Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, LINGUISTICS ( (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: 22 May 2017

Scope Marking at the Syntax-Semantics Interface

Summary and Keywords

Natural language allows questioning into embedded clauses. One strategy for doing so involves structures like the following: [CP-1 whi [TP DP V [CP-2 … ti …]]], where a wh-phrase that thematically belongs to the embedded clause appears in the matrix scope position. A possible answer to such a question must specify values for the fronted wh-phrase. This is the extraction strategy seen in languages like English. An alternative strategy involves a structure in which there is a distinct wh-phrase in the matrix clause. It is manifested in two types of structures. One is a close analog of extraction, but for the extra wh-phrase: [CP-1 whi [TP DP V [CP-2 whj [TP…t­j­…]]]]. The other simply juxtaposes two questions, rather than syntactically subordinating the second one: [CP-3 [CP-1 whi [TP…]] [CP-2 whj [TP…]]]. In both versions of the second strategy, the wh-phrase in CP-1 is invariant, typically corresponding to the wh-phrase used to question propositional arguments. There is no restriction on the type or number of wh-phrases in CP-2. Possible answers must specify values for all the wh-phrases in CP-2. This strategy is variously known as scope marking, partial wh movement or expletive wh questions. Both strategies can occur in the same language. German, for example, instantiates all three possibilities: extraction, subordinated, as well as sequential scope marking. The scope marking strategy is also manifested in in-situ languages. Scope marking has been subjected to 30 years of research and much is known at this time about its syntactic and semantic properties. Its pragmatics properties, however, are relatively under-studied. The acquisition of scope marking, in relation to extraction, is another area of ongoing research. One of the reasons why scope marking has intrigued linguists is because it seems to defy central tenets about the nature of wh scope taking. For example, it presents an apparent mismatch between the number of wh expressions in the question and the number of expressions whose values are specified in the answer. It poses a challenge for our understanding of how syntactic structure feeds semantic interpretation and how alternative strategies with similar functions relate to each other.

Keywords: (in)direct dependency, expletive wh-phrases, wh copy construction, extraction, co-ordinate structure constraint, scope freezing, bound variable readings, selection, negation, polar questions

1. Scope Marking

The scope marking construction is presented, with two approaches to the phenomenon outlined—the direct and the indirect dependency approaches. Three phases of the research on this topic are demarcated in terms of these two approaches.

1.1. Three Phases of Research

The Scope Marking construction, also known as Partial Wh-Movement and Expletive-Wh Construction, was once considered an exotic phenomenon. Over the years, it has lost that status as more and more languages have been discovered to have this construction and as numerous studies have documented its properties. However, its centrality in our understanding of the syntax-semantic interface has not diminished. We might even say that increased awareness of its properties has led to increased appreciation of its theoretical significance.

An example of a scope marking construction from English is given in (1a). Pronounced with an intonation that marks the two questions as a single unit, (1a) admits answers such as (1b):


Scope Marking at the Syntax-Semantics Interface

There are two points worth noting about (1b). It is a single sentence, with expressions from the first question in (1a) forming the matrix clause and expressions from the second question forming the complement clause. Furthermore, the wh-phrase in the first question has no obvious representation in (1b), while the wh-phrase in the second question has its value specified, giving (1b) the status of an answer to the question.

There are two ways of analyzing the paradigm in (1), dubbed the direct and indirect dependency approaches in Dayal (1994). The direct dependency approach takes a sentence like (1a) to have a Logical Form (LF) like (2a) and the interpretation in (2b):


Scope Marking at the Syntax-Semantics Interface

This approach makes scope marking an analog of questions involving extraction: where does John think we should go? The matrix wh forms a chain with the embedded wh, with the first marking scope and the second providing content. Interpreting this LF requires nothing beyond whatever semantics one needs to interpret basic questions. Here we have adopted what is known as the Hamblin-Karttunen semantics of questions for demonstration, but this choice is not critical (see Dayal, 2013, 2016, in press, for issues related to the syntax-semantics map). The analysis in (2) is, in essence, the analysis of van Riemsdijk (1983). A variant within direct dependency involves replacing the matrix wh with the embedded wh as an instance of expletive replacement (Beck, 1996; Müller & Sternefeld, 1996). Direct dependency has also been characterized in terms of feature percolation or feature driven movement. The essential feature of all direct dependency accounts involves mechanisms by which the embedded wh-phrase (or phrases) have matrix scope, while the matrix wh has no semantic content because it is not an argument of the matrix predicate.

The indirect dependency approach takes a question like (1a) to have an LF like (3a), with the two questions adjoined under a higher CP projection:


Scope Marking at the Syntax-Semantics Interface

Each CP in the adjunction structure is interpreted as a question in its own right, as shown in (3b). CP-1 is a question over propositional variables, in keeping with the analysis of what as the internal argument of the CP-1 verb think. CP-2 is a question about locations, as expected of the wh-phrase where. The connection between the two wh-phrases is indirect. CP-2 forms the restriction on the propositional variable in CP-1, playing the same role that common nouns play in wh-phrases like which N. Since the restriction is covert, a closer correspondent would be monomorphemic wh-phrases like who and what, where the restriction to human/animate or non-human/inanimate entities is implicit. Here that implicit restriction is represented by the variable T. It is identified with the second question through lambda abstraction, followed by lambda conversion. The structure in (3a) translates to something like which proposition in the set denoted by ‘where should we go’ does John think? This is, in essence, the analysis of Dayal (1994). It differs crucially from the direct dependency approach in ascribing semantic content to the wh-phrase in CP-1 as the internal argument of the predicate think, in analyzing CP-2 not as a complement of the matrix verb but as the restriction of what, and in interpreting the wh-phrases in CP-2 locally within CP-2. There are variants of this version of the indirect dependency accounts, which will be discussed in section 1.2.

As can be seen, both approaches end up with the same final denotation and therefore derive (1b) as a possible answer to (1a). As such, they are both equally plausible accounts of scope marking. Section 2 shows that substantive differences between the two emerge on closer inspection. For now, based on this simple equivalence, the direct and indirect dependency approaches will be used as a metric for characterizing research on this topic since the mid-1980s.

We can take the first phase of research to begin, naturally, with the first account of the phenomenon, namely the direct dependency account of van Riemsdijk (1983), and to end with the introduction of indirect dependency in Dayal (1994). During this phase scope marking had the status of an exotic phenomenon, restricted to a handful of languages and dialects: German (van Riemsdijk, 1983), Frisian (Hiemstra, 1986), Hungarian (Kiss, 1991), Romani (McDaniel, 1989), Bangla (Bayer, 1990), Iraqi Arabic (Wahba, 1991), and Hindi (Davison, 1984; Gurtu, 1985; Mahajan, 1990; Dayal, 1990, 1991, 1994). Crucially, these studies focused on cases of scope marking in which CP-2 was syntactically integrated with the first and displayed clear evidence of subordination. The German example (4a) is illustrative because CP-2 does not display the root clause property of having the verb in second position. The Hindi example (4b) illustrates the in-situ variant, with the appropriate scope configuration arising only at LF:


Scope Marking at the Syntax-Semantics Interface

The second phase of research on this topic is marked by two developments. On the one hand, the empirical domain of inquiry expanded considerably. Subordinated scope marking of the kind seen in (4) was noted in several unrelated languages. Additionally, sequential scope marking of the kind given in (1a) was shown to exist in English (Dayal, 1996) and German (Reis, 2000). This expanded empirical domain opened up the possibility of thinking of scope marking as a universally available option in natural language and of treating cross-linguistic variation in terms of direct versus indirect dependency in scope marking. Early proponents of the latter idea include von Stechow (2000) and Beck and Berman (2000), who argued that indirect dependency was appropriate for Hindi while direct dependency was right for German. The end of this phase is marked by Bruening (2004) with a rather dramatic presentation of both possibilities within a single language. The following examples are from Passamaquoddy:


Scope Marking at the Syntax-Semantics Interface

Bruening argues that (5a) uses the wh-expression keq(sey) ‘what,’ which is used to question propositions. Thus it fits in with the indirect dependency approach. The one in (5b) uses the wh-quantifier tan and is an expletive. On the basis of evidence from operator agreement in the language, Bruening further claims LF movement of the embedded wh tan to matrix position. This fits in with the direct dependency approach. He concludes, “Passamaquoddy is one of the few languages where scope marking has been described that actually uses a direct dependency; German, Hindi, Hungarian, etc., would all have to use an indirect dependency in their scope marking constructions.” (Bruening, 2004, p. 294)

The third phase is marked by the view, advocated in Dayal (2000), that there is no variation across languages in the type of dependency exhibited by scope marking. There is, however, room for variation in its syntactic expression. Later studies of German scope marking, which had been presented as an exemplar of direct dependency by Beck and Berman (2000) and von Stechow (2000), were shown to be better explained as instances of indirect dependency (Klepp, 2002; Felser, 2001, for example). Bruening (2006) re-analyzes the tan construction in (5b) as a variant of a different construction, the wh copy construction, rather than of scope marking. The scope marking construction has an invariant wh in CP-1, while the wh-copy construction repeats the same wh-phrase in both clauses. Classifying (5b) as the wh copy construction, Bruening then sides with the view that scope marking is to be thought of in terms of indirect dependency generally. The same is argued by Legate (2011) in relation to Walpiri and by Lipták and Zimmermann (2007) for Hungarian. Although there remain a few holdovers, it would be fair to say that the standard view of scope marking at this point is that it involves indirect dependency across languages.

1.2. The Syntax of Direct and Indirect Dependency

We will now take a closer look at the syntactic commitments of the two approaches, starting with the direct dependency approach. van Riemsdijk (1983) treated the matrix wh was in German as an expletive generated in matrix Spec position. As indicated earlier, the particulars of direct dependency analyses have changed over the years, in keeping with changes in the theory of wh movement and scope, but the characterization of the matrix wh as semantically vacuous has remained the defining feature of a direct dependency account. One advantage of treating the expletive wh as base-generated in Spec position is that it explains the following paradigm in German (von Stechow & Sternefeld, 1988):


Scope Marking at the Syntax-Semantics Interface

In contrast to the simple multiple wh question in (6a), the CP-1 of a scope marking construction requires was to be in Spec position rather than in the preverbal position where direct objects are generated. However, Klepp (2001) argues against this conclusion, noting that some speakers of the younger generation prefer the order in (6b). Klepp herself sees this as evidence for an indirect dependency account of German scope marking, but it is worth noting that the possibility of having the scope marker in the direct object position need not argue against direct dependency. Mahajan (2000) generates the wh in argument position but explains the contrast in (6) differently. He sees the need for was to appear in Spec as being due to its non D-linked status. Mahajan presents his account as a cross between direct and indirect dependencies, as do Fanselow and Mahajan (2000) and Horvath (1997), the latter for Hungarian scope marking. However, given the terms in which we have defined the difference between direct and indirect dependency, their accounts fall within direct dependency (see Dayal, 2000 for further discussion). Manetta (2010) proposes something similar while explicitly arguing for a direct dependency account. It is not strictly speaking required in a direct dependency account, then, that the scope marker be generated in Spec position.

The syntactic manifestation of indirect dependency is also varied. The original account in Dayal (1994) posited an adjunction structure of the kind shown in (3a). It emphasized the status of the matrix wh as the direct object, taking the cue from a corresponding declarative structure in which there is a pronominal in direct object position yeh ‘this.’ The scope marking construction in (4b) differs from (7a) in that it involves LF movement of kyaa ‘what.’ This is shown schematically in (7b)–(7c):


Scope Marking at the Syntax-Semantics Interface

Like Hindi, German has a pronominal counterpart es ‘this’ that occurs in preverbal position. Interestingly, was and es cannot co-occur even though they occupy different positions in the clause at the surface. German also places finite complements in post-verbal position but it differs from Hindi in having overt wh movement and overt V to C movement.

Herburger (1994) presents an analysis of German in which the scope marker is the head of a DP that takes CP-2 as its complement: [DP what [CP-2 who Mary has spoken]]. It is merged as an object in the matrix clause. The CP is later obligatory extraposed and adjoined to the right, and the remnant DP undergoes wh-movement from object position to the spec position, as in (8). Herburger sees this as an argument in favor of indirect dependency:


Scope Marking at the Syntax-Semantics Interface

Dayal (2000) capitalizes on these two possibilities to account for differences between subordinated and sequential scope marking with respect to variable binding. German (9a) represents adjunction as shown by V2 in CP-2, while (9b) shows subordination:


Scope Marking at the Syntax-Semantics Interface

Binding is not possible in (9a) because CP-2 has to undergo lambda conversion in order to function as the restriction of was (cf. 3). It violates a general ban against lambda conversion if it leads to the binding of free variables. In (9b) no lambda conversion is involved so binding is possible. Pafel (2000) offers a detailed discussion of scope within the direct dependency account while Lahiri (2002) presents an explicit account of the semantics involved in interpreting (9b) within indirect dependency.

Note that sequential scope marking can only be handled through indirect dependency. Even if one could argue for a syntactic reanalysis of the two questions of the sequential scope marking structure into a single question with subordination, the contrast in (9) argues against conflating the two structures. This establishes the existence of indirect dependency in natural language. The question to settle is whether direct dependency is also possible.

To return to the contrast in (9), it shows that an indirect dependency account is not tied to adjunction structures like (3a). In fact, it provides empirical evidence favoring the possibility of more than one syntactic structure—adjunction for sequential scope marking and DP complementation and extraposition for subordinated scope marking. Dayal (2000) shows that an indirect dependency account is possible even if a scope marker is merged in matrix Spec position, as suggested by proponents of the direct dependency approaches to explain the paradigm in (6). All that is required is that it be interpreted as the internal argument of a proposition selecting matrix verb. The take-home message is that the semantics of scope marking is fixed but its syntax is subject to variation. The view of variation presented in Dayal (2000) is influenced by Reis (2000), which shows the diachronic development of German scope marking from parenthetical constructions.

This might also be the place to mention languages like Walpiri that have different lexical items for questioning over objects (nyiya ‘what’) and propositions (nyarrpa ‘how’). The one used in scope marking is the one that is used to question over propositions. This observation is attributed to Ken Hale and Maria Bittner in Dayal (1994). Legate (2011) presents a detailed analysis of Walpiri scope marking where she makes the following points. Walpiri nyarrpa ‘how’ has a very limited distribution and cannot be considered the default wh. She further argues that it cannot be an expletive, because Warlpiri is a pro-drop language that gives no independent evidence of having expletives. She also notes that Walpiri finite clauses are islands for extraction. In light of these facts, she opts for an indirect dependency approach built on a syntactic account of the kind proposed by Herburger (1994). The analysis of Russian scope marking by Stepanov (2001) is also relevant to the issue of which wh expressions can occur in scope marking.

Other variants of the syntax of indirect dependency include Felser (2001) for German. She argues that the scope marking construction can be analyzed in terms of a predication relation between the scope marker and the embedded clause. She suggests furthermore that the scope marker and the embedded wh-clause both originate as complements of a matrix verb. The scope marker originates as an object in the spec of VP where it gets its theta-role from the matrix verb, whereas the embedded clause is introduced as an unselected complement of the matrix verb. Later in the derivation, the scope marker undergoes wh-movement to the spec of matrix CP. The dependency between scope marker and wh in the embedded clause thus is indirect, mediated as it is through the subject-predicate relation. Sulaiman (2014) analyzes Syrian Arabic along the same lines. This line of analysis has some similarities with the syntactic proposal in Manetta (2010), though the latter is cast in a direct dependency account.

2. Properties of Scope Marking

Some of the properties that scope marking is known to have are examined: those for which reasonable explanations exist in both the direct and the indirect dependency approaches and those that can be handled under only one approach, to the exclusion of the other. The list of properties discussed here is not exhaustive, and readers are requested to follow up on the original articles to get the full picture.

2.1. Neutral Properties

Scope marking has interesting selectional properties that have been noted since the earliest studies. The matrix verb must be able to take [-wh] complements, but the actual complement must be [+wh]. The German paradigm in (10) illustrates this: (10a) is unacceptable because the complement clause is not [+wh], while (10c) is unacceptable because the matrix verb selects only [+wh] complements:


Scope Marking at the Syntax-Semantics Interface

Under the direct dependency approach, the examples above would have the LFs in (11):


Scope Marking at the Syntax-Semantics Interface

(11a) is unacceptable because there is no wh in the embedded clause with which the matrix expletive wh can associate; (11c) is unacceptable because of competing demands on the embedded wh. It can either satisfy the selectional restrictions of the matrix verb by being interpreted locally or it can satisfy the need of the expletive wh for association with a contentful wh-phrase. Note that this explanation applies to subordinate scope marking only. A different explanation would be needed for sequential scope marking.

Under the indirect dependency approach, these results also follow straightforwardly. Consider the semantic types of the scope marker and the postverbal finite CP-2 in (10a)–(10c):


Scope Marking at the Syntax-Semantics Interface

We know that the verb ‘think’ takes propositional complements, so we take was ‘what’ to quantify over propositional variables in (10a). Its restriction must therefore be one level higher: a set of propositions. CP-2 in (10a) is a simple proposition; (10b) has appropriate types and is acceptable. In (10c), the matrix verb takes a question (sets of propositions) as its internal argument. Because the question in CP-2 denotes a set of propositions, not a set of sets of propositions (or equivalently a set of questions), we again have a mismatch. Thus, the seemingly arbitrary paradigm has a well-founded explanation here as well. And this explanation applies to sequential as well as subordinated scope marking, both of which show the same restrictions.

Note that under both approaches, the matrix wh is invariant: the default wh for the language in direct dependency; the wh that the language uses for questions over propositions in indirect dependency. In both approaches, there is no restriction on the type of the wh expression that occurs in the second question. The picture is less clear-cut for direct dependency when CP-2 has multiple wh-phrases because it involves having a single expletive with multiple associates. Crucially, it is not possible to interpret one wh-phrase locally and one with matrix scope, an option the direct dependency approach makes available. However, this problem has been addressed in different ways by various authors. See also related discussion of (18) in Section 2.2.

2.2. Properties Favoring One Type of Dependency

Some properties cannot be handled in the direct dependency approach, at least not without extra stipulations. Rather than provide a detailed analysis, the core of the explanations will be presented; the reader should consult the original articles for more details. A potential argument against indirect dependency is also presented The references mentioned in this section provide a multiplicity of perspectives , so readers can come to their own conclusions.

The first such property is what has been called anti-locality. von Stechow and Sternefeld (1988) show that was must bind a wh-phrase but the wh-phrase cannot come from the same clause:


Scope Marking at the Syntax-Semantics Interface

Because there is no connection between an expletive was and a complement clause in the direct dependency approach it is unclear why a clause-mate wh could not play the same role that an embedded wh plays. Indirect dependency, on the other hand, predicts the unacceptability of (13) given that was cannot be the internal argument of come in (13). And any wh expression that is indirectly associated with was must occur in its restriction, that is, in the finite clause that is referred to here at CP-2. See Beck and Berman (2000), Dayal (2000), von Stechow (2000), and Fanselow and Mahajan (2000) for discussion of anti-locality from different perspectives.

Another property that favors indirect dependency comes from the nature of parasitic gaps licensed by scope marking. Sabel (2000) notes examples like (14a), where the gap is of a propositional type. We draw on the discussion of the following paradigm in Dayal (2000) and note also that Sabel’s is a direct dependency account:


Scope Marking at the Syntax-Semantics Interface

Under indirect dependency it is expected that was should license a propositional type gap but not an individual type gap. This contrasts with the corresponding extraction structures, which show the opposite pattern of grammaticality. See also Horvath (1997) and Fanselow and Mahajan (2000). To the best of our knowledge this paradigm has not been taken up in the recent literature.

The third property favoring indirect dependency is the possibility of conjoined question complements, as in the German question-answer paradigm in (15) from Höhle (2000). The latter will not be characterized as falling in squarely with either the direct or the indirect dependency approach. The reader should consult the original article to get a sense of the development of theoretical interest in the topic of scope marking and for several relevant facts related to it:


Scope Marking at the Syntax-Semantics Interface

Dayal (2000) sees this as following from indirect dependency because a conjoined question denotes a set of propositions, as required of the restriction on the matrix was. This data does not appear to have been discussed in the direct dependency literature, where it would present a potential violation of the coordinate structure constraint.

A core feature of indirect dependency is that the second question is interpreted exactly as it would be interpreted if it were a root clause. This feature explains certain facts that can be characterized as scope freezing. Reis (2000) observes that the German extraction structure (16a) is ambiguous between a plausible and an implausible reading, while the scope marking structure in (16b) has only the implausible reading:


Scope Marking at the Syntax-Semantics Interface

von Stechow (2000) describes the plausible reading as: For which place x, in the agent’s belief worlds w’ is Fox more popular at x in w’ than Fox is popular at x in the real world w? The implausible reading is one where the levels of popularity at x are locked into the same world: For which place x, in the agent’s belief worlds w’ is Fox more popular at x in w’ than Fox is popular at x in w’? Note that the monoclausal (16c) has only the implausible reading because there is only a single world available, the actual one. The judgement for (16b), then, is as expected under indirect dependency. von Stechow takes on this challenge for direct dependency. He explains it as essentially an intervention effect due to the comparative morpheme -er, a quantifier, blocking the covert wh-movement of wo ‘where’ to matrix CP. This is based on the filter in Beck (1996), which only applies to covert movement. As such, the filter does not apply to (15a), which has overt movement. Note that this explanation only applies to subordinated scope marking but the effect is also visible in sequential scope marking.

Another case of scope freezing is discussed in Lahiri (2002). Kroch (1998) noted that questions like (17a) are ambiguous between a referential reading (what is the size of the set of books that Bill said John read?) and an amount reading (what is the number such that Bill said John read that many books?). Lahiri (2002) makes the important observation that Hindi scope marking structures have only the amount reading. That is, they have the reading that monoclausal amount questions have. The contrast can be demonstrated by the English extraction and sequential scope marking structures in (17a)–(17b):


Scope Marking at the Syntax-Semantics Interface

Lahiri argues that the scope freezing effect follows from an indirect dependency approach because the question in (17b) effectively asks: which proposition in the set denoted by how many books did John read? is such that Bill said that proposition.

This paradigm had until recently been thought to be a problem for direct dependency but Manetta (2010) suggests otherwise. She adopts Reinhart’s (1998) choice functional analysis of long-distance dependencies, where the wh-expression stays inside the embedded clause. She suggests that overt extraction allows for ambiguity while covert scope taking does not. Dayal (2017), however, shows that this suggestion does not hold up under scrutiny. She draws attention to a language like Japanese, which mirrors precisely the type of LF that Manetta posits for scope marking, one where the wh-expression remains in the embedded clause and is bound at a distance by the Q-binder in matrix Spec CP:


Scope Marking at the Syntax-Semantics Interface

The crucial point about examples like (18) is that they do not display the relevant scope freezing effect: (18) is ambiguous in exactly the way that the corresponding English extraction question in (17a) is. Dayal therefore concludes that Lahiri’s claim remains unchallenged. Note that scope freezing is seen also in sequential scope marking for which we must anyway resort to indirect dependency.

In fact, as pointed out by an anonymous reviewer and confirmed by Diti Bhadra (p.c.), the distinction under consideration can be illustrated in Bangla, a language that has two distinct strategies for questioning into finite complements. The scope marking strategy in (19a), which has the complement on its right, can be compared to the alternative in (19b) where the complement comes to the left and is embedded under the complementizer bole. As Dasgupta (1980) has established, the complementizer bole forces wide scope for embedded wh-phrases:


Scope Marking at the Syntax-Semantics Interface

While answers to both questions provide values for the embedded wh expression, (19a) and (19b) are not semantically equivalent. The scope marking structure only has an amount reading while the bole structure is ambiguous between an amount and a referential reading. This, then, provides further confirmation that scope freezing is due to the semantics of indirect dependency, and cannot be attributed to mechanisms for giving wide scope to embedded wh in-situ.

A final property of scope marking that favors indirect dependency has to do with examples like (10c), which disallow [+wh] selecting verbs in CP-1 of a scope marking construction. Horvath (1997) notes that such verbs are possible if CP-2 has more than one wh expression. She sees this as evidence for direct dependency: one wh satisfies the selectional restrictions of the matrix verb, the other takes matrix scope, via association with the expletive. Dayal (2000), however, notes that this is only possible if one wh is stressed:


Scope Marking at the Syntax-Semantics Interface

Dayal points to the account of echo questions in Comorovski (1996) and Dayal (1996). In brief, a multiple wh question with one echo wh denotes a set of questions of the kind shown in (20d). The type mismatch, which is the source of the unacceptability of (20a), is now resolved. This point is further discussed in Lahiri (2002). Dayal (2016) also discusses cases of scope marking where CP-2 is an echo question with one wh and one universal quantifier.

Another interesting argument comes from the possibility of scope marking with adjunct clauses in Hungarian, discussed by Lipták and Zimmermann (2007):


Scope Marking at the Syntax-Semantics Interface

Cases like these are interesting because the scope marker is already of the form what N. This raises the question of the relationship between CP-2 and the head noun iizenetet ‘message.’ They also give examples involving relative clauses. Lipták and Zimmermann argue that the direct dependency approach runs into a problem because the scope markers in these cases can by no means be considered expletive elements. Instead, they are full-blown argumental NPs/DPs, with a lexical meaning of their own. They provide instead an account within the indirect dependency approach, giving an explicit compositional semantics for it.

We will end this section by noting that the paradigm in (6a)–(6b) argues in favor of direct dependency, at least in languages with overt wh movement, though as discussed earlier, it is possible to account for the pattern within indirect dependency as well.

2.3. Properties Up for Debate

Let us now consider some properties that are not definitively settled by either approach. One involves CP-2 being a polar question. Dayal (1994) notes that this is possible in Hindi and that in such cases, the polarity is confined to CP-2. English sequential questions are noted in Dayal (1996, 2000) to have the same property:


Scope Marking at the Syntax-Semantics Interface

This is as expected under indirect dependency where CP-2 denotes the set {we should stay, we should not stay}. The problem is that polar questions are not acceptable as CP-2 in many languages, most notably German and Hungarian. Their unacceptability is seen by the proponents of direct dependency as evidence for that approach. Under this view, the yes/no operator cannot undergo wh-movement at LF (Beck & Berman, 2000; von Stechow, 2000). To see why neither approach can claim full explanatory power for this phenomenon, interested readers should familiarize themselves with the arguments and counterarguments given in the articles mentioned, as well as in Lahiri (2002), Horvath (1997) and Reis (2000). One related data point that has not been taken up in the literature is the unacceptability also of the German wh-phrase inwiefern ‘in what way’ as the associate of was (Beck & Berman, 2000).

Another aspect of scope marking that remains unsettled involves examples like (23), where there is more than one level of embedding. In such cases, each embedded clause must have a scope marker. One can frame the problem with reference to the potential LF in (23c) and ask why such a derivation is blocked:


Scope Marking at the Syntax-Semantics Interface

Note first of all, that both approaches have good explanations for the grammatical version in (23b) but they differ in their explanations for (23a). The direct dependency approach holds that the unacceptability follows from the fact that was (and its correlate in other languages) is an expletive that cannot be extracted. The indirect dependency approach suggests that such extraction is blocked because of the special relationship between was, which is the internal argument of the verb in CP-2, and its associate CP-3. An interesting wrinkle on this point is that the facts are less stable in some languages than in others. Some German speakers, for example, are willing to accept structures like (23a) while Hindi speakers resist it more uniformly. Beck & Berman (2000) provide a list of authors who do and do not consider structures like (23a) ungrammatical, to which the reader is referred. It is worth mentioning a point made by an anonymous reviewer in this connection. It seems that the Hindi counterpart of (23b) is possible but only if the scope marker is stressed in a way that supports long-distance scrambling, an option that is available in Hindi but only in specific discourse contexts. Under either approach, the difference between the variable status of such structures lies in the variable status of finite complements as islands for extraction.

Yet another open issue relates to the class of predicates that licenses scope marking. It has already been noted that a standard scope marking construction requires a predicate in CP-1 that selects [-wh] complements, but there are restrictions within this class that are poorly understood at this time. Bridge predicates and verbs of saying/thinking are the best, but there is cross-linguistic variation when we move out of the core group. Strong factive verbs, for example, are ruled out in German (though not in Hindi), which Müller and Sternefeld (1996) see as evidence for direct dependency. Factive verbs are islands for extraction and if scope marking involves covert extraction, this is to be expected. However, Reis (2000) makes the important observation that German sequential scope marking, where extraction could not play a role, does not allow such predicates either (see also related discussion in Dayal, 2000; Lahiri, 2002, among others).

The biggest open problem in scope marking, however, remains its resistance to negation, put on the theoretical center stage by Rizzi (1992) for German (24a). However, Höhle (2000) attributes this observation to earlier work by Kiss (1991):


Scope Marking at the Syntax-Semantics Interface

Rizzi’s account of (24a), within a direct dependency approach, treats this as a negative island effect. The precise details have evolved over the years, but the basic insight is to locate the problem in the interference by negation in the path between the wh expressions. Dayal (2000) notes that even if this general approach were correct, it would not necessarily be an argument for direct dependency. Negation could still play a blocking role, but it would block the path of CP-2, not the wh inside it, and the position inside DP that it is associated with: [was tCP-2 ], given the proposal in Herburger (1994). Dayal (1994), however, argued against a syntactic explanation for (24a) and suggested a pragmatic account instead. Under the indirect dependency view of the problem, the argument of verbs in the negative like not think is already restricted to a contextually relevant set. If the restriction T is already specified, CP-2 cannot replace it via lambda conversion. Although Dayal later revises her syntactic assumptions about (24a), her explanation for negation remains relevant, given a similar effect in sequential scope marking structures like (24b). Recall that the variable binding facts discussed earlier establish that the two clauses are not syntactically integrated into a single subordinated structure. Thus the issue to settle, once again, is not whether a pragmatic account of the negation facts is needed, but whether a syntactic account is needed in addition to it.

3. Looking Forward

Some further issues are discussed without specific reference to the type of dependency they may or may not give evidence for. These issues are of general relevance to the topic and worthy of further investigation. Because of space constraints, existing studies in these new domains will be highlighted rather than given full accounts.

3.1. Discourse

The discussion of negation provides a good transition to a discussion of the discourse properties of scope marking. Assuming that languages may have more than one way of questioning out of embedded clauses, are these options pragmatically constrained? Herburger (1994) observed a subtle meaning distinction between German extraction and scope marking structures. The first leaves it open whether the speaker accepts the presupposition behind the embedded question, a de dicto reading as she puts it, while the latter implies that the speaker is committed to it, a de re reading. Dayal (2000) illustrates this with reference to English (see also von Stechow, 2000; Reis, 2000):


Scope Marking at the Syntax-Semantics Interface

Lahiri (2002) compares the German scope marking structure (26a) with the corresponding extraction structure (26b) (the observation goes back to Herburger, 1994). He notes that the proposition presupposed by the embedded clause, Rosa kissed someone, is interpreted as the speaker’s belief, not as the agent George’s in (26a). This is not the case in (26b):


Scope Marking at the Syntax-Semantics Interface

While the authors previously mentioned discuss these effects in terms of direct/indirect dependency the point worth highlighting here is that the study of discourse appropriateness is independently significant. In addition to extraction, scope marking also alternates in some languages with the wh copy construction mentioned earlier. While the copy construction shares many properties with extraction, the resistance to negation is a common feature of scope marking and the wh copy construction. These issues seem worthy of further investigation.

3.2. Cross-linguistic Variants

We have referred to extraction and the copy construction as relevant alternatives to be used in situating the study of scope marking. We give an example of the copy construction from German in (27):


Scope Marking at the Syntax-Semantics Interface

There are other variants that have been noted as involving scope marking. The most oft quoted example is Malay (Bahasa Melayu), discussed by Cole and Hermon (2000):


Scope Marking at the Syntax-Semantics Interface

Cole and Hermon argue that this may be a case of a null expletive wh in matrix Spec that is replaced by the contentful wh in the embedded clause, as predicted by the direct dependency approach. This example is best handled by direct dependency, but we would pursue its study in relation to other examples of clausal pied-piping of the kind seen in (29a) and discussed by Ortiz de Urbina (1989, 1993) and Arregi (2003). Arregi makes a case for separating Basque clausal pied-piping from scope marking with respect to Herburger’s distinction between de dicto and de re readings:


Scope Marking at the Syntax-Semantics Interface

Another interesting construction, dubbed wh-slifting by Ross (1967), is illustrated in (29b). Haddican et al. (2014) note parallels with scope marking but show that there are also significant differences in how the two relate to discourse.

A consideration of the literature on these alternatives to extraction, then, establishes that there are several interesting properties they share. This may prompt us to think of wh scope in different ways. However, there should be no expectation that indirect dependency should apply to all constructions that are not readily amenable to an analysis in terms of extraction.

3.3. Acquisition Studies

Attention will now be shifted to scope marking in child language acquisition. Scope marking, as has been shown, comes in two syntactic guises. The canonical subordinated version is found in only a subset of the world’s languages. The sequential version, on the other hand, has a broader distribution.

There are several acquisition studies across a range of languages that suggest that scope marking may be universally attested in child grammar: English (Thornton, 1990; Thornton & Crain, 1994), French (Oiry, 2006; Oiry & Demirdache, 2006; Strik, 2008), Dutch (van Kampen, 1997), Basque (Gutierrez, 2004), Spanish (Gutierrez, 2006), and Japanese (Wakabayashi & Okawara, 2003), for example. Thornton (1990) and Thornton and Crain (1994) show that one-third of English speaking children in an elicited production task had acquired the scope marking construction before extraction:


Scope Marking at the Syntax-Semantics Interface

In fact, Gutierrez (2005) shows that Basque-Spanish children acquiring English as L3 produce scope marking even though their L1 and L2 lack the relevant structure. Wakabayashi and Okawara (2003) present similar results for Japanese students learning English as L2. It is worth noting that these studies use subordinated scope marking in adult grammar as their reference point. The results are less surprising if we consider sequential scope marking in adult grammar.

In this connection, another result from Thornton and Crain’s study proves instructive. According to them, English-speaking children initially manifest, along with the scope marking construction in (30), the wh copy construction in (31):


Scope Marking at the Syntax-Semantics Interface

They note that with the acquisition of extraction, the wh copy construction is lost while scope marking is retained. That the copy construction should be replaced by extraction is to be expected in a language where the adult grammar does not have the copy construction. That the scope marking construction should be retained is also to be expected because it is manifested in adult English grammar in the form of sequential scope marking.

This question remains open at this time: Is it possible for languages not to have sequential scope marking structures? After all, it requires nothing beyond intonation to club together two questions of the appropriate sort. As such, it stands a good chance of being a linguistic universal. If scope marking is universally available, we would expect scope marking to be retained in the process of language development but some of the studies mentioned seem to indicate that this is not the case. But if scope marking is lost in the process of acquisition because the adult grammar of the language lacks subordinated and sequential scope marking, what principles of grammar would be responsible for blocking the latter? Scope marking, we see, continues to pose interesting questions.

Further Reading

Crain, S. (2012). The emergence of meaning. Cambridge, U.K.: Cambridge University Press.Find this resource:

Karttunen, L. (1977). Syntax and semantics of questions. Linguistics and Philosophy, 1, 3–44.Find this resource:

Lee, I.-J., & Lee, D. (2012). On the acquisition of wh-scope marking in Korean-English interlanguage. Language and Linguistics, 57, 159–195. [Korean journal]Find this resource:

Lutz, U., Müller, G., & von Stechow, A. (Eds.). (2000). Wh-scope marking. Amsterdam: John Benjamins.Find this resource:

Martohardjono, G. (1993). Wh-movement in the acquisition of a second language: A cross linguistics study of 3 languages with and without overt movement (Doctoral dissertation). Cornell University, Ithaca, NY.Find this resource:

Thornton, R. (1995). Referentiality and Wh-movement in child English: Juvenile D-linkuency. Language Acquisition, 4, 139–175.Find this resource:

Umeda, M. (2007). Wh-scope marking in English-Japanese interlanguage. In A. Belikova, L. Meroni, & M. Umeda (Eds.), Proceedings of the 2nd Conference on Generative Approaches to Language Acquisition North America (GALANA 2006) (pp. 437–447). Somerville, MA: Cascadilla Proceedings Project.Find this resource:


Arregi, K. (2003). Clausal pied-piping. Natural Language Semantics, 11(2), 115–143.Find this resource:

Bayer, J. (1990). Directionality of government and logical form: A study of focussing particles and wh-scope (Habilitation thesis). University of Konstanz.Find this resource:

Beck, S. (1996). Quantified structures as barriers for LF movement. Natural Language Semantics, 4, 1–56.Find this resource:

Beck, S., & Berman, S. (2000). Wh-scope marking: Direct vs. indirect dependency. In U. Lutz, G. Müller, & A. von Stechow (Eds.), Wh-scope marking (pp. 17–44). Amsterdam: John Benjamins.Find this resource:

Bruening, B. (2004). Two types of wh-scope marking in Passamaquoddy. Natural Language & Linguistic Theory, 22, 229–305.Find this resource:

Bruening, B. (2006). Differences between the wh-scope-marking and wh-copy constructions in Passamaquoddy. Linguistic Inquiry, 37(1), 25–49.Find this resource:

Cole, P., & Hermon, G. (2000). Partial wh-movement: Evidence from Malay. In U. Lutz, G. Müller, & A. von Stechow (Eds.), Wh-scope marking (pp. 101–144). Amsterdam: John Benjamins.Find this resource:

Comorovski, I. (1996). Interrogative phrases and the syntax-semantics interface. Studies in Linguistics and Philosophy 59. Dordrecht, The Netherlands: Kluwer.Find this resource:

Dasgupta, P. (1980). Questions and relative and complement clauses in a Bangla grammar (PhD diss.). New York University.Find this resource:

Davison, A. (1984). Syntactic constraint on wh-in-situ: Wh-questions in Hindi-Urdu. Paper presented at LSA.Find this resource:

Dayal, V. (1990). See Srivastav.Find this resource:

Dayal, V. (1991). See Srivastav.Find this resource:

Dayal, V. (1994). Scope marking as indirect wh dependency. Natural Language Semantics, 2, 137–170.Find this resource:

Dayal, V. (1996). Locality in wh quantification: Question and relative clauses. Dordrecht, The Netherlands: Kluwer.Find this resource:

Dayal, V. (2000). Scope marking: Cross-linguistic variation in indirect dependency. In U. Lutz, G. Müller, & A. von Stechow (Eds.), Wh-scope marking (pp. 157–193). Amsterdam: John Benjamins.Find this resource:

Dayal, V. (2013). The syntax of scope and quantification. In M. den Dikken (Ed.), The Cambridge handbook of generative syntax (pp. 827–859). Cambridge, U.K.: Cambridge University Press.Find this resource:

Dayal, V. (2016). Questions. Oxford: Oxford University Press.Find this resource:

Dayal, V. (in press). Multiple Wh-Questions. In M. Everaert & Henk van Riemsdijk (Eds.), The companion to syntax (2d ed.). Oxford: Wiley. doi:10.1002/9781118396957.wbsyncom095Find this resource:

Dayal, V. (2017). Does Hindi-Urdu have feature driven wh movement to Spec vP? Linguistic Inquiry, 48(1), 159–172.Find this resource:

Fanselow, G., & Mahajan, A. (2000). Towards a minimalist theory of wh-expletives, wh copying, and successive cyclicity. In In U. Lutz, G. Müller, & A. von Stechow (Eds.), Wh-scope marking (pp. 195–230). Amsterdam: John Benjamins.Find this resource:

Felser, C. (2001). Wh-expletives and secondary predication: German partial wh-movement reconsidered. Journal of Germanic Linguistics, 13(1), 5–38.Find this resource:

Gurtu, M. (1985). Anaphoric relations in Hindi and English (PhD diss.). Central Institute of English and Foreign Languages, Hyderabad, India.Find this resource:

Gutierrez, M. J. (2004). The acquisition of L1 Basque long-distance wh-questions: A longitudinal investigation. Paper presented at BIDE04, Bilboa.Find this resource:

Gutierrez, M. J. (2005). The acquisition of English LD wh-questions by Basque/Spanish bilingual subjects in a school context (PhD diss.). University of the Basque Country.Find this resource:

Gutierrez, M. J. (2006). Acquiring long distance questions in L1 Spanish: A longitudinal investigation. In V. Torrens & L. Escobar (Eds.), The acquisition of syntax in romance languages (pp. 251–287). Amsterdam: John Benjamins.Find this resource:

Haddican, B., Holmberg, A., Tanaka, H., & Tsoulas, G. (2014). Interrogative slifting in English. Lingua, 138, 86–106.Find this resource:

Herburger, E. (1994). A semantic difference between full and partial wh-movement in German. Paper presented at the LSA Annual Meeting, Boston.Find this resource:

Hiemstra, I. (1986). Some aspects of Wh-questions in Frisian. Nowele, 8, 97–110.Find this resource:

Höhle, T. (2000). The w-…w-construction: Appositive or scope indicating? In U. Lutz, G. Müller, & A. von Stechow (Eds.), Wh-scope marking (pp. 249–270). Amsterdam: John Benjamins.Find this resource:

Horvath, J. (1997). The status of ‘wh-expletives’ and the partial wh-movement construction of Hungarian. Natural Language and Linguistic Theory, 15, 509–572.Find this resource:

Kiss, K. É. (1991). An argument for movement. In H. Haider & K. Netter (Eds), Representation and derivation in the theory of grammar (pp. 199–215). Dordrecht, The Netherlands: Kluwer.Find this resource:

Klepp, M. (2001). Partial wh-movement in German (Doctoral thesis). UCD.Find this resource:

Klepp, M. (2002). Much ado about was: Why German directly depends on indirect dependency. In. M. van Koppen, J. Sio, & M. de Vos (Eds.), Proceedings of ConSole X (pp. 111–125). Leiden, The Netherlands: Sole.Find this resource:

Kroch, A. (1998). Amount quantification, referentiality and long wh-movement. University of Pennsylvania Working Papers in Linguistics, 5, 21–36.Find this resource:

Lahiri, U. (2002). On the proper treatment of expletive wh in Hindi. Lingua, 112, 501–540.Find this resource:

Legate, J. A. (2011). Warlpiri wh-scope marking. Syntax, 14(2), 97–121.Find this resource:

Lipták, A., & Zimmermann, M. (2007). Indirect scope marking again: A case for generalized question formation. Natural Language & Linguistic Theory, 25(1), 103–155.Find this resource:

Mahajan, A. (1990). The A/A-bar distinction and movement theory (PhD diss.). MIT.Find this resource:

Mahajan, A. (2000). Toward a unified treatment of wh-expletives in Hindi and German. In U. Lutz, G. Müller, & A. von Stechow (Eds.), Wh-scope marking (pp. 317–323). Amsterdam: John Benjamins.Find this resource:

Manetta, E. (2010). Wh-expletives in Hindi-Urdu: The vP phase. Linguistic Inquiry, 41, 1–34.Find this resource:

McDaniel, D. (1989). Partial and multiple wh-movement. Natural Language and Linguistic Theory, 7, 565–604.Find this resource:

Müller, G., & Sternefeld, W. (1996). A-bar chain formation and economy of derivation. Linguistic Inquiry, 27, 480–511.Find this resource:

Oiry, M. (2006). Direct versus indirect wh-scope marking strategies in French child grammar. University of Massachusetts Occasional Papers 34. Amherst, MA: University of Massachusetts.Find this resource:

Oiry, M., & Demirdache, H. (2006). Evidence from L1 acquisition for the syntax of wh-scope marking in French. In V. Torrens & L. Escobar (Eds.), The acquisition of syntax in Romance languages (pp. 289–318). Language Acquisition and Language Disorders 41. Amsterdam: John Benjamins.Find this resource:

Ortiz de Urbina, J. (1989). Parameters in the grammar of Basque. Dordrecht, The Netherlands: Foris.Find this resource:

Ortiz de Urbina, J. (1993). Feature percolation and clausal pied-piping. In J. I. Hualde & J. Ortiz de Urbina (Eds.), Generative studies in Basque linguistics (pp. 189–219). Amsterdam: John Benjamins.Find this resource:

Pafel, J. (2000). Absolute and relative: On scope in German wh-sentences, W-.W construction included. In U. Lutz, G. Müller, & A. von Stechow (Eds.), Wh-scope marking (pp. 333–358). Amsterdam: John Benjamins.Find this resource:

Reinhart, T. (1998). Wh-in-situ in the framework of the Minimalist Program. Natural Language Semantics, 6, 29–56.Find this resource:

Reis, M. (2000). On the parenthetical features of German was …w constructions and how to account for them. In U. Lutz, G. Müller, & A. von Stechow (Eds.), Wh-scope marking (pp. 359–407). Amsterdam: John Benjamins.Find this resource:

Rizzi, L. (1992). Argument/adjunc atsymmetries. Proceedings of the North East Linguistic Society (NELS), 22, 365–381.Find this resource:

Ross, J. R. (1967). Constraints on variables in syntax (PhD diss.). MIT, Cambridge, MA (Published as Infinite syntax! Norwood, NJ: Ablex, 1986).Find this resource:

Sabel, J. (2000). Partial wh-movement and the typology of wh-questions. In U. Lutz, G. Müller, & A. von Stechow (Eds.), Wh-scope marking (pp. 409–446). Amsterdam: John Benjamins.Find this resource:

Srivastav, V. (1990). Hindi wh and pleonastic operators. Proceedings of the North East Linguistic Society (NELS), 20, 443–457.Find this resource:

Srivastav, V. (1991). Wh dependency in Hindi and the theory of grammar (PhD diss.). Cornell University.Find this resource:

Stepanov, A. V. (2001). Cyclic domains in syntactic theory (PhD diss.). University of Connecticut.Find this resource:

Strik, N. (2008). Syntaxe et acquisition des phrases interrogatives en francais et en néerlandais: une etude contrastive (PhD diss.). Université Paris 8.Find this resource:

Sulaiman, M. (2014). Wh-scope marking in Syrian Arabic: An indirect dependency. Paper presented at 2nd Asian and European Linguistics Conference, Newcastle University.Find this resource:

Thornton, R. (1990). Adventures in long-distance moving: The acquisition of complex wh-questions (PhD diss.). University of Connecticut.Find this resource:

Thornton, R., & Crain, S. (1994). Successful cyclic movement. In T. Hoekstra & B. Schwartz (Eds.), Language acquisition studies in generative grammar (pp. 215–253). Amsterdam: John Benjamins.Find this resource:

van Kampen, J. (1997). First steps in wh movement (PhD diss.). Rijksuniversiteit te Utrecht.Find this resource:

van Riemsdijk, H. (1983). Correspondence effects and the empty category principle. In Y. Otsu et al. (Eds.), Studies in generative grammar and language acquisition (pp. 5–16). Tokyo: International Christian University.Find this resource:

von Stechow, A. (2000). Partial wh-movement, scope marking and transparent logical form. In U. Lutz, G. Müller, & A. von Stechow (Eds.), Wh-scope marking (pp. 447–478). Amsterdam: John Benjamins.Find this resource:

von Stechow, A., & Sternefeld, W. (1988). Bausteine Syntaktischen Wissens. Opladen, Germany: Westdeutscher Verlag.Find this resource:

Wahba, W. A.-F. B. (1991). LF movement in Iraqi Arabic. In C.-T. J. Huang & R. May (Eds.), Logical structure and linguistic structure (pp. 253–276). Dordrecht, The Netherlands: Kluwer.Find this resource:

Wakabayashi, S., & Okawara, I. (2003). Japanese learners’ errors on long-distance wh-questions. In S. Wakabayashi (Ed.), Generative approaches to the acquisition of English by native speakers of Japanese (pp. 215–245). Berlin: Mouton de Gruyter.Find this resource: