<?xml version="1.0" standalone="yes"?>
<Paper uid="E95-1045">
  <Title>Towards an Account of Extraposition in HPSG*</Title>
  <Section position="1" start_page="0" end_page="0" type="metho">
    <SectionTitle>
2 Buccleuch Place, Edinburgh EH8 9LW, UK
</SectionTitle>
    <Paragraph position="0"> keller@cogsci, ed. ac. uk</Paragraph>
  </Section>
  <Section position="2" start_page="0" end_page="0" type="metho">
    <SectionTitle>
Abstract
</SectionTitle>
    <Paragraph position="0"> This paper investigates the syntax of extraposition in the HPSG framework.</Paragraph>
    <Paragraph position="1"> We present English and German data (partly taken from corpora), and provide an analysis using a nonlocal dependency and lexical rules. The condition for binding the dependency is formulated relative to the antecedent of the extraposed phrase, which entails that no fixed site for extraposition exists. Our account allows to explains the interaction of extraposition with fronting and coordination, and predicts constraints on multiple extraposition. null</Paragraph>
  </Section>
  <Section position="3" start_page="0" end_page="302" type="metho">
    <SectionTitle>
1 The Data
</SectionTitle>
    <Paragraph position="0"/>
    <Section position="1" start_page="0" end_page="301" type="sub_section">
      <SectionTitle>
1.1 Extraposition of S and PP
</SectionTitle>
      <Paragraph position="0"> In English, phrases can be extraposed, i.e., dislocated to the right boundary of a sentence. This phenomenon can be observed with adjuncts, such as relative clauses or PPs in (1)-(3), as well as with sentential and prepositional complements as in (4)-(6): 1 (1) An entirely new band rings today at Great Torrington, \[several of whom are members of the congregation\]. (UP) *Thanks go to Anette Frank, Tibor Kiss, Jonas Kuhn, Kai Lebeth, and Stefan Miiller for comments and suggestions in connection with the research reported here. Part of the work was carried out as part of the Verbmobil Project while the author stayed at the Institute for Logic and Linguistics, IBM Germany, Heidelberg.</Paragraph>
      <Paragraph position="1"> 1Extraposition data was acquired from the following corpora: UPenn Treebank (UP), London-Lund Corpus (LL), Stuttgart Newspaper Corpus (STZ). Other examples were taken from Culicover/Rochemont 1990 (CR), Gu~ron 1980 (CR), Haider 1994 (Hal), Nerbonne 1994 (Net), and Wiltschko 1994 (Wil).</Paragraph>
      <Paragraph position="2">  (2) Nobody must live here \[who is earning more than twenty pounds a week\]. (LL) (3) A man came into the room \[with blond hair\]. (CR) (4) There is very great public concern in Great  Britain today (...) \[whether the punishments which the courts are empowered to impose are adequate\]. (LL) (5) Extensive and intensive enquiries have been made \[into whether this fear of this penalty in fact deters people from murdering\]. (LL) (6) I don't see much argument myself any longer \[against differential rents\]. (LL) The antecedent (the category from which the dislocated element is extraposed) is a noun in these cases. Languages in which the right VP boundary is clearly marked (as e.g. by the non-finite verb in verb-second languages) can provide evidence for extraposition with verbal antecedents. Cf. the following German data, which include the extraposition of adjuncts in (7) and (8), and that of complements in (9) and (I0).</Paragraph>
      <Paragraph position="3"> (7) In der Nacht hatte es Tote gegeben \[in in the night had there victims been in Moskau (...)\]. (STZ) Moscow (8) Er hat den Nerv deutscher Nachkriegshe has the nerve of-German post-war geschichte getroffen \[mit seiner Romanhistory hit with his novel Triologie (...)\]. (STZ)  trilogy (9) Abet es wurde 5ffentlich aufmerksam but it was publicly attention gemacht \[auf eine prek~ire Situation\]. (STZ) called to a delicate situation (10) Er habe Schipke gesagt, \[dai\] man nicht mit  he have Schipke said that one not with Eiern werfen diirfe, schon gar nicht auf eggs throw be-allowed, PART PARTnot at den Bundeskanzler\]. (STZ) the chancellor  But also in English, we find evidence for extraposition from VP if we assume that adjuncts adjoin to the VP, and hence by default have to follow VP complements: (11) Florida National said yesterday \[that it remains committed to the merger\]. (UP) (12) Were hearing a lot these days \[about selling abroad, about the importance of Britain exporting abroad\]. (LL)</Paragraph>
    </Section>
    <Section position="2" start_page="301" end_page="301" type="sub_section">
      <SectionTitle>
1.2 Multiple Extraposition
</SectionTitle>
      <Paragraph position="0"> It is possible to have more than one extraposed phrase, as shown in (13) and (14): 2  (13) A man -i -j came in \[with blond halr\]i \[who was smiling\]j.</Paragraph>
      <Paragraph position="1"> (14) A paper _i _j just came out \[which talks about extraposition\]i \[which you might be  interested in\]j.</Paragraph>
      <Paragraph position="2"> In these examples, both extraposed elements are associated with the same antecedent. We observe that the serialization for multiple extraposed elements matters for PPs, but not for relative clauses:  (15)*A man _i _j came in \[who was smiling\]j \[with blond hair\]i.</Paragraph>
      <Paragraph position="3"> (16) A paper -i -j just came out \[which you  might be interested in\]/ \[which talks about extraposition\]i.</Paragraph>
      <Paragraph position="4"> We find a different pattern for multiple extraposition involving distinct antecedents:  (17) Its struck a grammarian _j last month * \[who analyzed it\]j \[that this clause is grammatical\]i. (Hal) (18)*Iti struck a grammarian _j last month \[that this clause is grammatical\]~ \[who analyzed it\]j. (Hal) (19) No one _i puts things _j in the sink \[that  would block it\]j \[who wants to go on being a friend of mine\]/. (Gue) (20)*No one _i puts things _j in the sink \[who wants to go on being a friend of mine\]i \[that would block it\]j. (Gue) It is plausible to assume that multiple extraposition with distinct antecedents is subject to a nesting requirement: The first extraposed phrase has to be associated with the last antecedent, the second one to the next-to-last antecedent, etc. Both types of constraints also apply for German, cf. Wiltschko (1994), who provides extensive evidence for the nesting requirement, including the :We use a trace-like notation to indicate the dependencies with extraposition and fronting phenomena. However, our account of extraposition involves no traces, cf. below.</Paragraph>
      <Paragraph position="5"> following data: (21) weil das Argument -i einen Mann _j because the argument a man anfgeregt hat, \[der das Fest besuchte\]j \[daft upset has who the party visited that Ranchen ungesund ist\]i. (Wil) smoking unhealthy is (22)*well das Argument _i einen Mann _j aufgeregt hat, \[daft Rauchen ungesund ist\]i, \[der das Fest besuchte\]j (Wil)</Paragraph>
    </Section>
    <Section position="3" start_page="301" end_page="302" type="sub_section">
      <SectionTitle>
1.3 Extraposition and Fronting
</SectionTitle>
      <Paragraph position="0"> The constraint of frozenness to further extraction, which states that no dislocation is possible out of an extraposed phrase, is widely accepted in the literature. The contrast between (23) and (24) illustrates this restriction: (23) Whoi did you see a picture of _i in the newspaper? (24)*Whoi did you see a picture in the newspaper of _~? Although this constraint seems to be valid for English, it is possible in German to have fronting of material from an extraposed phrase: 3 (25) Weni hast du geglaubt, daft Maria _~ who have you believed that Maria gekfiftt hat? (Wil) kissed has (26) \[Die Maria\]i hat Peter einem Mann _j the Maria has Peter to-a man gesagt, \[den er kannte\]j \[daft er -i gekiiftt said who he knew that he kissed hat\]. (Wil) has On the other hand, we can also observe extraposition from fronted phrases, as (27) and (28) show for fronted subjects and objects, respectively. (27) \[Ein Buch -j\]i hat er -i geschrieben \[das a book has he written which ihn weltberiihmt gemacht hat\]j.</Paragraph>
      <Paragraph position="1"> him world-famous made has.</Paragraph>
      <Paragraph position="2"> (28) \[Ein Buch _j\]i wa~.-i erschienen, \[das ihn a book had appeared which him weltberiihmt gemacht hat\]j.</Paragraph>
      <Paragraph position="3"> world-famous made has.</Paragraph>
      <Paragraph position="4"> We find similar data with extraposition from fronted objects in English: (29) \[Which book -j\]i did she write -i last year \[that takes only two hours to read\]j? (30) \[Which woman -j\]i did he meet -i yesterday \[from the south of France\]j? Therefore, we conclude that the phrase structure for extraposition cannot involve a hierarchiaThese examples are less acceptable to speakers of northern variants of German.</Paragraph>
      <Paragraph position="5">  cal constraint which states that extraposed elements are generally higher than fronted ones or vice versa. This is confirmed by the observation that fronted elements can be involved in multiple extraposition as in (26). Our analysis reflects this by avoiding the stipulation of a fixed location for extraposition.</Paragraph>
    </Section>
    <Section position="4" start_page="302" end_page="302" type="sub_section">
      <SectionTitle>
1.4 Islands and Boundedness
</SectionTitle>
      <Paragraph position="0"> Another common assumption is that extraposition is not subject to the islands constraints that hold for extraction to the left. The contrast between (3) and (31) makes clear that subjects are boundaries for fronting, but not for extraposition: (31)* \[With what color hair\]i did a man -i come into the room? (CR) Further, the restriction of upward boundeddeg hess applies to extraposition, i.e., in contrast to fronting, extraposition may not cross the sentence boundary: (32) Whoi did Mary say \[s that John saw a picture of_i in the newspaper\]? (CR) (33)*It was believed \[sthat John saw a picture -i in the newspaper by everyone\] \[of his brother\]~. (CR) We take both constraints as evidence that extraposition is different from fronting and should be handled using a separate nonlocal feature.</Paragraph>
    </Section>
  </Section>
  <Section position="4" start_page="302" end_page="304" type="metho">
    <SectionTitle>
2 An HPSG Account
</SectionTitle>
    <Paragraph position="0"/>
    <Section position="1" start_page="302" end_page="302" type="sub_section">
      <SectionTitle>
2.1 Nonlocal Dependencies
</SectionTitle>
      <Paragraph position="0"> We treat extraposition as a nonlocal dependency and introduce a new nonlocal feature EXTRA to establish the connection between an extraposed element and its antecedent. 4 A lexical rule is employed which removes prepositional or verbal complements from the SUBCAT list and introduces them into the EXTRA set: Complement Extraposition Lexical Rule (CELR) \[SUBCAT \[!\] ~ \[\] \] LNONLOCI,N~RIEXTRA \[\] U(\[\] }\] A similar rule is used to introduce adjuncts into EXTRA: 5 4We have to point out that the use of a nonlocal feature is not crucial to our analysis (as extraposition cannot cross the sentence boundary), but was chosen for technical convenience. Defining EXTRA in this way, we can rely on the Nonlocal Feature Principle for percolation; no additional mechanism is required.</Paragraph>
      <Paragraph position="1">  Note that the semantic contribution of the adjunct (standardly dealt with by the Semantics Principle) is incorporated into this lexical rule. The sharing \[\] states that the CONT-value of the output is identical with the CONT of the extraposed element, which in turn incorporates the semantics of the input via the sharing \[\].</Paragraph>
    </Section>
    <Section position="2" start_page="302" end_page="303" type="sub_section">
      <SectionTitle>
2.2 Periphery Marking
</SectionTitle>
      <Paragraph position="0"> Intuitively, our approach to the phrase structure of extraposition can be formulated as follows: An extraposed constituent has to be bound on top of a phrase that introduces intervening material between the extraposed constituent and its antecedent. 6 Since this constraint on the binding of an extraposed element is relative to its antecedent, we have no fixed site for extraposition, which explains the observed interaction between extraposition and fronting. It also entails a nesting requirement for multiple extraposition, as it triggers distinct binding sites for extraposition from distinct antecedents: The binding site reflects the relative position of the antecedent, b-hrthermore, we avoid spurious ambiguities which have been problematic for previous accounts/ Our requirement for EXTRA binding can be formulated in HPSG using the notion of periphery, which is defined for phrases containing an EXTRA element: A phrase has a left periphery iff it contains an EXTRA element which is inherited from (a) its phrasal rightmost daughter or (b) from its lexical head. Otherwise, the phrase has a right periphery, and EXTRA elements can be bound on is rather unusual in standard HPSG. But cf. van Noord/Bouma (1994) who show some other cases where recursive lexical rules are useful and deal with processing issues as well.</Paragraph>
      <Paragraph position="1"> 6Our analysis is inspired by the Locality Constraint for Identification (LCI) which Wiltschko (1994) proposes to account for extraposition in a GB framework. The LCI requires that an extraposed element is adjoined at the first maximal projection which dominates its antecedent.</Paragraph>
      <Paragraph position="2"> ZCf. Keller 1994, where we posited the S node as a fixed site for the binding of extraposed elements.</Paragraph>
      <Paragraph position="3"> Apart from leading to spurious ambiguities, this assumption is incompatible with the coordination data given in sec. 3.1.</Paragraph>
      <Paragraph position="4">  top of it.</Paragraph>
      <Paragraph position="5"> In case (a), no material exists to the right of the extraposed element which could intervene between it and an antecedent. In case (b), the EXTRA element originates directly from a lexical head and would be indistinguishable from a non-extraposed complement or adjunct if bound immediately. Intuitively, in both cases, the EXTRA element has to be bound further up the tree after it has found intervening material which identifies it as extraposed. null Our periphery definition entails that in a sentence which contain more than one projection with a right periphery, multiple locations for extraposition exist correspondingly. If a sentence contains no projection with a right periphery, no extraposition is possible.</Paragraph>
      <Paragraph position="6"> To formalize the notion of periphery, we introduce a new feature PERIPHERY (PER), which is located under LOCAL. Its value is of type periphery, defined as follows:  A headed phrase is marked \[PER left\] if it has a daughter D with a non-empty INHERIEXTRA set, and D is a. the rightmost daughter and phrasal; or b. the head daughter and lexical and marked \[PER left\].</Paragraph>
      <Paragraph position="7"> Note that (35b) allows for periphery marking to be specified lexically. We will return to this in sec. 2.6, where we formulate a parochial restriction for German. For English, however, we assume that all lexical entries are marked \[PER left\].</Paragraph>
    </Section>
    <Section position="3" start_page="303" end_page="303" type="sub_section">
      <SectionTitle>
2.3 Phrase Structure
</SectionTitle>
      <Paragraph position="0"> To implement the binding of extraposed elements, we introduce an additional immediate dominance schema, which draws on a new subtype of headstruc called head-extra-struc bearing the feature EXTRA-DTRS (taking a list of sign). As the binding of extraposed elements is only possible at the right periphery of a phrase, the head-extra schema specifies its head daughter as \[PEa right\] and marks its mother node as \[PEa extra\] (the latter is needed for the treatment of adjuncts, el. see. 2.5): s SHere loc(x) denotes a function which takes as x a list of sign and returns a set of loc containing the boc values of the elements of x.</Paragraph>
      <Paragraph position="2"> Note that the specification \[INHERIEXTRA { }\] requires all members of EXTRA to be bound at the same level. This ensures that extraposed elements originating from the same phrase axe sisters, and hence can be ordered by LPCs. We use LPCs to account for multiple extraposition from the same antecedent (cf. the data in (13)-(16)): (36) a. H &lt; E b. E \[HEAD prep\] &lt; E \[HEAD verb V rel\] The constraint in (36a) orders the EXTRA-DTRS (E) after the HEAD-DTR (H). With regard to the EXTRA-DTRS, PPs have to precede sentences or relative clauses, as stated in (36b).</Paragraph>
    </Section>
    <Section position="4" start_page="303" end_page="304" type="sub_section">
      <SectionTitle>
2.4 Examples
</SectionTitle>
      <Paragraph position="0"> The (simplified) tree structures for (6) and (3) are given in (37) and (38):</Paragraph>
    </Section>
    <Section position="5" start_page="304" end_page="304" type="sub_section">
      <SectionTitle>
2.5 Adjuncts
</SectionTitle>
      <Paragraph position="0"> The phrase structure for extraposition outlined so far has to be constrained further, since it allows adjuncts to adjoin higher than extraposed elements, which is clearly wrong. Cf. the following examples with extraposition from NP: (39)* An entirely new band rings today, \[several of whom are members of the congregation\] at Great Torrington.</Paragraph>
      <Paragraph position="1"> We conclude that the application of the head-adjunct schema has to be disallowed on top of a head-extra structure. This can be achieved straightforwardly by specifying adjuncts as \[MODILOClPER non-extra\].</Paragraph>
    </Section>
    <Section position="6" start_page="304" end_page="304" type="sub_section">
      <SectionTitle>
2.6 Extraposition from VP
</SectionTitle>
      <Paragraph position="0"> The AELR has to be restricted languagespecifically to account corectly for extraposition from VP: English has a head-initial VP, therefore the right periphery of the VP cannot be formed by the verb, but is provided by VP adjuncts (adverbs and PPs). As a consequence, extraposed VP adjuncts cannot be distinguished from VP adjuncts in base position, which is clearly undesirable. Therefore, we restrict the AELR to nouns on the input side, which disallows adjunct extraposition from VP and hence avoids spurious ambiguities.</Paragraph>
      <Paragraph position="1"> In German, in contrast, the AELR can apply in full generality. German has a head-final VP, which entails that a verb in final position can form the right periphery of a phrase, making extraposition of VP adjuncts and complements possible. We exploit of the lexical constraint in the PMC in (35b) to allow the binding of extraposed elements on top of verbs in final position, which we assume with Pollard (1990) to be marked \[INV --\]. We can therefore formulate the following lexical requirement:9 (40) \[INV --\] ~ \[PER right\] All other lexical entries are marked \[PER left\], and hence cannot introduce a right periphery.</Paragraph>
    </Section>
    <Section position="7" start_page="304" end_page="304" type="sub_section">
      <SectionTitle>
2.7 Fronting
</SectionTitle>
      <Paragraph position="0"> To account for the differences between English and German concerning the fronting from extraposed elements (cf. (24) vs. (25)) we restrict the head-extra schema as follows: For English we assume that both INHERISLASH and INHER\[EXTRA have to be empty for all elements of EXTRA-DTRS. This guarantees that nei9A similar rule has to be formulated for verbs with separable prefixes, where the prefix marks the right periphery.</Paragraph>
      <Paragraph position="1"> ther fronting nor further extraposition is possible from extraposed phrases.</Paragraph>
      <Paragraph position="2"> For German we assume that only INHERIEXTRA has to be empty for all elements of EXTRA-DTRS. Therefore, fronting but not extraposition is allowed from extraposed phrases.</Paragraph>
    </Section>
  </Section>
class="xml-element"></Paper>