta criterion that stipulates that theta roles and chains be biuniquely

On Control 7
related (viz. be in a one to one relation). Both properties are methodologically
suspect given minimalist commitments. Given Chomsky’s (1993) objections to
postulating levels like d-structure, it is natural to see if it is possible to elimin-
ate d-structure and its various requirements entirely from the grammar.
Moreover, from a purely conceptual point of view it is quite unclear why
chains should be barred from bearing more than one theta role. Eliminating
both d-structure and the biuniqueness condition on theta roles has obvious
conceptual attractions. Interestingly, neither assumption fits comfortably with
the assumption that control, like raising, is derived via movement. Thus re-
thinking the relation between movement and control immediately bears on
these methodological issues and thereby on questions of optimal grammatical
design.
This chapter reviews some of the recent hubbub surrounding these issues.
Before proceeding, however, one last caveat is in order. Like all reviews, this
one will suffer from the limited knowledge, perspective, interests and preju-
dices of the reviewer. The route taken here is but one of many. I hope that the
trip proves interesting. However, I am sure that it is not the only worthwhile
itinerary, nor the least contentious. Caveat lector!
1.2 Raising versus control
From the earliest work in Generative Grammar, Raising (1a) has been con-
trasted with Control (1b).
(1)a. John seems/appears/is likely to like Mary
b. John tried/hoped/expected to like Mary
The contrast has been drawn along several dimensions.
First, there are various empirical differences between the two kinds of con-
structions despite the obvious surface similarity. These include the following.
In (1a) John is understood as discharging a single thematic function. He is
the liker of Mary. This is coded by assigning a single theta role to John, that of
the external argument of the embedded verb like. In contrast to this, John in
(1b) is perceived as bearing two roles. He is at once Mary’s liker and also a
trier/hoper/expector. More, theoretically, John is related to a pair of thematic
positions, the external argument of the matrix predicate and that of the
embedded one.
This difference in the thematic properties of raising versus control brings
with it a host of others. Two are noteworthy.
First, the subject position of raising predicates can be filled with expletives
and idioms (all the while retaining the idiomatic interpretation) while those of
control predicates cannot be.
(2)a. There seemed/appeared/was likely to be a man here
b. *There tried/hoped/expected to be a man here

8 Norbert Hornstein
(3)a. The fat seemed/appeared/was likely to be in the fire
b. *The fat tried/hoped/expected to be in the fire
Second, passivizing the embedded clause in raising constructions leaves the
original meaning of the sentence largely intact. (4a,b) illustrate this property of
“voice transparency”; the two sentences being virtual paraphrases of one another.
(4)a. The doctor seemed to examine Mary
b. Mary seemed to be examined by the doctor
This contrasts sharply with what happens if we passivize in a control struc-
ture. (5a,b) are not voice transparent, not being even remote paraphrases.
(5)a. The doctor tried to examine Mary
b. Mary tried to be examined by the doctor
These two facts can be directly tied to the thematic differences between
raising and control verbs. One can explain the contrast between (2) and (3)
by observing that expletives like there and idioms cannot bear theta roles.
However, this is what they must do in examples like (2b) and (3b), on the
further assumption that all theta roles of a predicate must be assigned. The
trouble with (2b) and (3b) is that the theta roles that must be assigned cannot
be assigned to any appropriate element given the resistance it is natural to
assume that expletives and idioms have to bearing such roles. In (2a) and (3a)
there is only one theta role to assign and as such no similar problem arises.
Analogous considerations account for (4) and (5). In control structures like
those in (1b), the matrix subject fulfills two thematic requirements; that of the
matrix verb and that of the embedded one. Thus, in (5a) the doctor is both trier
and examiner, while Mary is merely the examinee. However, by passivizing the
embedded clause, Mary becomes the matrix subject and thereby assumes two
thematic functions, trier and examinee, while the doctor has just one thematic
role, the examiner. This accounts for the clear difference in meaning between
(5a,b). In contrast, the subject position of raising predicates is not associated
with any thematic role. Consequently, similar manipulation of the embedded
clause in (4) has no thematic repercussions. The thematic status of the doctor
in both (4a,b) is that of examiner and that of Mary is that of examinee. As the
thematic properties of the two constructions are undisturbed, the sentences
remain essentially paraphrases.
These empirical differences between raising and control structures have
generally been traced to an underlying difference in grammatical etiology.
Specifically, the operations underlying raising configurations are distinct from
those that generate control structures. The former are the province of move-
ment rules. In contrast, control structures are either formed by applications of
the non-movement rule Equi NP deletion (in the Standard Theory) or, in more
recent Government Binding (GB) analyses, control is the result of construal
rules that relate a phonetically null DP, “PRO,” to its antecedent. Structurally,
then, raising structures differ from control configurations as in (6).

On Control 9
(6)a. John1 seems [t1 to like Mary]
b. John1 seems [PRO1 to like Mary]
(6a) indicates that John has moved from the embedded subject position, leav-
ing a co-indexed trace in its movement site. In (6b) John controls a base gener-
ated null expression PRO. The indicated indexations, therefore, arise from two
different processes: movement for raising and construal for control. Similarly,
the empty categories arise from two different processes: movement in the case
of raising and lexical insertion in the case of control. The empirical differences
noted above trace back to these different derivational histories that invoke
different types of grammatical operations and employ different formatives.
It is worth considering the details. Specifically: what forces movement in
(6a) and requires construal in (6b)? The answer, in a GB (and also an Aspects)
style theory, is d-structure. The theoretical basis within GB for distinguishing
the two constructions relies on contrasting traces and PROs: PROs head chains,
traces do not; d-structure implements this difference. In fact, the classical dis-
tinction between raising and control follows seamlessly from the assumption
that d-structure exists. Consider the reasoning.
D-structure has two distinctive properties: it is input to the transformational
component and the locus of all thematic discharge; a representation of “pure
GF-θ.”3
Thus, prior to “displacement” operations (i.e. transformations) that
rearrange phrase markers, words/morphemes are assembled into d-structure
phrase markers by being lexically inserted into the available theta-positions.
After lexical insertion, transformations apply to map d-structure phrase markers
into others.
Given the requirements of d-structure, transformations cannot relate theta-
positions (via movement) as all theta-positions have been filled by lexical
insertion in forming the d-structure phrase marker. In particular, d-structure
prohibits the movement in (7) from the embedded subject to the matrix sub-
ject position as the matrix subject is a theta position and so must be filled at
d-structure. Consequently, it is unavailable for occupancy via movement.
(7) John1 tried [t1 to like Mary]
In this way, d-structure thematic requirements make movement between θ-
positions impossible and thus prohibit control relations (which involve multi-
ple theta-roles) from being the observed manifestations of movement operations.
Furthermore, if d-structure has only θ-positions filled (in addition to all such
positions being filled), then raising structures must be products of movement.
In particular, a structure like (8), a non-movement version of a raising con-
struction, is ill-formed at d-structure given that the matrix non-theta subject
position is filled by John. Given d-structure thematic requirements, this posi-
tion must be vacant at d-structure, that is it cannot be filled by lexical inser-
tion. The only remaining option is to fill it at some later phase of the derivation,
that is by movement.
(8) John1 seemed [PRO1 to like Mary]

10 Norbert Hornstein
In sum, the classical vision of d-structure as the representation of pure
GF-θ, that is the phrase marker where all and only thematic information is
grammatically rendered, theoretically forces the empirical distinction between
raising and control. Raising is required where d-structure thematic conditions
prohibit the insertion of lexical material (e.g. the subject position of a raising
verb as in (6a)) while it is prohibited where d-structure thematic conditions
require the presence of lexical insertion (e.g. the subject positions of the control
complement and matrix subjects in (6b)).
The Theta-Criterion further buttresses this view of d-structure, in par-
ticular the idea that all thematic information is discharged via lexical inser-
tion. The relevant feature of the Theta-Criterion is the requirement that there
be a biunique relation between θ-roles and chains, in particular, that every
chain bear at most one θ-role. This effectively prohibits all movement from
one θ-position to another. But if movement into θ-positions is forbidden,
yet all θ-roles must be discharged, then the only alternative is to fill each
θ-position via lexical insertion. The step from the Theta-Criterion to the postu-
lation of PRO and construal rules that relate PROs to their antecedents is a
short one.4
In sum, given the canonical view about d-structure and the Theta-Criterion,
we theoretically derive the fact that raising is due to movement while control
is due to construal. From this, the observed empirical differences follow. It’s a
nice story and it has had a lot of staying power. Nonetheless, there are prob-
lems given minimalist scruples. Here are three.
First, it requires the postulation of a theory internal formative PRO. PRO is
similar to, but different from a trace. In a GB style theory, at LF, both traces
and PROs are categories without phonetic contents. They are of the form
“[NP e]1.” PROs have the same structure at LF. However, whereas traces are
the residues of movement, PROs are lexical expressions which receive their
indices not in the course of the overt derivation but at LF via construal opera-
tions. That a lexical item (PRO) and a grammatical formative (trace) should be
essentially identical at LF is quite unexpected. Indeed, identity of structure
suggests that either both are lexical items or both are grammatical formatives.
Within a minimalist setting, things get worse. Nunes (1995) and Chomsky
(1998) have noted that traces are theory internal constructs which are best
avoided on conceptual grounds if possible. In place of traces, copies have been
pressed into service.5
This leaves the proposed structure of (6a) as (9).
(9) John seemed [John to like Mary]
However, (9) now raises a question for PRO. What is its structure? What kind
of lexical item do we have here? It cannot simply be a null pronoun, as in
many environments, for example (6b), it requires an antecedent. It might
simply be a null reflexive, but if so why must it be phonetically null? Recall
that minimalism resists the postulation of theory internal entities. Thus, the
more the features of PRO are idiosyncratic (i.e. the more PRO is distinguished
from more run of the mill lexical items) the less explanatory weight PRO has
in a minimalist setting. Humdrum is best. However, within GB, PRO is hardly

On Control 11
run of the mill given that it brings in its train a whole module of grammar, the
Control module, whose job it is to find its antecedent and provide an interpreta-
tion depending on its grammatical setting. We shall see that PRO continues to
enjoy a special status in some recent reanalyses.
A second worry arises from the fact that control requires construal rules.
Recall that Minimalism places a premium on simple theories. In the present
context, it favors theories of UG that minimize rule types. A theory that has
both movement rules and construal rules has two ways of establishing inter-
nominal dependencies in the grammar. Ockham dictates that a theory with
just one set of operations is preferred. On the assumption that movement is
independently required (e.g. for feature checking), this suggests that construal
processes should be eliminated, including those that underlie control.6
A third problem with the standard GB approach to control lies with its
reliance on d-structure. As outlined above, the distinction between raising and
control relies on the thematic requirements d-structure places on derivations.
However, d-structure is a theory internal level (see Chomsky, 1993) and should
be avoided if possible. If DS-centered accounts of control are suspect given
standard minimalist assumptions then this casts doubt on GB approaches to
control given their reliance on d-structure requirements for their explanatory
punch.
There are other more technical problems with the GB approach to control,
some of which we review below. However, even this brief discussion hope-
fully indicates that control theory as earlier conceived fits ill with the main
methodological emphases of the Minimalist Program. The question is how
radical a departure from traditional approaches is warranted empirically and
is desirable theoretically. Various answers to this question have been offered
and we review some below.
In sum, from a minimalist perspective, the standard approach to control
presents various difficulties and motivates a search for alternatives to the stand-
ard GB accounts. In what follows, I will assume that these problems are suffi-
cient reason to re-examine control phenomena.
1.3 Some basic properties of control
The theory of control must address two issues: (i) where PRO appears, and
(ii) how it is interpreted. Let us consider these in turn.
1.3.1 The distribution of PRO
PRO most conspicuously appears in the subject position of non-finite clauses.
(10a–e) are thus fully acceptable, while (10f–i) are not.
(10)a. John tried/hoped PRO to eat a bagel
b. John tried/preferred PRO eating a bagel

c. John thinks that [PRO eating/to eat a bagel] would be fun
d. John saw Mary before PRO leaving the party
e. John told Mary where PRO to eat a bagel
f. *John hoped that PRO eat a bagel
g. *John preferred for Bill to meet PRO
h. *John hated Bill meeting PRO
i. *John talked to PRO
What unifies the first four cases is that PRO sits in the subject position of a
non-finite clause, in contrast to the four examples that follow.
It is reasonable to assume that in (10f–i) PRO is governed by various lexical
heads. In (10f) it is the finite morpheme in Infl, in (10g,h) the verb meet and in
(10i) the preposition to. If one further assumes that non-finite clauses do not
contain lexical heads (at least of the relevant sort), then the presence of PRO
in (10a–e) correlates with the absence of (head) government of these Spec IP
positions.7
Thus, the distribution of PRO conforms to the descriptive general-
ization (11).
(11) PRO can only appear in ungoverned positions
Note that (11) does not prohibit generating PRO in a governed position.
This is fortunate as PRO can be base generated in object position so long as it
moves to an ungoverned position by s-structure (SS). (11) is a generalization
that holds at s-structure or later.
(12) John1 tried [PRO1 to be recognized t1]
As is well known, there were several attempts within GB to reduce (11) to
more basic principles. This is not the place to review these efforts.8
Suffice it to
say, that these analyses do not currently enjoy much support. In section 1.4,
we review two kinds of minimalist approaches to the distribution of PRO.
1.3.2 The interpretation of PRO
Control structures come in two varieties: local and long distance.
(13)a. John hopes [PRO to eat a bagel]
b. John hopes that [ [PRO eating a bagel] will be fun]
In the earliest treatments, two distinct rules were involved in the derivations
of (13a,b). Equi NP deletion applied to yield (13a) while Super Equi was in-
volved in the derivation of (13b). Equi interacted closely with the Principle
of Minimal Distance in determining the antecedent of PRO in (13a) while
this principle did not regulate applications of Super Equi.9
GB has honored
essentially this same analysis by endorsing a distinction between structures of
Obligatory versus Non-obligatory control (OC vs. NOC).10

On Control 13
OC and NOC differ in several important ways. Consider the following para-
digm illustrating the interpretive properties of obligatory control structures.11
(14)a. *It was expected PRO to shave himself
b. *John thinks that it was expected PRO to shave himself
c. *John’s campaign expects PRO to shave himself
d. John expects PRO to win and Bill does too (= Bill win)
e. *Johni told Maryj PROi+j to leave together/each other
f. The unfortunate expects PRO to get a medal
g. Only Churchill remembers PRO giving the BST speech
(14a) shows that an obligatory control PRO requires an antecedent. (14b) indic-
ates that this antecedent must be local and (14c) indicates that it must c-
command the PRO.12
(14d) shows that this PRO cannot have split antecedents.13
PRO in (14f) only has the “de se” interpretation in that the unfortunate believes
of himself that he will be a medal recipient. (14g) has the paraphrase (15a), not
(15b). On this reading only Churchill could have this memory for Churchill was
the sole person to give the speech. The two different readings follow on the
assumption that obligatory control PRO must have a c-commanding anteced-
ent. This requires “only Churchill” to be the binder. The unavailable reading
has “Churchill” as the antecedent. This is possible in (15b) where the pronoun
can have a non-c-commanding antecedent.
(15)a. Only Churchill remembers himself giving the BST speech
b. Only Churchill remembers that he gave the BST speech
PRO in non-obligatory control environments contrasts in every respect with
the obligatory control cases.
(16)a. It was believed that PRO shaving was important
b. Johni thinks that it is believed that PROi shaving himself is important
c. Clinton’si campaign believes that PROi keeping his sex life under
control is necessary for electoral success
d. John thinks that PRO getting his résumé in order is crucial and Bill
does too
e. John1 told Mary2 that PRO1+2 leaving together/each other was import-
ant to Bill
f. The unfortunate believes that PRO getting a medal would be boring
g. Only Churchill remembers that PRO giving the BST speech was
momentous
(16a) indicates that non-obligatory control PRO does not require an anteced-
ent. (16b) demonstrates that if it does have an antecedent it need not be local.
(16c) shows that the antecedent need not c-command this PRO. (16d) contrasts
with (14d) in permitting a strict reading of the elided VP, that is the reading in
which it is John’s resume which is at issue. (16e) can support split antecedents,

(16f) can have a non “de se” interpretation and (16f) is consistent with many
people other than Churchill recalling that the BST speech was a big deal. Note
that each non-obligatory control reading contrasts with those available in the
obligatory control examples in (14). The cases in (14) and (16) contrast in one
further interesting way; the former can be paraphrased with PRO replaced by
a reflexive while the interpretive doubles of (16) replace PRO with pronouns.
(17) illustrates this with the counterparts of (14c) and (16c).
(17)a. *John’si campaign expects himselfi to shave himself
b. Clinton’si campaign believes that hisi keeping his sex life under con-
trol is crucial for electoral success
In short, the differences in obligatory and non-obligatory control structures
duplicate, where applicable, what one finds with locally bound anaphors versus
pronouns. This makes sense if PRO is actually ambiguous – an anaphoric ex-
pression in obligatory control configurations and pronominal in NOC structures.
The question facing the theoretician is to explain, first, why PRO should dis-
play these two sets of properties and why we find the OC properties in “Equi”
configurations while we find the NOC cluster in “Super Equi” structures. In
other words, why do the properties cluster as they do and why do they dis-
tribute as they do.
1.4 Two approaches to the distribution of PRO
1.4.1 Null case
Chomsky and Lasnik (1993) argue that the standard GB analysis of control in
terms of government (see (11) above) cannot be reconciled with the last resort
motivation for movement within Minimalism.14
The empirical difficulty is
illustrated in (18).
(18)a. We never [PRO1 expected to be found t1]
b. *We never expected [PRO1 to appear to t1 [that Bill left]]
If movement is last resort and PRO must be ungoverned then the threat of
being governed should suffice to license PRO’s movement in (18a). But if it
suffices in (18a) why is it insufficient in (18b)? Chomsky and Lasnik argue that
the two cases fall together if we assume that PRO has a case that must be
checked.15
To get this to work technically, they assume that PRO has its own
(idiosyncratic) case, dubbed “null case,” that can only be checked by the T0
of
non-finite control clauses. PRO (and PRO alone) can and must check null case.
Given these assumptions, (18b) violates greed in that PRO moves from one
case position to another thereby violating last resort.
This same set of assumptions suffices to account for the facts in (10) re-
peated here.

On Control 15
In (10a–e), it is assumed that PRO is in the Spec of a T0
able to check null case.
Hence their acceptability. In contrast to this, PRO is in a nominative case
position in (10f), and accusative position in (10g–i). These are not positions in
which null case can be checked and so the structures underlying these sen-
tences are ungrammatical.
The most fully worked out version of this null case approach to obligatory
control is Martin (1996).16
Following Stowell (1982), he argues that control
infinitives differ from raising infinitives in that the former have tensed Infls.
In effect, Martin provides motivation for the assumption that some infinitives
can check case by assimilating them to the class of clauses that uncontroversially
do so: finite clauses. Both finite IPs and control infinitives are +Tense and so it
is not surprising that their respective Spec IPs are case positions. Given the
further assumption that only PRO bears null case, only it can appear in the
Spec IP position of tensed infinitives.
As mentioned, Martin follows Stowell in observing that some embedded
clauses appear to have tense specifications. Stowell distinguishes between
control clauses and raising clauses in that the latter always share the tense
specifications of the clause they are embedded under while the former do not.
The examples in (19) illustrate the point.
(19)a. John decided/remembered [PRO to go to the party]
b. John believed [Mary to be the best player]
In (19a) the embedded event of going to the party takes place in the future
with respect to John’s decision (or recollection). In contrast, in (19b) Mary
being the best player temporally coincides with John’s belief. Stowell (1982)
accounts for this by assuming that the embedded clause in (19a) has a T0
specified for tense and so the embedded event can be temporally located
independently of the main clause event. In contrast, the embedded T0
of rais-
ing clauses is not specified for tense and so its temporal specification cannot be
independent of the one in the main clause. Martin assumes that it is the pres-
ence of a temporal operator in control clauses that allows them to check (null)
case.17
Martin (1996) provides further interesting evidence for the assumption
that null case is a property of +tensed non-finite clauses. Here is one more
bit. Martin claims that event denoting predicates cannot occur under raising
predicates.18

(20)a. *Everyone believed Rebecca to win the game right then
b. *The defendant seemed to the DA to conspire against the govern-
ment at that exact time
Following Enç (1990), Martin assumes that event denoting predicates con-
tain event-variables which must be bound by “tense or some other (temporal)
operator in order to denote an individuated event” (Martin, 1996, p. 59). With-
out such an operator only a stative predicate is possible. The absence of a
tensed T0
in raising predicates accounts for the oddity of the examples in (20).
With this in hand, Martin (following Boskovic, 1997) observes the following
contrast between Romance croire and its English counterpart believe. The latter
is a raising predicate with the exceptionally case marked subject of the embed-
ded clause residing in its Spec VP.19
For example, (21a) has the structure (21b).
(21)a. John believes Mary to be tall
b. John1 [VP Mary2 [VP t1 v [believes [t2 to be tall]]]]
The former, however, shows the characteristics of a control predicate, as the
following French example indicates.
(22) Je crois [PRO avoir fait une erreur]
I believe to have made a mistake
Interestingly, in Romance, the restriction to statives/habituals noted on em-
bedded clausal complements of believe does not hold for croire. Contrast (23a)
and (23b), its English counterpart.
(23)a. Jean croit rêver
John believes to dream
b. John believes himself to dream
(23a) has the interpretation John believes himself to be dreaming. Note that this
reading is unavailable for (23b). (23b) is unacceptable in English because believe
is an ECM verb and so cannot take a control complement. What the data here
appear to indicate are two things: first, that in French croire is not an ECM verb
but a control verb and second, that when it is a control verb it allows eventive
predicates. Martin concludes from this that control infinitives have tensed T0
s
and that raising predicates do not.
One last interesting bit of corroborating evidence for this comes from a
further observation. Boskovic observes the following interesting fact. There
are cases of apparent ECM – like uses of croire, as in (24).20
(24) Qui1 Ana croyait-elle [t1 plaire à Pierre]
“Who did Ana believe to please Pierre”
In these cases, moreover, the eventive reading noted in (23) is unavailable.
Only a habitual interpretation is available.

On Control 17
(25) *Qui crois-tu rêver
“Who do you believe to be dreaming”
In sum, Martin provides interesting evidence that null case correlates with the
presence of temporally active T0
s. In effect, he argues that null case can be
checked by a T0
if and only if the T0
is +tense and −finite. If one further
assumes that PRO can only bear null case, then a full account for the distribu-
tion of PRO follows.
1.4.2 Some problems for null case
Before discussing other approaches to the distribution of PRO, let me voice
some skepticism regarding the proposed correlation between tense/eventive
properties and the nature of T0
in raising and control structures. Consider first
the claim regarding eventive predicates under raising versus control verbs.
The noted unacceptability of eventive predicates under raising verbs is
most pronounced for the ECM predicate believe (see (20a)). There are perfectly
acceptable raising constructions with embedded eventive predicates. So con-
trast the examples in (26) with (27).
(26)a. Rebecca seemed to win the game right then
b. John appeared to take the wrong medicine
c. John is likely/certain/sure to eat a bagel
(27)a. *John believed Rebecca to win the game right then
b. * John showed Bill to take the wrong medicine
c. * John believed Bill to eat a bagel
This contrast suggests that the tense property of relevance relates to ECM
verbs, rather than to raising predicates in general. But if so, then the conclu-
sions Martin draws are too broad. The distinction of theoretical interest is
raising versus control complements. Given this broad contrast, the cases in
(26) and (27) should be entirely parallel.
Note, incidentally, that (26c) above has one other curious property. The
complement clause is most naturally interpreted as in the future with respect
to the tense of the matrix. Thus, it means, more or less, that it is currently
likely/certain/sure that John will eat a bagel. This temporal ordering is unex-
pected given Stowell’s claim that raising predicates do not contain indepen-
dent tense specification.
Martin (1996, pp. 80–105) observes similar facts and concludes from this that
such cases involve control rather than raising. In fact, if one uses these temporal
diagnostics, it appears that all the standard (non-ECM) raising predicates are
actually ambiguous, with both a raising and a control structure. This assump-
tion runs into problems. Consider how.
Assumethatwheneveroneseesaneventiveembeddedpredicatethenthe clause
has a control structure. The sentences in (28) have embedded eventive predicates.

(28)a. The shit appeared to hit the fan then
b. It seemed to start to rain exactly then
c. ?There appeared to enter several men at that very moment
(28a) means, roughly, it appeared that pandemonium erupted then. The presence of
the punctual adverb reinforces this point. Given this, by assumption these sen-
tences all havecontrol structures, withthe matrix subject controllingan embedded
PRO. The problem is that were this so, we would expect idioms and expletives to
be barred from the subject matrix position. (28) indicates that this is incorrect.
Observe that we also would expect to find no voice transparency in raising
constructions where the embedded predicate had an eventive interpretation.
This also appears to be incorrect.
(29)a. The doctor seemed to then examine Mary
b. Mary seemed to then be examined by the doctor
In both (29a,b) the embedded clause is interpreted eventively. Nonetheless,
voice transparency seems to hold. Recall that the absence of voice transparency
and the limited distribution of expletives and idioms are the classical hall-
marks of control. The absence of these thematic diagnostics in (28) and (29)
suggests that they are standard raising, not control, constructions. If so, it can
be argued that tense diagnostics do not correlate well with the standard them-
atic diagnostics that distinguish raising from control and this, in turn, argues
against seeing some T0
s as assigners of null case.21
The data that Martin provides (following Boskovic, 1997) is also restricted to
a contrast between ECM predicates in Romance and English. It does not con-
trast non-ECM raising predicates (such as seem) with control predicates. This is
noteworthy for it is the latter contrast that is of obvious theoretical interest, not
the former.
It is unclear what makes ECM predicates exceptionally able to assign case.
However, it would not be remarkable if ECM verbs imposed semantic restric-
tions on their complements, selecting, for example, non-eventive clausal com-
plements. Were this the case, then one would expect the correlations noted by
Boskovic even in the absence of case assignment in the embedded clause in control
complements. In other words, if Boskovic’s observations pertain exclusively to
ECM verbs then their peculiarities tell us nothing about the thematic contrasts
between control and raising verbs. It is only if the properties that Boskovic
noted are properties of raising predicates in general that his observations sug-
gest that control predicates are null case markers.
A few last points. There are ECM verbs that do not display the diagnostics
noted in the text. For example, expect takes eventive predicates and requires a
temporal specification of the embedded event later than the matrix.
(30) John expected Mary to leave the party
This implies that in such a case the embedded clause has a +tense T0
. This then
further implies either that (i) the structure of (30) is similar to what one finds

On Control 19
with a persuade verb or (ii) that Mary is case marked by some element in C0
as
is the case, arguably, for want.
There are problems for assumption (ii). Were this so, we would expect
passivization to be blocked and for to appear overtly (in some configurations)
parallel to what we witness with want. Both expectations are unrealized.
(31)a. *John was wanted to leave
b. John was expected to leave
c. John wants very much for Bill to leave
d. *John expects strongly for Bill to leave
This leaves the (i)-option that what we have is a persuade-like configuration.
However, this suggestion also has its problems. Recall the standard thematic
differences between persuade and expect verbs. These are not suspended when
we find the embedded clause with an eventive/future interpretation. For ex-
ample, we find idioms and expletives with eventive/future readings (32a,b)
and we find voice transparency (32c,d) are paraphrases.
(32)a. John expected the shit to hit the fan at exactly 6
b. John expected there to erupt a riot
c. John expected the doctor to examine Mary then
d. John expected Mary to be examined by the doctor then
This is unexpected if these are control structures.
Finally, consider one last set of empirical difficulties.22
Stowell (1982)
observes that gerunds, in contrast to infinitives, vary their tense specifications
according to the properties of the matrix verb.
(33)a. Jenny remembered [PRO bringing the wine] (Stowell, 1982, (8b))
(J remembered a past event of bringing the wine)
b. Jim (yesterday) counted on [PRO watching a new movie (tonight)]
(J counted on a future event of movie watching)
In light of this, Stowell (1982) proposed that gerunds are generally marked
−tense.
If correct, this constitutes a problem for the view that only +tense T0
s can
assign null case and thereby license PRO given the presence of PRO in (33a,b).23
In fact, other gerunds provide potentially more serious problems. Pires (2001)
identifies a class of TP defective gerunds which with respect to their tense
properties are very similar to ECM infinitives in that their time interval must
coincide with the event time of the matrix.
(34) *Bill last night avoided [PRO driving on the freeway this morning]
In such cases the argument against postulating a +tensed T0
on Stowell-like
grounds is stronger still.

I have briefly reviewed some problems for the null case theory as proposed
by Chomsky and Lasnik (1993) and most fully elaborated and developed in
Martin (1996). Before moving on to consider another approach to the distribu-
tion of PRO (one that I personally prefer, I must confess), it is worth recalling
the virtues of the null case account. It directly addresses one of the main
characteristics of control phenomena; namely that it is subjects that are con-
trolled. Once one gives up government as a core concept of grammar (and
with it the generalization (11) that PRO only occurs in ungoverned positions)
it becomes a challenge to identify the subject position of non-finite clauses in
any sort of unified principled manner. Chomsky and Lasnik (1993) provide
one way of doing this: via a case marking property unique to non-finite clauses.
Martin’s (1996) suggestion that null case is a property of some +tense T0
s is a
reasonable attempt to provide a principled foundation for this idea and thereby
account for the limited distribution of PRO.
1.4.3 (OC) PROs as traces
Consider now a second approach to the distribution of PRO.24
It starts from
one observation and one methodological qualm. The observation is that con-
trol, like raising, affects subjects of non-finite clauses. The qualm is that the
theory of null case is stipulative and, hence, methodologically suspect. In fact,
null case singles out PRO for special treatment in two distinct ways. First, it is
the only lexical item able to check it or bear it. No other DP can bear null case,
not even phonetically null expressions like WH-traces.25
(35)a. John asked Bill *Mary/PRO to eat a bagel
b. *The man1 (who) John asked Bill t1 to eat a bagel
(35a) with Mary as the embedded subject is ungrammatical as Mary cannot
check nor carry null case. PRO is fine in this position because it can. Note that
(35b) is also unacceptable, presumably because t1 cannot check/carry null case
either. Thus, the only phonetically non-overt element able to bear null case is
PRO.
Second, to my knowledge, it is the only lexical item whose case properties
are grammatically specified. There are no other DPs, to my knowledge, that
can bear but a single case.
Null case does not pattern with other structural cases in other ways.
For example, it contrasts with WH-traces in not blocking wanna contraction.
(36a) cannot be contracted as in (36c) while contraction for (36b) as in (36d) is
fine.26
(36)a. Who1 do you t1 want to visit Mary
b. I1 want PRO1 to visit Mary
c. *Who do you wanna visit Mary
d. I wanna visit Mary

On Control 21
It thus seems that an accusative case marked non-phonetic WH-trace suffices
to block sandhi effects like wanna contraction while a null case marked non-
phonetic PRO does not.
In sum, whatever the other virtues null case may have, its peculiar char-
acteristics are evident. With this as prologomena, consider a movement based
approach to control.27
The core of such a theory is that OC PRO is identical to an NP-trace. It is the
residue of overt A-movement. Thus, the overt structure of (37a) is (37b).28
(37)a. John tried to win
b. [TP John [past [VP (John) [VP try [(John) to [(John) win]]]]]]
Note that this is entirely analogous to what one finds in raising constructions
with one caveat. In a raising construction, movement is from the embedded
clause to a matrix non-theta position while in control structures, movement
is via a matrix theta position. Note the trace in the spec of the matrix VP in
(37b).
This proposal accounts for the distribution of PRO as follows. Assume that
the subject positions of all non-finite clauses are not case marking positions
(clearly the null hypothesis given the unacceptability of cases like (38a)). Then
A-movement from this position is permitted. Further, as A-movement from
case positions is prohibited (see 38b) and if (OC) PRO is the residue of A-
movement, then we should never find PROs in case positions (see (10f–i),
repeated here).
(38)a. *John hopes [Frank to leave]
b. *John seems [t is nice]
(10f–i) are underivable as they all involve movement from a case position.
(10a–e) are all acceptable as each involves movement from a non-finite clause
via a matrix theta position ending in a matrix case position. If we assume that
non-finite clauses do not assign case to their subjects, then movement from the
embedded IP to the matrix IP will not be prohibited by last resort (i.e. Greed)
as it is in (10f–i).
In addition, note that the null phonetic value of PRO (here just an NP-t)
correlates with the fact that traces of A-movement are all phonetically null.29
In other words, whatever it is that renders copies due to A-movement

phonetically null will extend to explain the null status of PRO. We need invoke
no special null case properties and need invoke no special lexical item that is
only able to carry such a case. This grouping of PRO with NP-t also accounts
for why both behave similarly in wanna contraction contexts.
(39)a. I seemta (seem+to) eat bagels every morning
b. I wanna (want+to) eat bagels every morning
Thus, the fact that (OC) PRO is typically a subject of non-finite clauses,
follows from the basic premise that it is the residue of overt A-movement. This
is merely a descriptive generalization. The proposal here is that control, like
raising, is due to movement motivated by case concerns. It is not necessarily
restricted to non-finite subject positions. If some non-subject position is not a
case position, it can also be occupied by PRO. Lasnik (1995a,b), following a
proposal by Munn, suggests that examples like (i) involve control.
(i) John washed/shaved/dressed
Note that these do have the expected properties of control verbs. Thus, for
example, John is related to two theta-roles in the case of these reflexive predic-
ates. If one assumes that such verbs need not assign case to their objects, then
one can treat these as cases of A-movement (and so, control). Note that were
this a case of control, it would argue against the null case theory as here we
would have a PRO unrelated to a T0
.
One other point. Were this a case of control, then we would need to explain
why it is that sentences like (ii) are unacceptable.
(ii) John1 hoped Bill washed PRO1
This movement would not be blocked by case as the object position of wash
need not be case marked. However, minimality (shortest Move) would be
violated by moving over the intervening DP Bill. (40) lists the assumptions
required to permit the sort of movement advocated here.
(40)a. Theta roles are features
b. There is no upper bound on the number of theta features that a DP
can have
c. Movement is Greedy
d. Greed is understood as “enlightened self interest”
Of these four assumptions, (40a,b) are the most contentious. They amount to
rejection of the two basic ideas behind the standard theory of control: (i) that
DS (or any analog) regulates the possibilities of movement and (ii) that there is
any requirement limiting DPs to at most one theta-role. Both assumptions
seem natural enough in a minimalist setting. Consider why.
Chomsky (1993) argues that grammar internal (non-interface) levels have no
place in an optimal theory of grammar. DS, he observes, is such a level. As

On Control 23
such, it is methodologically suspect and should be eliminated. Chomsky (1993)
provides some empirical arguments to this same conclusion. Assume that these
considerations are decisive. Then one might expect the restrictions that DS
placed on grammatical operations to also disappear. Recall, that it was the
thematic restrictions that DS imposes that forced the distinction between rais-
ing and control in earlier theory. Without DS, however, we might expect the
two types of operations to fall together, as proposed in a movement approach
to control. Put more baldly, a movement theory of control is a natural con-
sequence of the minimalist elimination of DS.
This conclusion has been resisted. Chomsky’s proposed elimination of
DS does not entail that the thematic restrictions found in earlier DS based
theories play no role in the grammar. Chomsky (1995) achieves the func-
tional equivalent of DS by assuming that theta-roles are not features. Coupled
with the idea that all legitimate movement must check a feature, that is be
greedy, the claim that theta-roles are not features results in a system where
control cannot be reduced to movement as movement into theta positions
is unmotivated and so blocked by Greed. (40a) amounts to a rejection of this
way of re-introducing DS conditions once DS is abandoned. The assumption
that theta-roles are features is aimed at allowing movement between theta
positions.30
(40b) has a similar motivation. The prohibition against a DP’s bearing at
most one theta-role has no conceptual justification (in contrast to the idea that
a DP bear at least one, which plausibly follows from a principle of Full Inter-
pretation). The restriction, however, would follow were movement into theta
positions prohibited. Given that control always involves the relation of a DP to
at least two theta positions, a movement approach to control must counten-
ance a DP’s having more than one theta-role.
(40c,d) are standard minimalist assumptions. I here adopt the enlightened
self-interest interpretion of Greed as I take theta roles to be features of pre-
dicates that DPs obtain by merging with them (via either pure merge or
copy plus merge viz. Move). Theta features, then, are primarily properties of
predicates and only derivationally properties of arguments. If so, checking a
feature of one’s target must suffice to license movement. Greed as enlightened
self-interest permits this.
Given these four assumptions, then, a movement theory of control is viable.
I have argued that they are conceptually natural (if pressed, I might say optimal)
given a minimalist setting. I now want to consider empirical support for (40a,b)
that arises outside the domain of control phenomena.
1.4.4 Movement into theta positions
If control is movement then movement into theta positions must be possible.
There is independent evidence that such movement obtains. I here review two
cases in the literature.
Boskovic (1994, pp. 268 ff.) discusses the following examples from Chilean
Spanish.31

(41)a. Marta le quiere gustar a Juan
Martha clitic wants to please to Juan
“Marta wants Juan to like her”
b. A Juan le quiere gustar Marta
“Juan wants to like Marta”
Three facts regarding examples like (41) are of present interest.
First, a case marks Juan. That Juan bears this case is a lexical property of the
verbgustarwhich assigns a quirky/inherentcase to itsthematicexternal argument
much as Belletti and Rizzi (1988) argue with respect to similar verbs in Italian.
Second, (41a) differs from (41b), as the glosses indicate, in that Marta has
two thematic functions in the first sentence while Juan has two in the second.
This suggests that both are control structures with the matrix subject control-
ling the external argument in the embedded clause. Third, note that in (41b)
the matrix subject bears the inherent case marker a. Given that a is an inherent
case assigned by the embedded verb gustar, this implies that a Juan has moved
to the matrix subject position in (41b) from the embedded position in which a
and its corresponding theta-role were assigned to Juan. Such movement is also
observed in raising constructions with inherent case marked subjects (al is the
case marker).32
(42) Al professor le empezaron a gustar los estudiantes
“The professor began to like the students”
These three facts together implicate movement from one theta position to
another. The a case on Juan in (41b) indicates movement from the domain of
gustar, whence the inherent case marking was effected. However, in contrast
to (42), quiere has an external thematic argument to assign and in (41b) Juan
clearly bears it as the gloss indicates. The obvious implication is that Juan has
moved from the thematic position of gustar to a thematic position of quiere and
thereby obtained a second theta-role, as Boskovic (1994) concludes.33
Boskovic and Takahashi (1998) provides a second argument that movement
into theta positions is possible. The phenomenon analyzed is long distance
scrambling in Japanese.
As Boskovic and Takahashi note, long-distance scrambling – (43b) – is odd
given minimalist assumptions that it seems unmotivated. The base form from
which it is derived is well formed (43a).
(43)a. John-ga [Mary-ga sono hon-o katta to] ommotteiru (koto)
-nom -nom that book-acc bought that thinks fact
“John thinks that Mary bought that book”
b. Sono hon-o1 John-ga [Mary-ga t1 katta to] ommotteiru (koto)
“That book, John thinks that Mary bought”
The lack of semantic motivation for the movement is supported by the fact that
unlike Wh-movement and topicalization, scrambling does not seem to establish

On Control 25
an operator-variable relation at LF.34
In particular, it seems that the long scram-
bling must be undone at LF. (44) illustrates this.35
Boskovic and Takahashi note
that a (long) scrambled QP cannot take scope over a matrix QP subject.
(44) Daremo1-ni dareka-ga [Mary-ga t1 atta to] omotteiru (koto)
everyone-dat someone-nom Mary-nom met that thinks (fact)
“Everyone someone thinks that Mary met”
(44) must be interpreted with dareka (someone) scoping over daremo (everyone)
despite the fact that the latter c-commands the former in overt syntax. This
follows if the scrambled daremo must reconstruct at LF.
Where does the scrambled expression move to at LF? According to Boskovic
and Takahashi, it moves to its theta position. This movement is obligatory for
were it not to so move, it would violate full interpretation. In fact, as Boskovic
and Takahashi argue, virtually all of the characteristics commonly associated
with long scrambling follow if it is assumed that the scrambled expression is
Merged in overt syntax in its scrambled position and then moved covertly to
its theta position at LF. However, for this analysis to be viable, it must be
possible to move to a theta position and this movement must have some sort of
feature checking motivation given standard assumptions. Both desiderata are
met if it is assumed that theta roles are sufficiently like features to license
greedy movement, assumption (40a) above.36
To sum up, there is independent evidence for movement into theta posi-
tions. The existence of this movement indirectly enhances the proposal that
control is the overt manifestation of movement via multiple theta positions.
Why, after all, should UG have recourse to a specialized theory internal ex-
pression like PRO if movement can independently forge the requisite rela-
tions? The redundancy inherent in such an approach speaks against its adoption
on purely methodological grounds.37
1.4.5 Conclusion
This section has discussed the two dominant current minimalist approaches to
the distribution of PRO. Both attempt to respond to the fact that minimalist
assumptions leave little room for the standard GB approach to this issue. The
basic descriptive generalization is that PROs are subjects and are found in
non-finite sentences.38
How to account for this without theoretically alluding
to government is the basic challenge that both approaches try to meet.
1.5 The interpretation of PRO
There are two broad approaches to the interpretive properties of control in the
current literature. The first is a structural approach. Abstracting from many
interesting differences, structuralists generally see the interpretive properties

of control as reflecting structural facts concerning the syntax of control con-
figurations.39
These properties, then, are analyzed as reflecting the internal
workings of the grammar. The second approach is more eclectic. Eclectics see
the interpretive properties of control as being due to the complex interaction
of various modules, some of which are external to the syntax properly speak-
ing. Chomsky sums up the view well:
the theory of control involves a number of different factors: structural configura-
tions, intrinsic properties of verbs, other semantic and pragmatic considerations.
(Chomsky, 1981, pp. 78–9)
This section concentrates on the first approach. The next reviews objections
lodged against it from the more eclectically inclined. The main reason for
this is methodological; all things being equal, structuralism is preferable to
eclecticism. The reason is that all agree that grammatical structure is part of
any adequate approach to control. What distinguishes structuralists from
eclectics is whether this information exhausts what is needed. All things being
equal then, structuralism is preferable if attainable. The main arguments in
favor of eclecticism are the purported empirical inadequacies of structuralism.
As such, starting from the former position and considering possible empirical
problems that beset it should reveal the virtues and vices of both views.
As noted in section 1.3.2, not all PROs are interpreted in the same way. A
contrast exists between PROs in Obligatory Control configurations (OC PRO)
and those in non-obligatory structures (NOC PRO). The former have proper-
ties roughly equivalent to those of reflexives while the latter pattern inter-
pretively largely like pronouns.40
For illustration consider the two paradigms
first outlined in section 1.3 and repeated here.
Recall that (14) and (16) differ completely. The OC PROs in (14) require
local, c-commanding antecedents, require sloppy readings under ellipsis, forbid
split antecedents, only have de se readings and require bound readings with
only antecedents. The latter need no antecedents, need not be c-commanded
locally by an antecedent if one is present, allow a strict reading under ellipsis,
permit split antecedents, can have a non de se reading and can have a referential
reading with only antecedents.
(14)a. *It was expected PRO to shave himself
b. *John thinks that it was expected PRO to shave himself
c. *John’s campaign expects PRO to shave himself
d. John expects PRO to win and Bill does too (= Bill win)
e. *Johni told Maryj PROi+j to leave together/each other
f. The unfortunate expects PRO to get a medal
g. Only Churchill remembers PRO giving the BST speech
(16)a. It was believed that PRO shaving was important
b. Johni thinks that it is believed that PROi shaving himself is important
c. Clinton’si campaign believes that PROi keeping his sex life under
control is necessary for electoral success

On Control 27
d. John thinks that PRO getting his resumé in order is crucial and Bill
does too
e. John1 told Mary2 that PRO1+2 leaving together/each other was
important to Bill
f. The unfortunate believes that PRO getting a medal would be boring
g. Only Churchill remembers that PRO giving the BST speech was
momentous
There is an important fact worth bearing in mind in what follows. The data
illustrated in (14) and (16) illustrate that OC PRO has a proper subset of the
properties of NOC PRO. Thus, (16d) differs from (14d) in allowing a strict
reading which the latter forbids. (16d) can support a sloppy reading. This
holds for (most of) the other properties as well. This means that OC PRO and
NOC PRO cannot be in free variation. In particular, positions inhabited by OC
PRO must exclude NOC PRO. Were this not so, we would never be able to
observe OC PRO as its signature properties would never emerge, being, as
they are, a proper subset of NOC PRO. What this means theoretically is
that we need some way of excluding NOC PRO when OC PRO is present.
This has implications for any general theory of control and we return to it
at the end.
1.5.1 Some properties of a structural theory
Assume for the nonce that OC PRO is the residue of NP movement. If so, the
structure of a subject control verb is roughly as (45a) and an object control
verb looks like (45b).
(45)a. DP1 V [IP t1 [I −finite] [VP t1 V. . . . ]41
b. DP1 V DP2 [IP t2 [−finite] [VP t2 V. . . . ] 42
Several features of these configurations are of interest in relation to the para-
digm in (14).
First, the fact that OC configurations require local c-commanding anteced-
ents follows directly from their being residues of movement. In fact, it is the
same as the reason why A-traces require local, c-commanding, antecedents.
Second, the facts concerning sloppy readings under ellipsis, the prohibition
against split antecedents, the required de se reading for (14f) and the bound
reading in (14g) also follow given a movement analysis. The required sloppy
reading parallels what one finds in raising constructions.43
(46) Bill was expected t to win and Harry was too
Split antecedents are barred on a movement analysis. For an OC PRO to have
DP as its antecedent implies that PRO is the trace left by the movement of DP

from PRO’s position. Split antecedents would require having two distinct DPs
move from the very same position. This contemplated derivation is theoretic-
ally impossible and so split antecedents are ruled out.
It is worth noting that there is nothing that would rule out the possibility of
OC PRO taking split antecedents were PRO a base generated expression any
more than there is an explanation for why, given standard assumptions, loc-
ally bound reflexives cannot have split antecedents. Rather, the impossibility
of split antecedents for reflexives and OC PRO is a stipulation motivated
entirely on empirical grounds. The fact is tracked by an axiom explicitly for-
bidding split antecedents for reflexives and controlled PROs.44
Within a theory of control in which antecedents to PRO are functions of
idiosyncratic lexical features of particular verbs the problem is perhaps worse.
Why do control predicates designate only a single antecedent for PRO in OC
configurations? Why not two (or more) antecedents? Were this possible then
split antecedence would follow naturally. Thus, the impossibility of split ante-
cedents and the fact that control predicates “choose” but a single controller are
closely related facts. Both are immediately accounted for on a movement theory
of OC.
The de se interpretation found in (14f) also follows. As a result of move-
ment via multiple theta positions one ends up with a chain with multiple
theta-roles. The natural semantic interpretation of such a syntactic object is
of one expression, a chain, “saturating” several distinct argument positions.
Semantically, this yields a complex monadic predicate, roughly of the form
(47).
(47) DP[λx (Px,x)]
Such complex monadic predicates contrast with configurations of binding in
forcing de se readings.45
Thus, as movement via multiple thematic positions
leads to the formation of complex monadic predicates and given that these
only permit de se readings, it follows that if OC PRO is a residue of move-
ment that it will require a de se interpretation. It is, once again, worth noting
that this interpretive fact, though consistent with a base generation of a PRO
approach, does not follow from it.
The facts concerning (14g) also follow. On the movement theory, the PRO is
actually the residue of overt movement and so must have a c-commanding
antecedent. Only Churchill is a possible antecedent. Churchill alone is not. Thus,
the relevant structure must be one in which only Churchill is the antecedent of
PRO. The logical form of (14g) is (48).
(48) Only Churchill λx (x remembers x giving the BST speech)
In sum, if OC is formed via movement, then the interpretive properties
illustrated in (14) follow rather directly.
One technical point is worthy of note before considering further facts friendly
to the movement analysis. The above account relies on there being an LF chain
relating controller to OC PRO at LF. This structural property is compatible

On Control 29
with many different views of movement. Thus, Hornstein (2000), Manzini and
Roussou (2000), Martin (1996), and O’Neill (1995) implement a movement
analysis in rather different ways. Nonetheless, one common feature of all four
is that at LF PRO and its controller are links of a single common chain. This
suffices to derive the facts above.46
1.5.2 The Minimal Distance Principle
There are other characteristics of OC PRO constructions that follow naturally
if movement is involved. Rosenbaum (1967, 1970) observed that the controller/
PRO relation generally obeys the Minimal Distance Principle (MDP). Thus
examples like (49) must be controlled by the object, not the subject.
(49) John1 persuaded Mary2 PRO*1/2 to go home
The vast majority of control structures with matrix transitive verbs like persuade
require object control.47
Furthermore, when verbs can optionally alter their argument structures
Rosenbaum (1967) noted shift in the potential controllers.
(50)a. John1 asked/begged/got Mary2 PRO*1/2 to leave
b. John1 asked/begged/got PRO1 to leave
In (50a), John, the subject, cannot control PRO. In (50b), when the object is
not generated (at least in overt syntax), the subject can be (and must be) the
controller.
These sorts of cases can be accounted for in terms of the MDP. Assume for
the moment that the MDP is a descriptively adequate generalization. Why
does it hold? Note that it follows on a movement theory of control if one
assumes that movement is governed by minimality, a standard assumption.
To see this, consider what the derivation of (49) would have to be like were
John the antecedent of PRO.
(51)a. John [VP John persuaded Mary [IP John to [John go home]]]
The copies of John mark the history of derivation. Note that in moving from
the embedded Spec IP to the matrix Spec VP John crosses the intervening DP
Mary. This move violates minimality and is so barred. The only derivation not
prohibited by minimality is one in which the DP in Spec IP raises to the next
highest potential DP position, in this case the object. The derivation is illus-
trated in (51b).
(51)b. John [VP John persuaded Mary [IP Mary to [Mary go home]]]
So, if OC PRO is the residue of A-movement, the MDP follows.

1.5.3 Adjunct control and movement
Rosenbaum (1970) extends the MDP to cases of adjunct control.
(52) John saw Mary after/before/while PRO eating a bagel
In (52) only John can control the PRO in adjunct position. Control from the
object position is forbidden. This falls under the MDP on the assumption that
John but not Mary c-commands the adjunct.
Unfortunately, this assumption is not obviously correct. For example, it is
possible for objects to bind pronouns found within adjuncts, as Orson Wells
taught us.
(53) John will drink no wine1 before it1 is ready for drinking
If we assume that to be interpreted as a bound variable a pronoun must be c-
commanded by its antecedent, this implies that it in (53) is c-commanded by
no wine at least at LF. But then objects should be able to control into adjuncts,
contrary to fact.
Note that one might get around this difficulty if one assumed that the gen-
eralization described by the MDP holds for overt syntax. Were this so, then
one could argue that at LF, the object might c-command the adjunct but in
overt syntax it does not. The problem, given a minimalist grammatical archi-
tecture, is how to state this generalization. Hornstein (2000) offers a solution
along the following lines.48
Minimalism has generally taken movement to be a complex operation made
up of two simpler ones, copy and merge. Assume that this is so. Nunes (1995)
observes that if this is correct, then it should be possible to copy an expression
from one sub-tree and merge it into another.49
Call this sequence of operations
“sidewards movement.” It is possible to analyze adjunct control as involving
this sort of operation. In particular let us assume the following:50
(54)a. Sidewards movement is a species of movement and it is possible
b. Adjuncts headed by after, before, while, etc. are adjoined to VP (or
higher)
(54a) makes two claims: that sidewards movement is a grammatically viable
operation and that it falls under the same general restrictions that govern more
conventional kinds of movement. (54b) says where the adjuncts of interest are
merged. It assumes, contra Larson (1988) for example, that the adjuncts in (52)
are basically adjoined to VP and are not generated within the VP shell. Assum-
ing (54), consider a derivation of adjunct control structure like (55).
(55) John saw Mary after PRO eating lunch
The derivation proceeds as follows. First construct the adjunct; merge eating
and lunch then merge John to eating lunch then merge after. This yields (56).

On Control 31
(56) after [John eating lunch]
Next construct the matrix. Merge saw and Mary. This yields a derivation with
two unmerged sub-trees as in (57).
(57) [saw Mary] [after [John eating lunch]]
At this point, we have exhausted the numeration but the external argument
position of saw is still unfilled. To fill it we copy John and merge it into the
Spec VP position, as in (58).
(58) [John [saw Mary]] [after [John eating lunch]]
We then merge the two sub-trees and finish the derivation in the conventional
manner yielding (59).51
(59) [John T0
[[John [saw Mary]] [after [John eating lunch]]]]
There are several things to note about this derivation. First, John’s sidewards
movement is what allows its case to be checked. Were it not to move from the
adjunct, the derivation would crash as John would have an unchecked case.
Second, the movement only occurs after Mary has merged into the comple-
ment of saw. Why could John not move into this position? Were it to do so then
we could have object control so we need to discover what prevents this. I
assume that economy restricts sidewards movement and so one cannot move
if the derivation can proceed without movement. In this case, we can merge
Mary into this position so movement is blocked. This then forces John to move
only after Mary has merged into the object position. This is why the controller
of a PRO inside an adjunct cannot be an object.52
Thus, the spirit of the MDP can be captured in cases of adjunct control as
well, if we see movement as the composite of Copy and Merge and thereby
permit sidewards movement.
This is actually a very positive result. The reason is that adjunct control
displays all of the diagnostic properties of OC PRO.
PRO headed adjuncts require local, c-commanding antecedents.
(60)a. *Johni said [that Mary left after PROi dressing himself]53
b. *John’si picture appeared after PROi shaving himself
c. *[that Bill left] seemed true before PROarb noticing
The PRO in these adjuncts do not tolerate split antecedents:
(61)a. *Johni said that Maryj left after PROi+j washing themselves
b. *Johni told Maryj a story after PROi+j washing themselves
PRO headed adjuncts only have sloppy readings under ellipsis.

(62) John left before PRO singing and Bill did too
Thus, (62) only has the reading paraphrased in (63a). It cannot be understood
as (63b).
(63)a. . . . and Bill left before Bill sang
b. . . . and Bill left before John sang
In “Churchill” sentences – (64a) – they cannot take “Churchill” as antecedent.
In other words, (64b) is not an adequate paraphrase of (64a).
(64)a. Only Churchill left after PRO giving the speech
b. Only Churchill left after Churchill gave the speech
And within adjuncts of the appropriate type (e.g. purposives), they display the
obligatory de se interpretation.54
(65) The unfortunate sent his report (in order to) PRO to get a medal
All in all then, cases of adjunct control display both MDP effects as well as the
standard interpretive properties associated with OC structures. This is what
we expect if they are indeed formed by movement.
It is quite interesting that cases of adjunct control can be analyzed in terms
of movement. The reason is that such control is unlikely to be reducible to the
thematic requirements of an embedding control predicate. In other words, the
control one finds in such cases is not a function of the properties of the matrix
predicate. Hence the controller cannot be some designated argument of the
control predicate as has been proposed for cases of complement control. None-
theless, these cases of control display all the diagnostic properties of OC. This
suggests that, in at least some cases of obligatory control, the controller is
structurally specified. Moreover, if the properties of OC PROs within adjuncts
are structurally determined then it would be odd to treat cases of complement
control (where the very same properties appear) as derived by entirely differ-
ent operations and in entirely different ways.
This section has shown how movement accounts for some central features
of obligatory control configurations. The account covers standard cases of com-
plement control as well as adjunct control. The general characteristics of OC
PRO make sense once OC PRO is seen as the residue of overt A-movement. In
the next section, we consider some possible empirical problems for this sort of
approach.
1.6 Problems for movement
Several authors have noted empirical difficulties for the movement analysis of
control. This section reviews empirical challenges to the approach.

On Control 33
1.6.1 Promise: and markedness
One prominent argument rests on denying the general validity of Rosenbaum’s
MDP.55
The form of the argument is as follows: movement is subject to min-
imality. Minimality is an inviolable condition on movement. Thus, we should
find no cases of subject control in transitive control predicates. However, it is
well known that there are a class of predicates that appear to violate the MDP,
for example verbs like promise require subject control.
(66) John1 promised Mary2 PRO1/*2 to leave
Thus, a movement based approach to control cannot be correct as it is incom-
patible with violations of minimality.
There is something curious about this argument. It appears to concede
Rosenbaum’s observation that in general one finds object control in transitive
control predicates. However, the existence of a semantically coherent class of
exceptions (the promise class which includes vow and commit among others), is
taken to indicate that the reduction of the MDP to minimality cannot be cor-
rect. Here is Landau’s version of the argument.
Hornstein (1999) is aware of the exceptions to the MDP, but rather than attribut-
ing them to some hidden (double object) structure, he proposes to view the MDP
as [a, sic] markedness condition. (Landau, 1999, pp. 231–2)
The problem is that the MDP is not a primitive in Hornstein’s system:
“the MDP reduces to the MLC [Minimal Link Condition/i.e. Minimality
NH].” This reduction is taken to be a strong argument in favor of the whole
approach. But then any properties of the MDP should follow from properties
of the MLC. In particular, if the MDP is a markedness condition, so should the
MLC be. This is clearly not the case, however; the MLC is exceptionless, and
its violations are sharply ungrammatical, whereas the “marked” violations of
the MDP (i.e. subject control) are perfect.
There are several replies that one can make to this argument. First, let us be
clear what it would mean to say that the minimality should be understood as
a “markedness condition.” One way to interpret this is that verbs that fail to
respect minimality are marked in the sense that environments that appear to
allow it will necessarily be hard to acquire. What “hard” means is that its
acquisition will be data-driven with the language acquisition device requiring
considerable evidence before it abandons the view that the witnessed opera-
tions are actually permitted in the observed environment. That is all that
“marked” means.56
In this sense, then, a movement theory treats control verbs
like promise as marked in that they should be harder to acquire than the nor-
mal run of transitive control verbs like persuade. As Hornstein (1999) observed
(and Landau, 1999 recognizes), there is acquisition data from C. Chomsky
(1969) indicating that it is indeed the case that promise predicates are acquired
late.57

Exploring the Variety of Random
Documents with Different Content

V. Persons who tell an artist that he is equal to Claude, or a writer
that he is as great as Bacon, do not add to the satisfaction of their
hearers, but pay themselves a left-handed compliment, by supposing
that their judgment is equivalent to the suffrage of posterity.
VI. A French artist advised young beginners against being too fond
of a variety of colours, which might do very well on a smaller scale,
but when they came to paint a large picture they would find they had
soon lavished all their resources. So superficial writers may deck out
their barren round of common-places in the finest phrases
imaginable; but those who are accustomed to work out a subject by
dint of study, must not use up their whole stock of eloquence at once,
they must bring forward their most appropriate expressions as they
approach nearer to the truth, and raise their style with their
thoughts. A good general keeps his reserve, the élite of his troops, to
charge at the critical moment.
VII. ‘Procrastination is the thief of time.’ It is singular that we are
so often loth to begin what gives us great satisfaction in the progress,
and what, after we have once begun it, we are as loth to leave off. The
reason is, that the imagination is not excited till the first step is taken
or the first blow is struck. Before we begin a certain task, we have
little notion how we shall set about it, or how we shall proceed: it is
like attempting something of which we have no knowledge, and
which we feel we are incapable of doing. It is no wonder, therefore,
that a strong repugnance accompanies this seeming inaptitude: it is
having to make bricks without straw. But after the first effort is over,
and we have turned our minds to the subject, one thing suggests
another, our ideas pour in faster than we can use them, and we
launch into the stream which bears us on with ease and pleasure to
ourselves. The painter who did not like to mix his colours or begin on
a new canvas in the morning, sees the light close in upon him with
unwilling eyes; and the essayist, though gravelled for a thought, or at
a loss for words at the outset of his labours, winds up with alacrity
and spirit.
VIII. Conversation is like a game at tennis, or any other game of
skill. A person shines in one company who makes no figure in
another—just as a tolerably good cricketer, who might be an
acquisition to a country club, would have his wicket struck down at
the first bowl at Lord’s-ground. The same person is frequently dull at

one time and brilliant at another: sometimes those who are most
silent at the beginning of an entertainment are most loquacious at
the end. There is a run in the luck both in cards and conversation.
Some people are good speakers but bad hearers: these are put out,
unless they have all the talk to themselves. Some are best in a tête-à-
tête; others in a mixed company. Some persons talk well on a set
subject, who can hardly answer a common question, still less pay a
compliment or make a repartee. Conversation may be divided into
the personal or the didactic: the one resembles the style of a lecture,
the other that of a comedy. There are as many who fail in
conversation from aiming at too high a standard of excellence, and
wishing only to utter oracles or jeux-d’esprit, as there are who expose
themselves from having no standard at all, and saying whatever
comes into their heads. Pedants and gossips compose the largest
class. Numbers talk on without paying any attention to the effect
they produce upon their audience: some few take no part in the
discourse but by assenting to everything that is said, and these are
not the worst companions in the world. An outcry is sometimes
raised against dull people, as if it were any fault of theirs. The most
brilliant performers very soon grow dull, and we like people to begin
as they end. There is then no disappointment nor false excitement.
The great ingredient in society is good-will. He who is pleased with
what he himself has to say, and listens in his turn with patience and
good-humour, is wise and witty enough for us. We do not covet those
parties where one wit dares not go, because another is expected. How
delectable must the encounter of such pretenders be to one another!
How edifying to the bye-standers!
IX. It was well said by Mr. Coleridge, that people never improve by
contradiction, but by agreeing to differ. If you discuss a question
amicably you may gain a clear insight into it; if you dispute about it
you only throw dust in one another’s eyes. In all angry or violent
controversy, your object is not to learn wisdom, but to prove your
adversary a fool; and in this respect, it must be admitted, both
parties usually succeed.
X. Envy is the ruling passion of mankind. The explanation is
obvious. As we are of infinitely more importance in our own eyes
than all the world beside, the chief bent and study of the mind is
directed to impress others with this self-evident but disputed

distinction, and to arm ourselves with the exclusive signatures and
credentials of our superiority, and to hate and stifle all that stands in
the way of, or obscures, our absurd pretensions. Each individual
looks upon himself in the light of a dethroned monarch, and the rest
of the world as his rebellious subjects and runaway slaves, who
withhold the homage that is his natural due, and burst the chains of
opinion he would impose upon them: the madman in Hogarth (sooth
to say), with his crown of straw and wooden sceptre, is but a type and
common-place emblem of every-day life.
XI. It has been made a subject of regret that in forty or fifty years’
time (if we go on as we have done) no one will read Fielding. What a
falling off! Already, if you thoughtlessly lend Joseph Andrews to a
respectable family, you find it returned upon your hands as an
improper book. To be sure, people read ‘Don Juan’; but that is in
verse. The worst is, that this senseless fastidiousness is more owing
to an affectation of gentility than to a disgust at vice. It is not the
scenes that are described at an alehouse, but the alehouse at which
they take place that gives the mortal stab to taste and refinement.
One comfort is, that the manners and characters which are objected
to as low in Fielding have in a great measure disappeared or taken
another shape; and this at least is one good effect of all excellent
satire—that it destroys ‘the very food whereon it lives.’ The generality
of readers, who only seek for the representation of existing models,
must therefore, after a time, seek in vain for this obvious
verisimilitude in the most powerful and popular works of the kind;
and will be either disgusted or at a loss to understand the
application. People of sense and imagination, who look beyond the
surface or the passing folly of the day, will always read Tom Jones.
XII. There is a set of critics and philosophers who have never read
anything but what has appeared within the last ten years, and to
whom every mode of expression or turn of thought extending beyond
that period has a very odd effect. They cannot comprehend how
people used such out-of-the-way phrases in the time of Shakspeare;
the style of Addison would not do now—even Junius, they think,
would make but a shabby thread-bare figure in the columns of a
modern newspaper—all the riches that the language has acquired in
the course of time, all the idiomatic resources arising from study or
accident, are utterly discarded—sink under-ground: and all that is

admired by the weak or sought after by the vain, is a thin surface of
idle affectation and glossy innovations. Even spelling and
pronunciation have undergone such changes within a short time,
that Pope and Swift require a little modernizing to accommodate
them to ‘ears polite;’ and that a bluestocking belle would be puzzled
in reciting Dryden’s sounding verse with its occasional barbarous,
old-fashioned accenting, if it were the custom to read Dryden aloud
in those serene, morning circles. There is no class more liable to set
up this narrow superficial standard, than people of fashion, in their
horror of what is vulgar and ignorance of what really is so; they have
a jargon of their own, but scout whatever does not fall in with it as
Gothic and outré; the English phrases handed down from the last age
they think come east of Temple-bar, and they perform a sedulous
quarantine against them. The Times, having found it so written in
some outlandish depêche of the Marquis of Wellesley’s, chose as a
mark of the haute literature, to spell dispatch with an e, and for a
long time he was held for a novice or an affected and absolute writer
who spelt it otherwise. The Globe, with its characteristic good sense
and sturdiness of spirit has restored the old English spelling in
defiance of scandal. Some persons who were growing jealous that the
author of Waverley had eclipsed their favourite luminaries may
make themselves easy; he himself is on the wane with those whose
opinions ebb and flow with the ‘inconstant moon’ of fashion, and has
given way (if Mr. Colburn’s advertisements speak true, ‘than which
what’s truer?’) to a set of titled nonentities. Nothing solid is to go
down, or that is likely to last three months; instead of the standing
dishes of old English literature we are to take up with the nicknacks
and whipt syllabubs of modern taste; are to be occupied with a
stream of titlepages, extracts, and specimens, like passing figures in a
camera obscura, and are to be puzzled in a mob of new books as in
the mob of new faces in what was formerly the narrow part of the
Strand.
XIII. Never pity people because they are ill-used. They only wait
the opportunity to use others just as ill. Hate the oppression and
prevent the evil if you can; but do not fancy there is any virtue in
being oppressed, or any love lost between the parties. The
unfortunate are not a jot more amiable than their neighbours,
though they give themselves out so, and our pity takes part with
those who have disarmed our envy.

XIV. The human mind seems to improve, because it is continually
in progress. But as it moves forward to new acquisitions and
trophies, it loses its hold on those which formerly were its chief boast
and employment. Men are better chemists than they were, but worse
divines; they read the newspapers, it is true, but neglect the classics.
Everything has its turn. Neither is error extirpated so much as it
takes a new form and puts on a more artful disguise. Folly shifts its
ground, but finds its level: absurdity is never left without a
subterfuge. The dupes of dreams and omens in former times, are
now the converts to graver and more solemn pieces of quackery. The
race of the sanguine, the visionary, and the credulous, of those who
believe what they wish, or what excites their wonder, in preference to
what they know, or can have rationally explained, will never wear
out; and they only transfer their innate love of the marvellous from
old and exploded chimeras to fashionable theories, and the terra
incognita of modern science.
XV. It is a curious speculation to take a modern belle, or some
accomplished female acquaintance, and conceive what her great-
great-grandmother was like, some centuries ago. Who was the Mrs.
—— of the year 200? We have some standard of grace and elegance
among eastern nations 3000 years ago, because we read accounts of
them in history; but we have no more notion of, or faith in, our own
ancestors than if we had never had any. We cut the connexion with
the Druids and the Heptarchy; and cannot fancy ourselves (by any
transformation) inmates of caves and woods, or feeders on acorns
and sloes. We seem engrafted on that low stem—a bright, airy, and
insolent excrescence.
XVI. There is this advantage in painting, if there were no other,
that it is the truest and most self-evident kind of history. It shows
that there were people long ago, and also what they were, not in a
book darkly, but face to face. It is not the half-formed clay, the old-
fashioned dress, as we might conceive; but the living lineaments, the
breathing expression. You look at a picture by Vandyke, and there
see as in an enchanted mirror, an English woman of quality two
hundred years ago, sitting in unconscious state with her child playing
at her feet, and with all the dove-like innocence of look, the grace and
refinement that it is possible for virtue and breeding to bestow. It is
enough to make us proud of our nature and our countrywomen; and

dissipates at once the idle, upstart prejudice that all before our time
was sordid and scarce civilised. If our progress does not appear so
great as our presumption has suggested, what does it signify? With
such models kept in view, our chief object ought to be not to
degenerate; and though the future prospect is less gaudy and
imposing, the retrospect opens a larger and brighter vista of
excellence.
XVII. I am by education and conviction inclined to republicanism
and puritanism. In America they have both; but I confess I feel a
little staggered in the practical efficacy and saving grace of first
principles, when I ask myself, ‘Can they throughout the United
States, from Boston to Baltimore, produce a single head like one of
Titian’s Venetian nobles, nurtured in all the pride of aristocracy and
all the blindness of popery?’ Of all the branches of political economy,
the human face is perhaps the best criterion of value.

COMMON SENSE
The Atlas.]
[October 11, 1829.
Common sense is a rare and enviable quality. It may be truly said
that ‘its price is above rubies.’ How many learned men, how many
wits, how many geniuses, how many dull and ignorant people, how
many cunning knaves, how many well-meaning fools are without it!
How few have it, and how little do they or others know of it, except
from the infallible results—for one of its first requisites is the utter
absence of all pretension! The vulgar laugh at the pedant and
enthusiast for the want of it, while they themselves mistake bigotry
and narrow-minded notions for it. It is not one of the sciences, but
has been well pronounced to be ‘fairly worth the seven.’ It is a kind of
mental instinct, that feels the air of truth and propriety as the fingers
feel objects of touch. It does not consist with ignorance, for we
cannot pronounce on what we do not know; and on the other hand,
the laying in a stock of knowledge, or mastering any art or science,
seems to destroy that native simplicity, and to warp and trammel the
unbiassed freedom of mind which is necessary to its receiving and
giving their due weight to ordinary and casual impressions. Common
sense is neither a peculiar talent nor a laborious acquirement, but
may be regarded as a sound and impartial judgment operating on the
daily practice of life, or on what ‘comes home to the business and
bosoms of men’; combined with great attainments and speculative
inquiries, it would justly earn the title of wisdom; but of the latter we
have never known a single instance, though we have met with a few
of the former; that is, we have known a number of persons who were
wise in the affairs of the world and in what concerned their own
interest, but none who, beyond this, and in judging of general

questions, were not the dupes of some flaw of temper, of some
weakness or vanity, or even striking advantage of their own. To give
an example of two in illustration. A person may be an excellent
scholar, a good mathematician, well versed in law and history, a first-
rate chess-player, a dazzling fencer, in a word, a sort of admirable
Crichton—you are disposed to admire or envy so many talents united
—you smile to see him wanting in common sense, and getting into a
dispute about a douceur to a paltry police-officer, and thinking to
interest all Europe and both Houses of Parliament in his success. It is
true, he has law and reason on his side, has Grotius and Puffendorf
and the statutes at large doubled down in dog-ears for the occasion,
has a vast and lively apparatus of well-arranged premises and
conclusions ready to play off against his adversaries; but he does not
consider that he has to deal with interest and custom, those
impalpable, intangible essences, that ‘fear no discipline of human
wit.’ Does he think to check-mate the police? Will he stop the mouth
of a hungry tide-waiter with a syllogism? Or supersede a perquisite
by the reductio ad absurdum? It is a want of common sense, or the
not distinguishing properly between the definite and the indefinite.
No one can have arrived at years of discretion without knowing or
feeling that he cannot take a single step without some compromise
with existing circumstances; that the path of life is intercepted with
innumerable turnpike-gates, at which he must pay down the toll of
his own convictions and of strict justice; that he cannot walk the
streets but by tacit allowance; and that to disregard all impediments
in the right line of reason and written forms is to imitate the conduct
of Commodore Trunnion, who mistook the land for the sea, and
went to be married by the wind and compass. The proofs of this
occur every hour of the day—they may not be registered, they may
not be remembered, but they are virtually and effectively noted down
by the faculty of common sense, which does not feel its way the less
surely because it proceeds often mechanically and blindly. There may
be exceptions indeed to ordinary rules, on which a man may go to
martyrdom and a stake (such as that of Hampden and ship-money),
but these occur once in a century, and are only met with at the
corners of streets by those who have an excess of logical
discrimination, and have to pay a certain tax for being too clever by
half. It is the fashion at present among the philosophical vulgar to
decry feeling, both the name and the thing. It would be difficult,

however, to do without it: for this word embraces all that mass of
knowledge and of common sense which lies between the extremes of
positive proof or demonstration and downright ignorance; and those
who would pragmatically confine their own convictions or those of
others to what is absolutely known and understood, would at best
become scientific pedants and artificial barbarians. There are some
persons who are the victims of argument; as there are others who are
the slaves of minute details and matters of fact. One class will have a
reason for every thing, and will admit the greatest absurdities that
are formally proposed to them; the other must have facts to support
every conclusion, and can never see an inch beyond their noses. The
last have the organ of individuality largely developed, and are
proportionably deficient in common sense. Their ideas are all local
and literal. To borrow the language of a great but obscure
metaphysician, their minds are epileptic; that is, are in perpetual
throes and convulsions, fasten on every object in their way not to
help but to hinder their progress, and have no voluntary power to let
go their hold of a particular circumstance, to grasp the whole of any
question, or suspend their judgment for an instant. The fact that is
before them is every thing; the rest goes for nothing. They are always
at cross-purposes with themselves, for their decisions are the result
of the last evidence, without any corrective or qualifier in common
sense; in the hunt after proofs, they forget their principles, and gain
their point, though they lose their cause.
The Scotch have much of this matter-of-fact understanding, and
bigotry to personal and actual statistics. They would persuade you
that there is no country but Scotland, nothing but what is Scotch. Mr.
Mac Alpine shifts the discourse from the metropolis, hurries rapidly
over the midland counties, crosses the border, and sits down to an
exordium in praise of the ‘kindly Scot.’ Charity has its home and
hearth by Tweed-side, where he was born and bred, Scotch beggars
were quite different from English beggars: there was none of the
hard-heartedness towards them that was always shown in England.
His mother, though not a rich woman, always received them kindly,
and had a bag of meal out of which she always gave them something,
as they went their rounds. ‘Lord! Mr. Mac Alpine!’ says Mrs. Mac
Alpine, ‘other people have mothers as well as you, and there are
beggars in England as well as Scotland. Why, in Yorkshire, where I
was brought up, common beggars used to come round just as you

describe, and my mother, who was no richer than yours, used to give
them a crust of bread or broken victuals just in the same way; you
make such a fuss about nothing.’ Women are best to set these follies
to rights:—
‘They have no figures nor no fantasies,
Which busy thought draws in the brains of men.’
If no great philosophers, they do not want common sense; and are
only misled in what lies beyond their sphere of feeling and
observation, by taking up the opinions of their better halves. The
common people in like manner do not want common sense in what
falls under their especial cognizance and daily practice. A country-
shoemaker or plough-man understands shoemaking, and can ‘crack
of ploughs and kine,’ though he knows nothing of the Catholic
question. If an old woman in a country-town believes she shall be
burnt at a stake, now that this question is settled, it is because she is
told so by those who ought to know better, and who impose their
prejudices upon her ignorance. Vulgar errors which are taken on
trust, or are traditional, or are the blunders of ignorance on points of
learning, have nothing to do with common sense, which decides only
on facts and feelings which have come under its own notice.
Common sense and common-place are also the antipodes of each
other: the one is a collection of true experiences, the other a routine
of cant phrases. All affectation is the death of common sense, which
requires the utmost simplicity and sincerity. Liars must be without
common sense, for instead of considering what things really are,
their whole time and attention are taken up in imposing false
appearances on themselves and their neighbours. No conceited
person can have the faculty we have been speaking of, since all
objects are tinged and changed from their proper hue by the idle
reflection of their fancied excellence and superiority. Great talkers
are in the same predicament, for they sacrifice truth to a fine speech
or sentiment, and conceal the real consequences of things from their
view by a cloud of words, of empty breath. They look at nature not to
study what it is, but to discover what they can say about it.
Passionate people are generally thought to be devoid of judgment.
They may be so, when their passions are touched to the quick; but
without a certain degree of natural irritability, we do not conceive
truth leaves sufficient stings in the mind, and we judge correctly of

things according to the interest we take in them. No one can be a
physiognomist, for example, or have an insight into character and
expression, without the correspondent germs of these in his own
breast. Phlegmatic C——, with all his husbandry acquirements, is but
half a philosopher, half a clown. Poets, if they have not common
sense, can do very well without it. What need have they to conform
their ideas to the actual world, when they can create a world
according to their fancy? We know of no remedy for want of tact and
insight into human affairs, any more than for the defect of any other
organ. Tom Jones is, we think, the best horn-book for students in
this way; and if the novice should rise up no wiser from its repeated
perusal, at least such an employment of his time will be better than
playing the fool or talking nonsense. After all, the most absurd
characters are those who are so, not from a want of common sense,
but who act in defiance of their better knowledge. The capricious and
fickle who change every moment, the perverse who aim only at what
is placed out of their reach, the obstinate who pursue a losing cause,
the idle and vicious who ruin themselves and every one connected
with them, do it as often with their eyes open as from blind
infatuation; and it is the bias of their wills, not the deficiency of their
understandings, that is in fault. The greatest fools in practice are
sometimes the wisest men in theory, for they have all the advantage
of their own experience and self-reflection to prompt them; and they
can give the best advice to others, though they do not conceive
themselves bound to follow it in their own instance. Video meliora
proboque, etc. Their judgments may be clear and just, but their
habits and affections lie all the wrong way; and it is as useless as it
would be cruel to expect them to reform, since they only delight and
can only exist in their darling absurdities and daily and hourly
escapades from common sense and reason.

THE SPIRIT OF CONTROVERSY
The Atlas.]
[January 31, 1830.
The Spirit of Controversy has often been arraigned as the source of
much bitterness and vexation, as productive of ‘envy, malice, and all
uncharitableness’: and the charge, no doubt, is too well founded. But
it is said to be an ill wind that blows nobody good; and there are few
evils in life that have not some qualifying circumstance attending
them. It is one of the worst consequences of this very spirit of
controversy that it has led men to regard things too much in a single
and exaggerated point of view. Truth is not one thing, but has many
aspects and many shades of difference; it is neither all black nor all
white; sees something wrong on its own side, something right in
others; makes concessions to an adversary, allowances for human
frailty, and is nearer akin to charity than the dealers in controversy
or the declaimers against it are apt to imagine. The bigot and
partisan (influenced by the very spirit he finds fault with) sees
nothing in the endless disputes which have tormented and occupied
men’s thoughts but an abuse of learning and a waste of time: the
philosopher may still find an excuse for so bad and idle a practice.
One frequent objection made to the incessant wrangling and collision
of sects and parties is, What does it all come to? And the answer is,
What would they have done without it? The pleasure of the chase, or
the benefit derived from it, is not to be estimated by the value of the
game after it is caught, so much as by the difficulty of starting it and
the exercise afforded to the body and the excitement of the animal
spirits in hunting it down: and so it is in the exercises of the mind
and the pursuit of truth, which are chiefly valuable (perhaps) less for
their results when discovered, than for their affording continual

scope and employment to the mind in its endeavours to reach the
fancied goal, without its being ever (or but seldom) able to attain it.
Regard the end, is an ancient saying, and a good one, if it does not
mean that we are to forget the beginning and the middle. By
insisting on the ultimate value of things when all is over, we may
acquire the character of grave men, but not of wise ones. Passe pour
cela. If we would set up such a sort of fixed and final standard of
moral truth and worth, we had better try to construct life over again,
so as to make it a punctum stans, and not a thing in progress; for as
it is, every end, before it can be realised, implies a previous
imagination, a warm interest in, and an active pursuit of, itself, all
which are integral and vital parts of human existence, and it is a
begging of the question to say that an end is only of value in itself,
and not as it draws out the living resources, and satisfies the original
capacities of human nature. When the play is over, the curtain drops,
and we see nothing but a green cloth; but before this, there have
been five acts of brilliant scenery and high-wrought declamation,
which, if we come to plain matter-of-fact and history, are still
something. According to the contrary theory, nothing is real but a
blank. This flatters the paradoxical pride of man, whose motto is, all
or none. Look at that pile of school divinity! Behold where the demon
of controversy lies buried! The huge tomes are mouldy and worm-
eaten:—did their contents the less eat into the brain, or corrode the
heart, or stir the thoughts, or fill up the void of lassitude and ennui in
the minds of those who wrote them? Though now laid aside and
forgotten, if they had not once had a host of readers, they would
never have been written; and their hard and solid bulk asked the
eager tooth of curiosity and zeal to pierce through it. We laugh to see
their ponderous dulness weighed in scales, and sold for waste paper.
We should not laugh too soon. On the smallest difference of faith or
practice discussed in them, the fate of kingdoms hung suspended;
and not merely so (which was a trifle) but Heaven and Hell trembled
in the balance, according to the full persuasion of our pious
forefathers. Many a drop of blood flowed in the field or on the
scaffold, from these tangled briars and thorns of controversy; many a
man marched to a stake to bear testimony to the most frivolous and
incomprehensible of their dogmas. This was an untoward
consequence; but if it was an evil to be burnt at a stake, it was well
and becoming to have an opinion (whether right or wrong) for which

a man was willing to be burnt at a stake. Read Baxter’s Controversial
Works: consider the flames of zeal, the tongues of fire, the heights of
faith, the depths of subtlety, which they unfold, as in a darkly
illuminated scroll; and then ask how much we are gainers by an utter
contempt and indifference to all this? We wonder at the numberless
volumes of sermons that have been written, preached, and printed
on the Arian and Socinian controversies, on Calvinism and
Arminianism, on surplices and stoles, on infant or adult baptism, on
image-worship and the defacing of images; and we forget that it
employed the preacher all the week to prepare his sermon (be the
subject what it would) for the next Lord’s day, with infinite collating
of texts, authorities, and arguments; that his flock were no less
edified by listening to it on the following Sunday; and how many
David Deans’s came away convinced that they had been listening to
the ‘root of the matter’! See that group collected after service-time
and pouring over the gravestones in the churchyard, from whence, to
the eye of faith, a light issues that points to the skies! See them
disperse; and as they take different paths homeward while the
evening closes in, still discoursing of the true doctrine and the glad
tidings they have heard, how ‘their hearts burn within them by the
way’! Then again, we should set down, among other items in the
account, how the schoolboy is put to it to remember the text, and
how the lazy servant-wench starts up to find herself asleep in church-
time! Such is the business of human life; and we, who fancy
ourselves above it, are only so much the more taken up with follies of
our own. We look down in this age of reason on those controverted
points and nominal distinctions which formerly kept up such ‘a coil
and pudder’ in the world, as idle and ridiculous, because we are not
parties to them; but if it was the egotism of our predecessors that
magnified them beyond all rational bounds, it is no less egotism in us
who undervalue their opinions and pursuits because they are not
ours; and, indeed, to leave egotism out of human nature, is ‘to leave
the part of Hamlet out of the play of Hamlet.’ Or what are we the
better with our Utilitarian Controversies, Mr. Taylor’s discourses
(delivered in canonicals) against the evidence of the Christian
religion, or the changes of ministry and disagreements between the
Duke of Wellington and the Duke of Newcastle?
‘Strange! that such difference should be

’Twixt Tweedledum and Tweedledee!’
But the prevalence of religious controversy is reproached with
fomenting spiritual pride and intolerance, and sowing heart-
burnings, jealousies, and fears, ‘like a thick scurf o’er life;’ yet, had it
not been for this, we should have been tearing one another to pieces
like savages for fragments of raw flesh, or quarrelling with a herd of
swine for a windfall of acorns under an oak-tree. The world has never
yet done, and will never be able to do, without some apple of discord
—some bone of contention—any more than courts of law can do
without pleadings, or hospitals without the sick. When a thing ceases
to be a subject of controversy, it ceases to be a subject of interest.
Why need we regret the various hardships and persecutions for
conscience-sake, when men only clung closer to their opinions in
consequence? They loved their religion in proportion as they paid
dear for it. Nothing could keep the Dissenters from going to a
conventicle while it was declared an unlawful assembly, and was the
highroad to a prison or the plantations—take away tests and fines,
and make the road open and easy, and the sect dwindles gradually
into insignificance. A thing is supposed to be worth nothing that
costs nothing. Besides, there is always pretty nearly the same
quantity of malice afloat in the world; though with the change of time
and manners it may become a finer poison, and kill by more unseen
ways. When the sword has done its worst, slander, ‘whose edge is
sharper than the sword,’ steps in to keep the blood from stagnating.
Instead of slow fires and paper caps fastened round the heads of the
victims, we arrive at the same end by a politer way of nicknames and
anonymous criticism. Blackwood’s Magazine is the modern version
of Fox’s Book of Martyrs. Discard religion and politics (the two
grand topics of controversy), and people would hate each other as
cordially, and torment each other as effectually about the preference
to be given to Mozart or Rossini, to Malibran or Pasta. We indeed fix
upon the most excellent things, as God, our country, and our King, to
account for the excess of our zeal; but this depends much less upon
the goodness of our cause than on the strength of our passions, and
our overflowing gall and rooted antipathy to whatever stands in the
way of our conceit and obstinacy. We set up an idol (as we set up a
mark to shoot at) for others to bow down to, on peril of our utmost
displeasure, let the value of it be what it may——

‘Of whatsoe’er descent his Godhead be,
Stock, stone, or other homely pedigree,
In his defence his servants are as bold
As if he had been born of beaten gold.’
It is, however, but fair to add, in extenuation of the evils of
controversy, that if the points at issue had been quite clear, or the
advantage all on one side, they would not have been so liable to be
contested about. We condemn controversy, because we would have
matters all our own way, and think that ours is the only side that has
a title to be heard. We imagine that there is but one view of a subject
that is right; and that all the rest being plainly and wilfully wrong, it
is a shocking waste of speech, and a dreadful proof of prejudice and
party spirit, to have a word to say in their defence. But this is a want
of liberality and comprehension of mind. For in general we dispute
either about things respecting which we are a good deal in the dark,
and where both parties are very possibly in the wrong, and may be
left to find out their mutual error; or about those points, where there
is an opposition of interests and passions, and where it would be by
no means safe to cut short the debate by making one party judges for
the other. They must, therefore, be left to fight it out as well as they
can; and, between the extremes of folly and violence, to strike a
balance of common sense and even-handed justice. Every sect or
party will, of course, run into extravagance and partiality; but the
probability is, that there is some ground of argument, some
appearance of right, to justify the grossest bigotry and intolerance.
The fury of the combatants is excited because there is something to
be said on the other side of the question. If men were as infallible as
they suppose themselves, they would not dispute. If every novelty
were well founded, truth might be discovered by a receipt; but as
antiquity does not always turn out an old woman, this accounts for
the vis inertiæ of the mind in so often pausing and setting its face
against innovation. Authority has some advantages to recommend it
as well as reason, or it would long ago have been scouted. Aristocracy
and democracy, monarchy and republicanism, are not all pure good
or pure evil, though the abettors or antagonists of each think so, and
that all the mischief arises from others entertaining any doubt about
the question, and insisting on carrying their absurd theories into
practice. The French and English are grossly prejudiced against each
other; but still the interests of each are better taken care of under this

exaggerated notion than if that vast mass of rights and pretensions,
which each is struggling for, were left to the tender mercies and
ruthless candour of the other side. ‘Every man for himself and God
for us all’ is a rule that will apply here. Controversy, therefore, is a
necessary evil or good (call it which you will) till all differences of
opinion or interest are reconciled, and absolute certainty or perfect
indifference alike takes away the possibility or the temptation to
litigation and quarrels. We need be under no immediate alarm of
coming to such a conclusion. There is always room for doubt, food
for contention. While we are engrossed with one controversy, indeed,
we think every thing else is clear; but as soon as one point is settled,
we begin to cavil and start objections to that which has before been
taken for gospel. The Reformers thought only of opposing the
Church of Rome, and never once anticipated the schisms and
animosities which arose among Protestants: the Dissenters, in
carrying their point against the Church of England, did not dream of
that crop of infidelity and scepticism which, to their great horror and
scandal, sprung up in the following age, from their claim of free
inquiry and private judgment. The non-essentials of religion first
came into dispute; then the essentials. Our own opinion, we fancy, is
founded on a rock; the rest we regard as stubble. But no sooner is
one out-work of established faith or practice demolished, than
another is left a defenceless mark for the enemy, and the engines of
wit and sophistry immediately begin to batter it. Thus we proceed
step by step, till, passing through the several gradations of vanity and
paradox, we came to doubt whether we stand on our head or our
heels, alternately deny the existence of spirit and matter, maintain
that black is white, call evil good and good evil, and defy any one to
prove the contrary. As faith is the prop and cement that upholds
society by opposing fixed principles as a barrier against the inroads
of passion, so reason is the menstruum which dissolves it by leaving
nothing sufficiently firm or unquestioned in our opinions to
withstand the current and bias of inclination. Hence the decay and
ruin of states—then barbarism, sloth, and ignorance—and so we
commence the circle again of building up all that it is possible to
conceive out of a rude chaos, and the obscure shadowings of things,
and then pulling down all that we have built up, till not a trace of it is
left. Such is the effect of the ebb and flow and restless agitation of the
human mind.

ENVY
The Atlas.]
[February 14, 1830.
Envy is the grudging or receiving pain from any accomplishment
or advantage possessed by another. It is one of the most tormenting
and odious of the passions, inasmuch as it does not consist in the
enjoyment or pursuit of any good to ourselves, but in the hatred and
jealousy of the good fortune of others and the debarring and
defrauding them of their due and what is of no use to us, on the dog
in the manger principle; and it is at the same time as mean as it is
revolting, as being accompanied with a sense of weakness and a
desire to conceal and tamper with the truth and its own convictions,
out of paltry spite and vanity. It is, however, but an excess or
excrescence of the other passions (such as pride or avarice) or of a
wish to monopolise all the good things of life to ourselves, which
makes us impatient and dissatisfied at seeing any one else in
possession of that to which we think we have the only fair title. Envy
is the deformed and distorted offspring of egotism; and when we
reflect on the strange and disproportioned character of the parent,
we cannot wonder at the perversity and waywardness of the child.
Such is the absorbing and exorbitant quality of our self-love, that it
represents us as of infinitely more importance in our own eyes than
the whole universe put together, and would sacrifice the claims and
interest of all the world beside to the least of its caprices or
extravagances; need we be surprised then that this little, upstart,
overweening self, that would trample on the globe itself and then
weep for new ones to conquer, should be uneasy, mad, mortified,
eaten up with chagrin and melancholy, and hardly able to bear its
own existence, at seeing a single competitor among the crowd cross

its path, jostle its pretensions, and stagger its opinion of its exclusive
right to admiration and superiority? This it is that constitutes the
offence, that gives the shock, that inflicts the wound, that some poor
creature (as we would fain suppose) whom we had before overlooked
and entirely disregarded as not worth our notice, should of a sudden
enter the lists and challenge comparison with us. The presumption is
excessive; and so is our thirst of revenge. From the moment,
however, that the eye fixes on another as the object of envy, we
cannot take it off; for our pride and self-conceit magnify that which
obstructs our success and lessens our self-importance into a
monster; we see nothing else, we hear of nothing else, we dream of
nothing else, it haunts us and takes possession of our whole souls;
and as we are engrossed by it ourselves, so we fancy that all the rest
of the world are equally taken up with our petty annoyances and
disappointed pride. Hence the ‘jealous leer malign’ of envy, which,
not daring to look that which provokes it in the face, cannot yet keep
its eyes from it, and gloats over and becomes as it were enamoured of
the very object of its loathing and deadly hate. We pay off the score
which our littleness and vanity has been running up, by ample and
gratuitous concessions to the first person that gives a check to our
swelling self-complacency, and forces us to drag him into an
unwilling comparison with ourselves. It is no matter who the person
is, what his pretensions—if they are a counterpoise to our own, we
think them of more consequence than anything else in the world.
This often gives rise to laughable results. We see the jealousies
among servants, hackney-coachmen, cobblers in a stall; we are
amused with the rival advertisements of quacks and stage-coach
proprietors, and smile to read the significant intimation on some
shop window, ‘No connection with the next door;’ but the same folly
runs through the whole of life; each person thinks that he who stands
in his way or outstrips him in a particular pursuit, is the most
enviable, and at the same time the most hateful character in the
world. Nothing can show the absurdity of the passion of envy in a
more striking point of view than the number of rival claims which it
entirely overlooks, while it would arrogate all excellence to itself. The
loftiness of our ambition and the narrowness of our views are equal,
and indeed both depend upon the same cause. The player envies only
the player, the poet envies only the poet, because each confines his
idea of excellence to his own profession and pursuit, and thinks, if he

could but remove some hapless competitor out of his way, he should
have a clear stage to himself, or be a ‘Phœnix gazed by all:’ as if,
though we crushed one rival, another would not start up; or as if
there were not a thousand other claims, a thousand other modes of
excellence and praiseworthy acquirements, to divide the palm and
defeat his idle pretension to the sole and unqualified admiration of
mankind. Professors of every class see merit only in their own line;
yet they would blight and destroy that little bit of excellence which
alone they acknowledge to exist, except as it centres in themselves.
Speak in praise of an actor to another actor, and he turns away with
impatience and disgust: speak disparagingly of the first as an actor in
general, and the latter eagerly takes up the quarrel as his own: thus
the esprit de corps only comes in as an appendage to our self-love. It
is perhaps well that we are so blind to merit out of our immediate
sphere, for it might only prove an additional eye-sore, increase the
obliquity of our mental vision, multiply our antipathies, or end in
total indifference and despair. There is nothing so bad as a cynical
apathy and contempt for every art and science from a superficial
smattering and general acquaintance with them all. The merest
pedantry and the most tormenting jealousy and heart-burnings of
envy are better than this. Those who are masters of different
advantages and accomplishments, are seldom the more satisfied with
them: they still aim at something else (however contemptible) which
they have not or cannot do. So Pope says of Wharton—
‘Though wondering senates hung on all he spoke,
The club must hail him master of the joke.
Shall parts so various aim at nothing new?
He’ll shine a Tully and a Wilmot too.’
The world, indeed, are pretty even with these constellations of
splendid and superfluous qualities in their fastidious estimate of
their own pretensions, for (if possible) they never give any individual
credit for more than one leading attainment. If a man is an artist, his
being a fine musician adds nothing to his fame. When the public
strain a point to own one claim, it is on condition that the fortunate
candidate waives every other. The mind is prepared with a plausible
antithesis in such cases against the formidable encroachments of
vanity: one qualification is regularly made a foil to another. We allow
no one to be two things at a time: it quite unsettles our notions of

Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade
Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.
Let us accompany you on the journey of exploring knowledge and
personal growth!
ebookfinal.com

[FREE PDF sample] Minimalist Syntax Randall Hendrick ebooks

More Related Content

Similar to [FREE PDF sample] Minimalist Syntax Randall Hendrick ebooks (20)

Recently uploaded (20)

[FREE PDF sample] Minimalist Syntax Randall Hendrick ebooks