Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, LINGUISTICS (linguistics.oxfordre.com). (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy).

date: 17 August 2017

Derivational Economy in Syntax and Semantics

Summary and Keywords

Economy considerations have always played an important role in the generative theory of grammar. They are particularly prominent in the most recent instantiation of this approach, the Minimalist Program, which explores the possibility that Universal Grammar is an optimal way of satisfying requirements that are imposed on the language faculty by the external systems that interface with the language faculty which is also characterized by optimal, computationally efficient design. In this respect, the operations of the computational system that produce linguistic expressions must be optimal in that they must satisfy general considerations of simplicity and efficient design. Simply put, the guiding principles here are (a) do something only if you need to and (b) if you do need to, do it in the most economical/efficient way. These considerations ban superfluous steps in derivations and superfluous symbols in representations. Under economy guidelines, movement takes place only when there is a need for it (with both syntactic and semantic considerations playing a role here), and when it does take place, it takes place in the most economical way: it is as short as possible and carries as little material as possible. Furthermore, economy is evaluated locally, on the basis of immediately available structure. The locality of syntactic dependencies is also enforced by minimal search and by limiting the number of syntactic objects and the amount of structure accessible in the derivation. This is achieved by transferring parts of syntactic structure to the interfaces during the derivation, the transferred parts not being accessible for further syntactic operations.

Keywords: binding, economy, EPP, Inclusiveness Condition, labels, last resort, multiple spell-out, phases, scope

1. General Economy Considerations

Economy considerations have always played an important role in the generative theory of grammar. Early on, they took the form of an evaluation metric for selecting grammars from the format permitted for rule systems. The role of economy has changed in the most recent instantiation of this approach, the Minimalist Program (MP), where there is no need for such an evaluation metric due to the restrictiveness of the theory. Still, there has been no decrease in the importance of Economy Principles. In fact, they play a central role in the foundational works of MP (see Chomsky, 1995).

MP explores the possibility that Universal Grammar, a genetic endowment that helps children acquire language, is an optimal way of satisfying requirements imposed on the language faculty by the external systems the language faculty interfaces with and is also characterized by optimal, computationally efficient design. Language consists of a lexicon and a computational system, which is embedded in two performance systems: articulatory-perceptual and conceptual-intentional, with Phonological Form (PF) and Logical Form (LF) interfacing with the performance systems. A derivation converges at the interface levels if it contains only legitimate PF and LF objects. However, defining linguistic expressions simply as pairs (P, L) formed by a convergent derivation and satisfying interface conditions does not suffice; the operations of the computational system that produce linguistic expressions must be optimal in that they must satisfy some general considerations of simplicity and efficient design. These considerations ban superfluous steps in derivations and superfluous symbols in representations, favor short movements over long movements, etc. Simply put, the guiding principles here are do something only if you need to, and if you do need to, do it in the most economical/efficient way.

2. Derivational Economy in Syntax

Consider first the former, as it applies to movement. It has led to the last resort nature of movement: in line with the leading idea of economy, movement must happen for a reason. One such driving force may be provided by Case. Consider (1).

(1)

Derivational Economy in Syntax and Semantics

Since the subject position of raising infinitives is not a Case-licensing position, Jane cannot be Case-licensed in the position of t. Raising to matrix SpecIP, a nominative-licensing position, rectifies its Case inadequacy. Once Jane has been Case-licensed, it can no longer undergo A-movement, to a Case or a non-Case position. This follows from Last Resort, if A-movement is driven by Case considerations. Jane being Case-licensed in the position of t in (2), Last Resort blocks its movement in (2a-b).

(2)

Derivational Economy in Syntax and Semantics

The efforts to identify the precise reason for each movement have led to raising the more general theoretical issue of where the formal inadequacy driving movement lies, in the target (Attract) or the moving element (Greed). Greed was the earliest approach (Chomsky, 1993, revived in Bošković, 2007, based on considerations of successive-cyclic movement). Under Greed, X can move only if X has a formal inadequacy, and if the movement will help rectify the inadequacy. Under Attract, the target head always triggers movement (Chomsky, 1995), which means it always has a formal inadequacy that is rectified by the movement (see also Lasnik, 1995, for a combination of these two approaches to Last Resort). Under this approach, movement of Jane in (1) is driven by I, which has a property (the EPP) that is satisfied by a DP—this property triggers movement of Jane. (Jane’s Case-licensing is merely a beneficial side effect of the satisfaction of the target’s requirement.) Returning to (2), while the Greed approach easily handles (2), such examples require additional assumptions in the target-driven system. Thus, Chomsky (2000) posits the Activation Condition, where X can move only if X has an uninterpretable feature, i.e., a formal inadequacy (the need to license its Case feature in the case of A-movement). This essentially reintroduces Greed into a system where movement is supposed to be target-driven. However, the target-driven approach has a conceptual advantage in that under this approach the driving force for movement can be satisfied as soon as the relevant element enters the structure, rather than indefinitely later in the derivation, as in the moving-element driven approach. As discussed in Bošković (2011), there are, however, cases of movement that are driven by the properties of moving elements, like quantifier raising (which does not occur because the target of quantifier raising would require an adjoined quantifier; see also Section 2) or multiple movements to the same position such as multiple wh-fronting in languages like Bulgarian, where all wh-phrases must move to SpecCP (see below). This in itself suggests that there is something wrong with wh-phrases that do not undergo movement.

Another natural principle of efficient computation that has the same flavor as Last Resort—namely, do nothing if you don’t have to—is the Inclusiveness Condition (Chomsky, 1995), which confines the power of syntax to (re)arrangements of lexical items, banning syntax from creating new objects. The condition is very appealing conceptually due to its restrictive nature. It has led to elimination of a number of mechanisms from the Government and Binding theory, like traces, indices, and bar-levels, which in turn has led to a new conception of phrase structure (Bare Phrase structure) and the copy theory of movement.

One question regarding Last Resort is whether lexical insertion, i.e., pure Merge, should be subject to Last Resort. Chomsky (1995) suggests it shouldn’t be, the reasoning being that if cost were assigned to it, no lexical insertion would ever take place. On the other hand, Chomsky (2000) suggests that pure Merge is subject to Last Resort, with pure merge being driven by selectional requirements. This, however, led to a considerable enrichment of selectional requirements. Bošković (1997) takes a position in between Chomsky (1995) and Chomsky (2000). A number of authors have proposed economy-of-representation principles intended to ban unnecessary projections (e.g., Bošković, 1997; Chomsky, 1995; Grimshaw, 1997; Law, 1991; Safir, 1993; Speas, 1994). Noticing such principles typically ban only unnecessary functional structure, Bošković (1997) suggests tying them to Last Resort. The suggestion is that only lexical (as opposed to functional) items are present in the numeration, which is defined as an array of lexical items mapped by the computational system into a linguistic expression (Chomsky, 1995). Assuming that derivations that do not exhaust numerations do not converge, inserting an element from a numeration into the structure is a step toward a well-formed derivation, hence in accordance with Last Resort. Lexicon is then accessed during the derivation only when a certain functional category becomes necessary in structure building (e.g., feature-checking of a feature of a lexical category may require presence of a particular functional category), again in accordance with Last Resort. This makes lexical insertion uniformly subject to Last Resort.1

Returning to movement, the conception of movement as a last resort operation brings up the question of optionality. Truly optional movements do not fit well with a framework where movement takes place for a reason: in such a framework either there is a reason for a movement to take place, in which case it must take place, or there isn’t, in which case it in fact cannot take place, given Last Resort. Everything else being equal, we then would not expect to find truly optional movements. In a case like French (3), where wh-movement appears to be optional, it must then be the case that everything else is not equal. Consider then the broader French wh-movement/wh-in-situ paradigm in (3)–(5), which can be used to illustrate several issues regarding the minimalist guideline that language is economical.

(3)

Derivational Economy in Syntax and Semantics

(4)

Derivational Economy in Syntax and Semantics

(5)

Derivational Economy in Syntax and Semantics

Bošković (1998) provides an analysis of (3) where everything is indeed not equal in (3) and where (3a) and (3b) do not merely differ in an optional application of a single movement operation. Traditionally, it has been assumed that lexical insertion can only take place in overt syntax, otherwise the sound–meaning relation would collapse. Chomsky (1995) proposes a deduction of this assumption from Full Interpretation: suppose a lexical item like John is inserted in LF. The derivation would then crash due to the presence of phonological features of John, which LF cannot interpret. On the other hand, if John is inserted in PF, PF will not be able to interpret the semantic features of John. The only way for the derivation to proceed is to insert John before spell-out: PF then strips off the phonological features of John, with the semantic features of John proceeding into LF. This analysis allows LF insertion of lexical items, as long as they are phonologically null and the insertion occurs at the top of the tree, to satisfy the cycle. Bošković (1998) argues for an LF C-insertion analysis of French wh-in-situ. Wh-in-situ is in fact possible in French only in the contexts where LF insertion is in principle possible, namely, when the C is phonologically null and located at the top of the tree. (This is the case in (3) but not in (4)–(5). Not all dialects of French allow overt C examples like (5a).) Under this analysis, (3a) and (3b) do not merely differ in an optional application of wh-movement. In fact, the paradigm in (3)–(5) indicates that we cannot simply assume that French interrogative C optionally triggers wh-movement. If that were the case, we would not be able to enforce the +wh-movement option in (4) and (5). Bošković (1998) then argues that French interrogative C requires wh-movement. The difference between (3a) and (3b) is in the timing of C-insertion. In both cases, the C triggers wh-movement immediately upon its insertion (note French wh-in-situ shows locality-of-movement effects): however, the C is inserted overtly in (3a), hence it triggers overt wh-movement, and it is inserted covertly in (3b), hence it triggers covert wh-movement here. (4b) and (5b) are ruled out because the LF C-insertion derivation is blocked for independent reasons (in (4b) because the C is overt and in (5b) because of the cycle, i.e., because it does not take place at the top of the structure). The above discussion illustrates what one would expect to find in the cases involving seemingly optional movements: not everything is equal in the movement and non-movement options.

The French data in question also show that economy should be evaluated locally, without look-ahead. Chomsky (1995) argues that covert movement is more economical than overt movement, hence should be preferred to overt movement when both are in principle possible. The reason here concerns the general economy requirement to carry as little material as possible under movement. The underlying assumption is that movement takes place in order to license formal features. Given the requirement to carry as little material as possible under movement, we may then expect movement to affect only formal features that need licensing. Such movement would, however, cause a problem in PF under the assumption that scattered lexical items, where parts of a lexical item are located in different places due to feature movement, cannot be pronounced in PF. Overt movement then has to carry full lexical items for PF reasons. On the other hand, in LF, where the considerations of pronunciation do not apply, movement of parts of lexical items, i.e., only the relevant features, is in principle possible, hence forced by the economy requirement to carry as little material as possible under movement. LF movement is then favored over overt movement (when both are in principle possible) by the economy requirement to carry as little material as possible under movement, which deduces the Procrastinate principle of Chomsky (1993) (for further discussion of the requirement, which minimizes pied-piping, see Stateva, 2002, and Bošković, 2004; the latter applies it to quantifier float, arguing that quantifier float is obligatory due to the requirement in question whenever it is possible, the cases where it is not possible involving a different structure which disallows quantifier float).

Returning to French wh-in-situ, given these consideration, the LF C-insertion derivation should be more economical than the overt C-insertion derivation, since it triggers covert wh-movement, which is more economical than overt wh-movement. It seems then that (3a), the LF C-insertion derivation, should block (3b), the overt C-insertion derivation. However, this is the case only if look-ahead is allowed. C-insertion itself works in exactly the same way whether it is covert or overt; there is no reason to prefer one of the two options. It is true that later down the road one of these operations will lead to a more economical movement. However, if economy considerations must be evaluated locally, without look-ahead, rather than globally, with look-ahead, we get exactly the right outcome, allowing both (3a) and (3b). The data under consideration thus favor local over global economy, as one would expect from a system with efficient design. In fact, Collins (1997) argues for a conception of economy where economy is quite generally evaluated locally based only on immediately available structure and syntactic objects within a derivation; this conception excludes comparison of different derivations or look-ahead within a single derivation when it comes to economy considerations.

The guidelines of local economy have played a prominent role in syntactic theorizing within MP, even regarding fundamental questions like the nature of the interaction with the interfaces. MP assumes that the syntactic computation is driven by the need to eliminate from the computation objects that are illegitimate at the interfaces, since the interfaces cannot deal with them. The exact implementation of the need has taken various forms, from syntactic operations (like movement) taking place to eliminate features that would not be interpretable at the interfaces to movement taking place to make labeling of syntactic structure possible, under the assumption that labels are needed for interpretation (Chomsky, 2013). In this context, Epstein, Groat, Kawashima, and Kitahara (1998) raise an important question which is relevant to the issue of look-ahead, discussed above: How can a syntactic operation at some intermediate point of the derivation be driven by the notion of legitimacy at the interfaces if these interfaces access only the final PF and LF representations? Both GB and early Minimalism assumed that there is a single point of spell-out, where the derivation separates into the phonology and the semantics branch. A number of authors (e.g., Chomsky, 2000, 2001; Epstein, 1999; Epstein et al., 1998; Uriagereka, 1999) have, however, argued that spell-out may apply multiple times, arguing for a radically derivational nature of the computation where the interfaces access syntactic computations as the derivation proceeds without the mediation of PF and LF levels of representation. Syntactic derivation itself plays the central role in this approach, as there are no real levels of representation. This approach alleviates the look ahead problem regarding syntactic operations being driven by the notion of legitimacy at the interfaces since in the derivational multiple spell-out model the interfaces interface with the syntax throughout the syntactic derivation.

As noted above, when it comes to economy of derivation considerations within MP, the guiding principles are (a) do something only if you need to and (b) if you do need to, do it in the most economical/efficient way. The former has resulted in the last resort nature of movement, with the Inclusiveness Condition also having this flavor. The above discussion has also noted one instantiation of the latter: movement carries as little material as possible. We now turn to another illustration of the economy effect on movement where when movement does take place, it must take place in the most efficient way. In particular, it is as short as possible.2 The effect is illustrated by traditional superiority effects, which provide a rather strong confirmation of the idea of economy in language. Consider (6), whose structure prior to wh-movement is given in (7).

(6)

Derivational Economy in Syntax and Semantics

(7)

Derivational Economy in Syntax and Semantics

The GB account of the (6) was based on Chomsky’s (1973) Superiority Condition (see (8)), which says that if X and Y are candidates for a particular movement operation, and X asymmetrically c-commands Y, then X rather than Y undergoes the movement in question: Who is superior to what in (7), hence (8) requires that who undergoes wh-movement in (6) (there are many accounts of the Superiority Condition as well as attempts to reduce it to independently motivated principles—see, e.g., Cheng & Demirdache, 1990; Hornstein, 1995; Kayne, 1984; Lasnik & Saito, 1992; Pesetsky, 1982;Watanabe, 1992).

(8)

Derivational Economy in Syntax and Semantics

In MP, Superiority effects follow from general considerations of economy of derivation; in fact, they provide perhaps the most straightforward illustration of economy of derivation, which requires that every movement requirement be satisfied through the shortest movement possible. English interrogative C requires its Spec position to be filled by a wh-phrase, however this requirement is to be implemented. In (7), the requirement can be satisfied by moving either who or what. Since the former results in shorter movement, (6a) is preferred to (6b). (6b) is thus ruled out because the formal inadequacy of the interrogative C has not been satisfied in the most economical way, i.e., through the shortest movement possible.

The economy account is simpler hence conceptually preferable to the account based on (8). There are also empirical reasons that favor the economy account. Consider the following Bulgarian multiple wh-fronting (MWF) data:

(9)

Derivational Economy in Syntax and Semantics

(10)

Derivational Economy in Syntax and Semantics

All wh-phrases move to SpecCP in Bulgarian MWF. The wh-phrase that is highest prior to wh-movement (the subject in (9) and the indirect object in (10)) must be first in the linear order. This has been analyzed as a Superiority effect, which means the highest wh-phrase moves first, and the second wh-phrase either right-adjoins to the first wh-phrase (Rudin, 1988) or moves to a lower SpecCP (Richards, 2001). As discussed in Bošković (1999), like English +wh-C, Bulgarian +wh-C requires movement to its Spec. Additionally, there is a focusing requirement on wh-phrases, which is also satisfied by moving wh-phrases to SpecCP. Consider now constructions involving three wh-phrases. Assuming that the wh-phrase that moves first to SpecCP satisfies the requirement of the interrogative C to have a filled SpecCP, the wh-phrase that is highest prior to wh-movement must move first to SpecCP—this way the formal inadequacy in question is satisfied in the most economical way. The other wh-phrases then move to SpecCP to satisfy their focusing requirements. Whichever order the wh-phrases move in, the requirements are satisfied through the movements of the same length, i.e., they cross the same maximal projections. We would then expect that the highest wh-phrase will move first, and then the other two wh-phrases can move in any order. As discussed in Bošković (1999), the prediction is borne out: the highest wh-phrase must move first (hence is first in the linear order), the order of the second and the third wh-phrase is free. Thus, while the dative must move before (and precede) the accusative in (10), this is not the case in (11), where their order of movement is free. (Note the subject must be first.)

(11)

Derivational Economy in Syntax and Semantics

On the other hand, the account based on (8) will always force the higher wh-phrase to move first, which means the order of all three fronted wh-phrases should be fixed, reflecting their order prior to wh-movement; i.e., under this account (and the same holds for other accounts of Superiority cited above), we would expect the superiority effect from (10a–b), i.e., the strict ordering of the wh-phrases, to be maintained in (11a–b). These data thus provide an argument for the economy account of Superiority and for the general notion of economy in language. They also show that economy does not always rule out all but one option. When several options are equally economical, they are all allowed.

Bulgarian MWF may provide another argument for the shortest movement requirement. Consider again (9) under the multiple-specifier analysis, where each wh-phrase is located in a SpecCP. As discussed above, the nominative wh-phrase must move first, so that the strong +wh-feature of C can be licensed in the most economical way. Richards (2001) interprets (9) as indicating that the second wh-phrase must move to a lower SpecCP, i.e., it must tuck in under the already existing Spec of CP, not create a SpecCP on top of the already existing one. This can be interpreted as a result of the Shortest Move requirement, an analysis Richards actually argues for: the tucking-in movement is shorter than movement on top of the existing SpecCP, hence preferred by economy of derivation.

A final note on Superiority: the superiority effect is semantically conditioned. Consider (12):

(12)

Derivational Economy in Syntax and Semantics

Who, rather than what, must undergo wh-movement in (12a–b), a familiar Superiority effect. However, (12d) is grammatical in spite of what wh-moving over who. Importantly, (12d) is not ambiguous: it is acceptable only if who takes the matrix scope; it is unacceptable on the multiple indirect question reading, where both what and who are interpreted in the embedded SpecCP. Apparently, superiority effects are conditioned on the relevant wh-phrases being interpreted in the same interrogative SpecCP. It is not clear how to capture this effect since bringing in semantic considerations here appears to require a considerable look ahead: at the point of movement of what to SpecCP, which takes place before the matrix clause is built in (12d), (12b), and (12d) are in fact identical. It is not clear how to make this step of movement sensitive to what will happen later in the derivation and the final semantics of the examples in question (for discussion, see Fox, 1995, 2000; Golan, 1993; and Reinhart, 2006, who argue that only constructions with essentially the same semantic interpretation can be compared regarding economy of derivation).

We have seen above that under economy guidelines, movement takes place only when there is a need for it, and when it does take place, it takes place in the most economical way: it is as short as possible and carries as little material as possible. Furthermore, economy is evaluated locally on the basis of immediately available structure. These economy considerations leave little room for truly optional movements, since options are in principle allowed only when they are equally economical.

We now turn to the role of phases as an efficient design mechanism. The early research within the generative paradigm has noticed that syntactic dependencies can span only a limited amount of structure. In the current theory, the locality of syntactic dependencies is treated in terms of phases, the goal here being to have an optimal and efficient computational system. This is accomplished in the phase theory, combined with multiple spell-out, by limiting the number of syntactic objects and the amount of syntactic structure that the syntactic derivation is working on. This is achieved by transferring parts of syntactic structure to the interfaces during the derivation, the transferred parts not being accessible for further syntactic operations (Uriagereka, 1999), with phases determining the transfer points. This approach captures the successive-cyclic character of long-distance movement with a new twist. It is well known that long-distance movements do not take place in one fell swoop, but rather through a series of shorter movement steps.3 As noted above, in the phase system, phases determine when syntactic structure is sent to spell-out—in particular, what is sent to spell-out is the complement of a phasal head (Chomsky, 2000, 2001). Given that what is sent to spell-out is no longer accessible to syntactic computation, this means that a moving element needs to move to the phasal edge and out of the phasal complement before the complement is sent to spell-put. Successive-cyclic movement then must target phasal edges. In fact, given the ban on superfluous movement, i.e., last resort, unless there are reasons that are independent of the phase-related locality considerations just discussed for movement to stop in any other intermediate position, we would expect successive-cyclic movement to target only phasal edges, nothing else (the issue is controversial though; see Kang, 2014, for an approach where successive-cyclic movement indeed targets only phasal edges).4

A rather dramatic illustration of the ban on superfluous intermediate steps is provided by local subject questions. Consider the following paradigm (for discussion of the paradigm, see Bošković, 2016; Messick, 2016, and references therein).

(13)

Derivational Economy in Syntax and Semantics

(14)

Derivational Economy in Syntax and Semantics

(15)

Derivational Economy in Syntax and Semantics

(16)

Derivational Economy in Syntax and Semantics

(17)

Derivational Economy in Syntax and Semantics

(18)

Derivational Economy in Syntax and Semantics

Questions like (13) are sometimes assumed not to involve wh-movement (e.g., Carstens et al., 2016; Chomsky, 1986). There is, however, evidence that the wh-phrase in (13) does not remain in SpecIP. Taking (14a–b) to indicate that the hell can only modify wh-phrases in SpecCP, (14c) provides evidence that who in (13) does not stay in SpecIP. Consider also (15)–(16). Everyone can take scope over the subject in (16) but not in (15). This is unexpected if the subject in (15) could stay in SpecIP. (14)–(16) thus provide evidence that who does not stay in SpecIP in (13). The WUE data in (17)–(18) provide evidence that who does not even pass through SpecIP on its way to SpecCP. WUE allows quantifier float under wh-movement, as in (17). Importantly, (18), which is unacceptable in standard English, is also unacceptable in WUE. If the wh-phrase were to move to SpecIP on its way to SpecCP, the quantifier in (17) would be floated under movement to SpecIP, but (18) indicates that this is not possible. Based on this, McCloskey (2000) concludes the wh-phrase moves directly to SpecCP. But what happens then with the usual requirement that SpecIP be filled in English?

There are two approaches to the traditional EPP effect: (a) movement to SpecIP is driven by an inadequacy of the target (I), which requires a Spec; (b) the movement is triggered by a problem in the structure prior to the movement to SpecIP, i.e., by a problem in the base-generated position where who is located in SpecvP (see, e.g., Bošković, 2007; Chomsky, 2013; Epstein & Seely, 2006). Chomsky (2013) provides a labeling account of this issue (see Messick, 2016). He provides a labeling algorithm for determining the nature of an object formed by Merge when in the case where a head and a phrase are merged, the head projects (see below). When two phrases (i.e., non-heads) merge, labeling is possible if the elements in question undergo agreement, otherwise one of the two elements has to move. Since traces are ignored for labeling, the remaining element then determines the label. Consider then (13). Under the vP-internal subject hypothesis, who is merged with vP. Since we are dealing here with the merger of two phrases that do not agree, who needs to move away to resolve the labeling problem (the movement is not driven by a property of I). With subject wh-questions, the external argument must move anyway to SpecCP, vacating SpecvP. Wh-movement to SpecCP then enables the labeling of vP. The object formed by the merger of I and vP is labeled by I since this is the case of a head-phrase merger. Since a wh-phrase and an interrogative CP undergo agreement (for the Q-feature), CP is labeled by the feature-sharing of the wh-phrase and the interrogative CP. Chomsky’s (2013) system thus derives (13) via direct movement of the wh-phrase to SpecCP.

But why is it that wh-movement in (13) cannot even proceed via SpecIP, as indicated by WUE (17)–(18)? Economy of derivation—in particular, the ban on superfluous steps—is relevant here. Given the above discussion, this step of movement, i.e., movement to SpecIP, is not necessary in (13). As a result, given the ban on superfluous operations, it should not take place.5 Notice also that nothing would in fact go wrong under the above discussion if John in (19) moves to a position above SpecIP.

(19)

Derivational Economy in Syntax and Semantics

Bošković (2007) argues the reason why this does not happen is also economy, this time the shortest move requirement. John needs to move from its base-position, as discussed above (under Bošković’s analysis, John needs to move to a position that c-commands I); movement to SpecIP results in shorter movement than movement to a higher position, hence this derivation is preferred.

What the discussion of (13) and (19) indicates is that economy minimizes derivational steps (see the lack of movement to SpecIP in (13)) and requires them to be as short as possible. The guiding ideas behind these two have been applied to other domains. Thus, one case of the former is provided by Nunes (2004). Under the copy theory of movement, where movement leaves behind a copy rather than a trace (in accordance with the Inclusiveness Condition), a question arises which copies will be realized in PF. The issue is approached through copy deletion: copies that are not deleted are realized in PF. Nunes (2004) develops a system where the reason why the highest copy is typically realized in PF—this is not always the case; see, e.g., Bošković and Nunes (2006)—has to do with considerations of word ordering (see Kayne’s 1994 LCA) and, crucially, economy: Nunes shows that deletion applies in the most economical way, with the fewest instances of deletion applying if the highest copy fails to be deleted, hence receiving PF realization (when this option is in principle possible); this is basically another example of fewest steps, though rather different from the movement cases discussed above. Regarding shortest move, as discussed above, this economy notion is responsible for superiority effects, where in (7) the interrogative C attracts the closest wh-phrase to minimize the movement length. This guiding economy idea is also present in the labeling algorithm of Chomsky (2013), though in a rather different form. Chomsky (2013) argues that labeling is not an automatic part of the Merge operation, which combines syntactic objects, proposing an algorithm that provides labels (at the point when labeling is required, which is the phasal level for Chomsky). As noted above, the base part of the algorithm is that when a head and a phrase merge, the head determines the label of the resulting object. Thus, when arrive merges with that woman in (20), the object in question is verbal in nature, not nominal, because arrive, the head, projects).

(20)

Derivational Economy in Syntax and Semantics

What is behind this is Minimal Search, which in (20) searches for the first computational atom, i.e., a lexical item/head, which is arrive. This element then labels the result of the merger of arrive and that woman in (20). A similar mechanism is taken to be involved in agreement, where an agreeing head searches for the closest element it can Agree with, thus always agreeing with the closest element. Abstractly, these cases illustrate the same notion of economy as the shortest move from the paradigm in (6)–(7).

3. Derivational Economy in Semantics

In the previous section we showed that overt movement operations are constrained by economy considerations. Covert movement operations also exhibit such effects. Take, e.g., scopally ambiguous sentences involving two quantificational elements such as (21a). A standard analysis of the inverse scope reading (i.e., the reading paraphrasable as “for all movies x there exists some critic y, such that y likes x”) is that the DP every movie undergoes covert movement above the DP some critic (e.g., May, 1985), referred to as Quantifier Raising (QR). The LF for the inverse scope reading is (21b).

(21)

Derivational Economy in Syntax and Semantics

Like the movement operations discussed in the previous section, QR is subject to economy conditions. Thus, Fox (1995, 2000) proposes (22).

(22)

Derivational Economy in Syntax and Semantics

The intuition behind (22) is the same as with the last resort cases of overt movement discussed in Section 1: movement takes place only if there is motivation for it. However, the motivations for the overt and covert movement cases are distinct. For overt movement the motivation is to eliminate crash-inducing formal features, while the motivation for covert movement (i.e., scope shifting operations) is to give access to new semantic interpretations.

The evidence Fox gives in favor of (22) comes from elliptical constructions with quantificational elements. Fox begins with the observation that in Verb Phrase Ellipsis (VPE) constructions that involve two quantificational elements in both the antecedent clause (AC) and the elided clause (EC) the structure is only two ways ambiguous out of four possible readings. Either both the antecedent and the elided clause have surface scope or both have inverse scope. The readings where the two clauses mismatch in scope are unavailable, as (23) shows (elided material is given in <…>):

(23)

Derivational Economy in Syntax and Semantics

Fox accounts for this by appealing to an independently motivated constraint on ellipsis known as Parallelism. We can assume a simplified definition of Parallelism in (24).

(24)

Derivational Economy in Syntax and Semantics

(24) allows us to account for the observed pattern in (23), assuming that QR leaves a trace in its base position that is interpreted as a variable. In order for the AC to get an inverse scope reading as in (23a), the DP every teacher must undergo QR to a position higher than a boy. In order for Parallelism to be satisfied, a parallel application of QR must take place in the EC, moving every teacher to a structurally higher position than a girl. If no scope shifting operations take place in the AC (resulting in surface scope), as in (23b), then Parallelism can only be satisfied if likewise no scope shifting operations take place in the EC. The interpretations that require an instance of QR in only one of the AC or EC (23c–d) violate Parallelism, and hence are unavailable.

With this in mind, consider a minimally different example in (25). As in the previous example, both the subject and the object in the AC are quantificational DPs, but unlike the previous example, only the object in the EC is quantificational; the subject in this case is a non-quantificational DP, Mary. Crucially, unlike the previous example, (25) is unambiguous. It only allows the surface scope reading in the AC.

(25)

Derivational Economy in Syntax and Semantics

Fox argues that this follows from the interaction of the Scope Economy constraint in (22) and Parallelism in (24). For the inverse scope to obtain in the AC, every teacher must undergo QR. There are two potential representations to consider. The first, given in (26a), has every teacher undergoing QR in the AC, but not in the EC, where the DP stays in its base position. This representation satisfies Scope Economy as the application of QR of every teacher to a position over a boy results in an interpretation distinct from the surface scope. However, Parallelism is violated here. No parallel application of QR is present in the EC, hence the representation is ruled out. Now consider the representation in (26b). This representation satisfies Parallelism as there are parallel applications of QR of every teacher in the AC and the EC. As mentioned above, the application of QR in the AC also satisfies Scope Economy as it results in an interpretation distinct from the surface reading; however, the application of QR in the EC crucially violates Scope Economy as the movement of every teacher above the non-quantificational DP Mary does not result in a new interpretation, hence this representation is also ruled out. The interaction of these two constraints then allows us to correctly account for the unambiguous nature of (25).

(26)

Derivational Economy in Syntax and Semantics

Following Fox’s work, many authors have adopted the Scope Economy constraint and used it to account for a number of other restrictions on QR (e.g., Cecchetto, 2004; Mayr & Spector, 2011; Takahashi, 2010; Wurmbrand, 2013).

Many other economy constraints have been proposed to account for scope shifting operations. A number of authors have proposed economy constraints that favor structures where linear precedence indicates scope such that if A linearly precedes B at PF, then A scopes over B at LF, maximizing LF transparency in this respect (Bobaljik, 2002; Bobaljik & Wurmbrand, 2012; Diesing, 1997; Müller, 2002; Reinhart, 2006; Williams, 2003,). Take, for example, Bobaljik & Wurmbrand’s (2012) (27).

(27)

Derivational Economy in Syntax and Semantics

ScoT is a “soft” violable constraint, meaning that ScoT can be violated as a last resort. As we have seen previously, English allows inverse scope readings as in (21), where an object quantifier can scope over a subject quantifier even though in the surface word order, the subject precedes the object. Compare this to Japanese (28). In (28a) the inverse scope reading where the object scopes over the subject is impossible. When the object overtly moves to a position above the subject via scrambling as in (28), the reading is possible.

(28)

Derivational Economy in Syntax and Semantics

The difference between English and Japanese is tied to the ability to overtly move the object over the subject via scrambling. In English, there is no scrambling, so the only word order available is the one where the subject precedes the object. Because of this, ScoT can be violated as a last resort allowing for a mismatch between linear precedence and scope. In Japanese, on the other hand, scrambling can create word orders where the object precedes the subject overtly. So unlike English, ScoT cannot be violated as doing so would not be a last resort (see Bobaljik and Wurmbrand (2012) for detailed discussion).

We have seen that the semantic economy constraint Scope Economy has the same intuition as the Last Resort view of syntactic movement operations: an operation only occurs if there is motivation for it. We now turn to economy constraints that restrict binding possibilities. We will see that the intuition behind these constraints is similar to the Shortest Move constraint discussed above: long dependencies are blocked by more local dependencies. To illustrate, consider Fox’s (2000) Rule H (for similar proposals and extensions, see, e.g., Büring, 2005; Grodzinsky & Reinhart, 1993; Reinhart, 1983; Roelofsen, 2011; Rule H is also conceptually similar to Rizzi’s (1990) relativized minimality).

(29)

Derivational Economy in Syntax and Semantics

The effect of Rule H is illustrated by (30), where there are two bound pronouns. Consider the two LFs for (30) in (31). In (31a), the λ-expression appended directly below John binds both pronouns in the embedded clause in what is termed co-binding. In (31b) (which involves transitive binding), the λ-expression appended directly below John only binds the pronoun he; the pronoun his, in turn, is bound by the λ-expression appended directly below he. As both of these LFs result in the same interpretation, Rule H rules out the co-binding representation in favor of the transitive binding representation.

(30)

Derivational Economy in Syntax and Semantics

(31)

Derivational Economy in Syntax and Semantics

The evidence for Rule H again comes from possible interpretations of VPE constructions. Bound variables in elliptical constructions can be interpreted in two ways, as shown in (32). In (32a) (the so-called strict interpretation), the elided pronoun refers to John. In (32b) (the so-called sloppy reading), the elided pronoun is bound by Bill.

(32)

Derivational Economy in Syntax and Semantics

Now consider a VPE construction that takes a sentence like (30) with two pronouns as its antecedent. Assuming that both pronouns have the ability to be strict or sloppy, we expect the sentence to be ambiguous in four ways, but as indicated in the judgments in (33), only three of the four readings are available.

(33)

Derivational Economy in Syntax and Semantics

Recall that there is an independent constraint on ellipsis that forces the antecedent and the elided clause to be parallel regarding the position of variables and their binders. So if the pronouns in the elided clause are interpreted as bound by Bill (i.e., sloppy interpretation), then that binding relation must be matched by a binding relation in the antecedent. With this in mind, the LFs for the possible interpretations of the elided clause are given in (34) (assuming that the assignment function maps the index 1 to John).

(34)

Derivational Economy in Syntax and Semantics

For the LFs that involve bound variable interpretations (34b–d), there must be parallel binding relations in the antecedent in order to satisfy Parallelism. In the LF (34b), the binding relations between λ4 and he4 and λ3 and his3 are parallel to the binding relations in the LF of the antecedent in (31b). The binding between λ4 and he4 in (34c) is likewise parallel to the binding found in (31b). The binding between λ4 and his4 in (34d), on the other hand, is not parallel to any binding relation in the transitive binding LF in (31b). The only LF for the antecedent that would allow for a parallel binding with (34d) would be the co-binding representation in (31a), but such an LF is blocked by Rule H. Since a parallel LF for the antecedent cannot be constructed, the LF in (34d) is not licit; hence we correctly capture the fact that the interpretation in (33d) is ruled out. Without Rule H, an antecedent with parallel binding relations to (34d) could be generated, hence the LF would be licit and we would incorrectly predict that (33d) is a possible interpretation of the construction.

A number of other proposals to deduce or simplify binding theory by appealing to economy considerations have been made in the literature (e.g., Reinhart, 1983; Safir, 2004, 2014; Schlenker, 2005). Take, for example, Schlenker’s (2005) Minimize Restrictors! from (35).

(35)

Derivational Economy in Syntax and Semantics

This constraint essentially forces short definite descriptions over longer ones—a similar intuition to the Minimize Structure constraints (see Section 1). Minimize Restrictors! can account for Condition C of the Binding Theory (and its exceptions) if one assumes that pronouns are short definite descriptions. Take, for example, (36). (36a) is a Condition C violation where an R-expression (John) is c-commanded by a coreferent expression. Under (35), this construction is ruled out by the availability of (36b), where a pronoun is used in place of John, because the pronoun is a shorter description than the R-expression. Turning to (36c), (36c) is acceptable despite the idiot being c-command by a coreferent R-expression. This is also captured by (35). If a pronoun was used in place of the idiot, then information would be lost (i.e., the information that the speaker thinks John is an idiot), so the use of the idiot serves a pragmatic purpose that the shorter description cannot capture. Because of this, (35) does not force the use of the pronoun in this example, and (36c) is correctly predicted to be grammatical. If Condition C were to be applied blindly to (36c), it would incorrectly rule it out.

(36)

Derivational Economy in Syntax and Semantics

In summary, the considerations of economy of derivation require movement to take place for a reason (with both syntactic and semantic considerations playing a role here), also banning superfluous steps/operations. When movement does take place, it is as short as possible and carries as little material as possible. Economy should be evaluated locally, on the basis of immediately available structure. The locality of syntactic dependencies is also enforced by minimal search and by limiting the number of syntactic objects and the amount of structure accessible (or present) in the derivation.

Further Readings

Boeckx, C. (2007). Understanding minimalist syntax. Oxford: Blackwell.Find this resource:

Bošković, Ž. (1998). LF movement and the minimalist program. In P. N. Tamanji & K. Kusumoto (Eds.), NELS 28 (pp. 43–57). Amherst: University of Massachusetts.Find this resource:

Bošković, Ž. (2011). Last resort with move and agree in derivations and representations. In C. Boeckx (Ed.), The handbook of linguistic minimalism (pp. 327–353). Oxford University Press.Find this resource:

Chomsky, N. (1995). The minimalist program. Cambridge, MA: MIT Press.Find this resource:

Chomsky, N. (2000). Minimalist inquiries. In R. Martin, D. Michaels, & J. Uriagereka (Eds.), Step by step: Essays on minimalist syntax in honor of Howard Lasnik (pp. 89–155). Cambridge, MA: MIT Press.Find this resource:

Chomsky, N. (2001). Derivation by phase. In M. Kenstowicz (Ed.), Ken Hale: A life in language (pp. 1–52). Cambridge, MA: MIT Press.Find this resource:

Chomsky, N. (2013). Problems of projection. Lingua, 130, 33–49.Find this resource:

Collins, C. (1997). Local economy. Cambridge, MA: MIT Press.Find this resource:

Epstein, S. D., Groat, E., Kawashima, R., & Kitahara, H. (1998). A derivational approach to syntactic relations. New York: Oxford University Press.Find this resource:

Epstein, S. D., Kitahara, H., Obata, M., & Seely, D. T. (2015). Economy of derivation and representation. In M. Dikken (Ed.), The Cambridge handbook of generative syntax (pp. 487–514). Cambridge, U.K.: Cambridge University Press.Find this resource:

Fox, D. (2000). Economy and semantic interpretation. Cambridge, MA: MIT Press.Find this resource:

Fukui, N. (1996). On the nature of economy in language. Cognitive Studies, 3(1), 151–171.Find this resource:

Reinhart, T. (2006). Interface strategies: Optimal and costly computations. Cambridge, MA: MIT Press.Find this resource:

Safir, K. (2015). Syntax, binding, and patterns of anaphora. In M. Dikken (Ed.), The Cambridge handbook of generative syntax (pp. 515–576). Cambridge, U.K.: Cambridge University Press.Find this resource:

Schlenker, P. (2005). Minimize restrictor! (notes on definite descriptions, condition c and epithets). In Buiry, Haitink, & Maier (Eds.), Proceedings of Sinn und Bedeutung (p. 385–416).Find this resource:

References

Abels, K. (2003). Successive cyclicity, anti-locality, and adposition stranding. Unpublished doctoral dissertation, University of Connecticut, Storrs.Find this resource:

Bobaljik, J. D. (2002). A-chains at the PF-interface: Copies and ‘covert’ movement. Natural Language and Linguistic Theory, 20(2), 197–267.Find this resource:

Bobaljik, J. D., & Wurmbrand, S. (2012). Word order and scope: Transparent interfaces and the ¾ signature. Linguistic Inquiry, 43, 371–421.Find this resource:

Bošković, Ž. (1994). D-structure, theta-criterion, and movement into theta-positions. Linguistic Analysis, 24, 247–286.Find this resource:

Bošković, Ž. (1997). The syntax of nonfinite complementation: An economy approach. Cambridge, MA: MIT Press.Find this resource:

Bošković, Ž. (1998). LF movement and the minimalist program. In P. N. Tamanji & K. Kusumoto (Eds.), NELS 28 (pp. 43–57). Amherst: University of Massachusetts.Find this resource:

Bošković, Ž. (1999). On multiple feature-checking: Multiple wh-fronting and multiple head-movement. In S. Epstein & N. Hornstein (Eds.), Working minimalism (pp.159–187). Cambridge, MA: MIT Press.Find this resource:

Bošković, Ž. (2004). Be careful where you float your quantifiers. Natural Language and Linguistic Theory, 22, 681–742.Find this resource:

Bošković, Žo. (2007). On the locality and motivation of move and agree: An even more minimal theory. Linguistic Inquiry, 38, 589–644.Find this resource:

Bošković, Ž. (2008). On successive cyclic movement and the freezing effect of feature checking. In J. M. Hartmann, V. Hegedüs, & H. van Riemsdijk (Eds.), Sounds of silence: Empty elements in syntax and phonology (pp. 195–233). Amsterdam: Elsevier.Find this resource:

Bošković, Ž. (2011). Last resort with move and agree in derivations and representations. In C. Boeckx (Ed.), The handbook of linguistic minimalism. Oxford University Press.Find this resource:

Bošković, Ž. (2016). On the timing of labeling: Deducing comp-trace effects, the subject condition, the adjunct condition, and tucking in from labeling. The Linguistic Review, 33(1), 17–66.Find this resource:

Bošković, Ž., & Nunes, J. (2006). The copy theory of movement: A view from PF. In N. Corver & J. Nunes (Eds.), The copy theory of movement (pp. 13–74). Amsterdam: John Benjamins.Find this resource:

Büring, D. (2005). Bound to bind. Linguistic Inquiry, 36(2), 259–274.Find this resource:

Carstens, V., Hornstein, N., & Seely, D. T. (2016). Head-head relations in problems of projection. The Linguistic Review, 33(1), 67–86.Find this resource:

Cecchetto, C. (2004). Explaining the locality conditions on QR: Consequences for the theory of phases. Natural Language Semantics, 12(4), 345–397.Find this resource:

Cheng, L. L.-S., & Demirdache, H. (1990). Superiority violations. In L. Cheng & H. Demirdache (Eds.), MIT Papers in Linguistics 13 (pp. 27–46). Cambridge, MA: Department of Linguistics and Philosophy, MIT.Find this resource:

Chomsky, N. (1973). Conditions on transformations. In S. Anderson & P. Kiparsky (Eds.), A festschrift for Morris Halle (pp. 232–286). New York: Holt, Rinehart and Winston.Find this resource:

Chomsky, N. (1986). Barriers. Cambridge, MA: MIT Press.Find this resource:

Chomsky, N. (1993). A minimalist program for linguistic theory. In K. Hale & S. J. Keyser (Eds.), The view from building 20 (pp. 1–52). Cambridge, MA: MIT Press.Find this resource:

Chomsky, N. (1995). The minimalist program. Cambridge, MA: MIT Press.Find this resource:

Chomsky, N. (2000). Minimalist inquiries. In R. Martin, D. Michaels, & J. Uriagereka (Eds.), Step by step: Essays on minimalist syntax in honor of Howard Lasnik (pp. 89–155). Cambridge, MA: MIT Press.Find this resource:

Chomsky, N. (2001). Derivation by phase. In M. Kenstowicz (Ed.), Ken Hale: A life in language (pp. 1–52). Cambridge, MA: MIT Press.Find this resource:

Chomsky, N. (2013). Problems of projection. Lingua, 130, 33–49.Find this resource:

Collins, C. (1997). Local economy. Cambridge, MA: MIT Press.Find this resource:

Diesing, M. (1997). Yiddish VP order and the typology of object movement in Germanic. Natrual Language and Linguistic Theory, 15(2), 369–427.Find this resource:

Epstein, S. D. (1999). Un-principled syntax: The derivation of syntactic relations. In S. Epstein & N. Hornstein (Eds.), Working minimalism (pp. 317–345). Cambridge, MA: MIT Press.Find this resource:

Epstein, S. D., Groat, E., Kawashima, R., & Kitahara, H. (1998). A derivational approach to syntactic relations. New York: Oxford University Press.Find this resource:

Epstein, S. D., & Seely, T. D. (2006). Derivations in minimalism. Cambridge, U.K.: Cambridge University Press.Find this resource:

Fox, D. (1995). Economy and scope. Natural Language Semantics, 3, 283–341.Find this resource:

Fox, D. (2000). Economy and semantic interpretation. Cambridge, MA: MIT Press.Find this resource:

Fukui, N. (1996). On the nature of economy in language. Cognitive Studies, 3(1), 151–171.Find this resource:

Golan, Y. (1993). Node crossing economy, superiority and D-linking. Unpublished manuscript, Tel Aviv University.Find this resource:

Hornstein, N. (1995). Logical Form: From GB to minimalism. Malden, MA: Blackwell.Find this resource:

Grimshaw, J. (1997). Projection, heads and optimality. Linguistic Inquiry, 28, 373–422.Find this resource:

Grodzinsky, Y., & Reinhart, T. (1993). The innateness of binding and coreference. Linguistic Inquiry, 24(1), 69–101.Find this resource:

Grohmann, K. (2003). Prolific domains: On the anti-locality of movement dependencies. Amsterdam: John Benjamins.Find this resource:

Kang, J. (2014). On the lack of TP and its consequences. Unpublished doctoral dissertation, University of Connecticut, Storrs.Find this resource:

Kayne, R. (1984). Connectedness and binary branching. Dordrecht, The Netherlands: Foris.Find this resource:

Kayne, R. (1994). The antisymmetry of syntax. Cambridge, MA: MIT Press.Find this resource:

Lasnik, H. (1995). Case and expletives revisited: On greed and other human failings. Linguistic Inquiry, 26, 615–633.Find this resource:

Lasnik, H., & Saito, M. (1992). Move α: Conditions on its application and output. Cambridge, MA: MIT Press.Find this resource:

Law, P. (1991). Effects of head movement on theories of subjacency and proper government. Unpublished doctoral dissertation, MIT, Cambridge, MA.Find this resource:

May, R. (1985). Logical form: Its structure and derivation. Cambridge, MA: MIT Press.Find this resource:

Mayr, C., & Spector, B. (2011). Generalized scope economy—not too strong! Unpublished manuscript, ZAS and Institut Jean Nicod-CNRS-ENS-EHESS.Find this resource:

McCloskey, J. (2000). Quantifier float and wh-movement in an Irish English. Linguistic Inquiry, 31, 57–84.Find this resource:

Messick, T. (2016). Matrix subject questions and the nature of the EPP. Unpublished manuscript, University of Connecticut, Storrs.Find this resource:

Müller, G. (2002). On feature movement. In A. Alexiadou, E. Anagnostopoulou, S. Barbiers, & H.-M. Gärtner (Eds.), Dimensions of movement: From features to remnants (pp. 209–242). Amsterdam: John Benjamins.Find this resource:

Nunes, J. (2004). Linearization of chains and sideward movement. Cambridge, MA: MIT Press.Find this resource:

Pesetsky, D. (1982). Paths and categories. Unpublished doctoral dissertation, MIT, Cambridge, MA.Find this resource:

Reinhart, T. (1983). Anaphora and semantic interpretation. London: Croom Helm.Find this resource:

Reinhart, T. (2006). Interface strategies: Optimal and costly computations. Cambridge, MA: MIT Press.Find this resource:

Richards, N. (2001). Movement in language: Interactions and architectures. Oxford: Oxford University Press.Find this resource:

Rizzi, L. (1990). Relativized minimality. Cambridge, MA: MIT Press.Find this resource:

Roelofsen, F. (2011). Free variable economy. Linguistic Inquiry, 42(4), 682–697.Find this resource:

Rudin, C. (1988). On multiple questions and multiple wh-fronting. Natural Language and Linguistic Theory, 6, 445–501.Find this resource:

Safir, K. (1993). Perception, selection, and structural economy. Natural Language Semantics, 2, 47–70.Find this resource:

Safir, K. (2004). The syntax of anaphora. Oxford: Oxford University Press.Find this resource:

Safir, K. (2014). One true anaphor. Linguistic Inquiry, 45(1), 91–124.Find this resource:

Saito, M., & Murasugi, K. (1999). Subject predication within IP and DP. In K. Johnson & I. Roberts (Eds.), Beyond principles and parameters (pp. 167–188). Dordrecht, The Netherlands: Kluwer.Find this resource:

Schlenker, P. (2005). Minimize restrictor! (notes on definite descriptions, condition c and epithets). In C. Buiry, J. Haitink, & E. Maier (Eds.), Proceedings of Sinn und Bedeutung (pp. 385–416). Nijmegen, The Netherlands: Nijmegen Centre of Semantics.Find this resource:

Speas, M. (1994). Null arguments in a theory of economy of projection. In University of Massachusetts Occasional Papers 17:179–208. Department of Linguistics, University of Massachusetts, Amherst.Find this resource:

Stateva, P. (2002). Possessive clitics and the structure of nominal expressions. Lingua, 112, 647–690.Find this resource:

Takahashi, M. (2010). Case, phases, and nominative/accusative conversion in Japanese. Journal of East Asian Linguistics, 19, 319–355.Find this resource:

Uriagereka, J. (1999). Multiple spell-out. In S. Epstein & N. Hornstein (Eds.), Working minimalism (pp. 251–282). Cambridge, MA: MIT Press.Find this resource:

Watanabe, A. (1992). Wh-in-situ, subjacency, and chain formation. In MIT Occasional Papers in Linguistics 2. MITWPL, Department of Linguistics and Philosophy, MIT, Cambridge, MA.Find this resource:

Williams, E. (2003). Representation theory. Cambridge, MA: MIT Press.Find this resource:

Wurmbrand, S. (2013). QR and selection; covert evidence for phasehood. In S. Kiene & S. Sloggett (Eds.), Proceedings of the North Eastern Linguistics Society Annual Meeting 42 (NELS 42) (pp. 277–290). Amherst: University of Massachusetts.Find this resource:

Notes:

(1.) The tacit assumption here is that numeration formation itself is costless. This is natural; numerations, which are comprised of lexical categories, determine what we want or choose to say, which is not determined by the computational mechanism of the human language.

(2.) It is worth noting here that Fukui (1996) observes a number of cases involving the same notion of economy in physics and combinatorial mathematics.

(3.) The intuition here is that movement cannot be too long. It has also been argued that movement cannot be too short, which is referred to as antilocality (see, e.g., Abels, 2003; Bošković, 1994; Grohmann, 2003; Saito & Murasugi, 1999). In other words, movement needs to be just right, not too short and not too long (see Abels, 2003, for a suggestion to derive antilocality from Last Resort).

(4.) This approach to successive-cyclic movement eliminates a tension between two notions of economy: shortest move and fewest derivational steps. Consider successive-cyclic movement. If such movement keeps to shortest move, it will have more steps, and if it minimizes the number of steps, it will have longer moves. Chomsky (1993) resolved this tension by adopting Form Chain as the basic transformational operation. A long-distance movement operation has only one Form Chain operation, with shortest move minimizing chain links internal to that Form Chain operation. In later work the Form Chain operation was eliminated. The tension noted in this footnote has dissolved: shortest move can no longer apply internal to Form Chain, which has been eliminated. Rather, the steps of successive-cyclic movement are determined by phases, i.e., derivational spell-out. Movement must target phasal edges: there is no choice between fewest steps and shortest moves at phasal edges since the latter are now required. A one-fell-swoop movement across a phasal boundary is not an option.

(5.) What we are dealing with here is essentially a ban on movement to the subject position feeding wh-movement. In more traditional minimalist terms, we are dealing here with the ban on A-feature-checking movement feeding A’-feature-checking movement. Bošković (2008) argues this ban is much more general: no feature-checking movement can feed another feature-checking movement, which Bošković (2008) argues ultimately follows from economy considerations.