| ISSN: 1139-8756 |
INTERPRETABLE FEATURES OF THE OBJECT POSITION: OPTIONS FOR PARAMETERS
Montserrat Sanz
Kobe City University of Foreign Studies (Japan)
The present article is a summary of several ideas introduced in Sanz (2000), Events and Predication. A New Approach to Syntactic Processing in English and Spanish. Current Issues in Linguistic Theory 207. John Benjamins. The book contains further details of the analysis, as well as a chapter on the application of the theory to phenomena concerning processing issues of traces and garden path sentences. Constructions beyond those analyzed in this article are presented in the book. The book also includes a more thorough exposition of event theory and an analysis of the functional category Event Phrase, which is omitted here.
A parameter between two languages is most notable when one of the languages allows for a construction that is completely absent in the other. Another way in which a parameter can make itself visible is by imposing different syntactic restrictions and properties to a construction that both languages share. There exist numerous phenomena related to objects that distinguish Spanish and English. These phenomena point out at the existence of a parameter concerning the object position, which manifests itself in both of the above mentioned ways. Resultative predicates of the English type, for instance, exemplify the former, since they are impossible in Spanish.
(1) The blacksmith pounded the metal flat
(2) *El herrero golpeó el metal plano
"The blacksmith pounded the metal flat"
Regarding the latter way of identifying a parameter, I will focus for the purpose of this paper on middle constructions (see Sanz (2000) for more constructions). The middle construction exists in both languages, but shows different properties in English and Spanish. Whereas it is subject to several restrictions in English, no restrictions apply to Spanish middles with regard to type of verb, but middles of any kind take the clitic se obligatorily.
(3)
a. These chickens kill easily
b. *Planets see easily/ *Planets see better at night
c. *This wall hits easily
d. *John considers a fool easily anywhere he goes
(4)
a. Los planetas *(se) ven mejor por la noche
the planets se see-PL better at night
"Planets are seen better at night"
b. Este muro *(se) golpea muy bien
this wall se hits very well
"This wall hits well", "This wall is easy to hit"
c. A Juan *(se) le considera tonto por donde va
to John se him considers a fool everywhere goes
"John is considered a fool everywhere he goes"
The first set of data (examples (1) and (2)) concerns sentences with an overt object, whereas the sentences in (3) and (4) involve a detransitive mechanism in which the internal argument must be promoted to the subject position. Thus, the root of the parameter must be sought in the properties of objects and object positions.
This article presents an analysis of these phenomena based on an extended version of the theory of the object position in Spanish by Torrego (1998) and in English by Basilico (1998). These theories about VP configuration and object properties are modified here to include a semantic feature, following the extensive research on event types of the last decades (Vendler 1967, Dowty 1979, 1991; Tenny 1987, 1992, 1994, 1995; Pustejovsky 1988, 1991; van Voorst 1988; Parsons 1990, Verkuyl 1993; Smith 1991, 1996; McClure 1994), in particular the work of Tenny and Dowty.
The approach followed in this paper assumes that semantic phenomena that affect syntax do so because there are semantic features embedded in functional projections. The object parameter in question will be attributed to the semantic feature [measure], which will be shown to belong to different projections in both languages. This differs from Tenny's (1994) account of object effects on Aktionsart in the following way: I do not attribute the measuring properties of an object to the aspectual role of that object, but to a semantic feature that the object checks in a functional projection. The main objective of the analysis presented in this paper is to understand the semantic features of object positions and the possibilities for parameterization of those features.
When the syntactic effects of an interpretable (i.e., semantic) feature diverge in two languages, two possible explanations arise. First, the feature may be associated with a 'strong' functional projection in one of the languages and with a 'weak' one in the other. A strong functional projection is taken to be one that contains uninterpretable features as well as interpretable ones. Uninterpretable features cause visible syntactic operations, and therefore the semantic feature in question may be involved in some visible syntactic process. Alternatively, the feature may be embedded in a functional projection that exists only in one of the languages but not in another (Thráinsson 1996). That is to say, both the feature composition of functional categories and the configuration may be parameterized, and the result is that an operation that is visible in a language is not visible in another.
The view taken here is that these two possibilities are a consequence of each other. Languages that do not allow for the derived object functional projection to embed both intepretable and uninterpretable features must count on an extra projection in the structure where the interpretable features of the object can be hosted. This shows in the relative restrictions that apply to the constructions that materialize in a language. The proposal is explained in detail in the following section.
Torrego (1998) analyzes direct objects in Spanish marked with the preposition a, the dative preposition. She concludes that the position where objects check Case in this language can host inherent as well as structural Case properties. This proposal is extended in Sanz (2000) and hence in the present article to include Aktionsart properties. Under the account being argued for here, the functional projection where objects check Case is enriched with a semantic feature [measure]. The inherent Case that Torrego refers to is considered to be measure.
Basilico (1998) focuses on objects in English and concludes that there are two possible positions for object generation in this language, as well as the derived position for Case checking (similar to Spanish). One of these two projections is called the Transitive Phrase. I will argue that the feature [measure] is part of that functional projection, which is internal to the VP. If a semantic feature is embedded in a projection where lexical items merge, this feature is, naturally, checked by Merge, and not by movement. In the case of English, every time an object merges in the Transitive Phrase, the feature [measure] is checked overtly. Hence, inherent Case is independent from structural Case in this language, and the derived position for objects hosts only structural Case properties.
The derived position in Spanish may or may not contain an interpretable feature ([measure]), because it is a site where both structural and inherent Case can be checked. This option (or underspecification of the position for the semantic feature) exists in Spanish because there is no alternative position where an object could measure out the event. Objects merge (generate) as sisters of verbs, an internal position where partitive Case is assigned, but where no measuring out is possible. Measuring out occurs only when the object raises outside of the VP to the specifier of v.
In English, however, the specifier of v never contains the interpretable feature [measure]. The measure inherent Case can only be checked internally to the VP, whereas the derived position is limited to assigning accusative Case. In other words: inherent and structural Cases are separated in English and this language has an additional internal position for object merging. In Spanish, on the contrary, the inherent Case (measure) and the structural Case (accusative) must be dealt with in the same functional projection.
In sum, this study is intended as a contribution to the understanding of parameters on the basis of features of functional projections. It is also a contribution to the literature on Aktionsart, since it presents a comparison between event phenomena in English and Spanish and explores the roots of the syntactic differences that cause those phenomena. Consequently, it is a study about the nature of the object position and of objects in general.
The skeptical reader might have the feeling that an analysis of this type is a fancy way of describing the phenomena rather than explaining them. Accounting for the observed meaning of the constructions by positing a semantic feature is nothing but re-stating the descriptive facts. I believe that most contemporary research in theoretical linguistics could be viewed as such a re-definition. However, the descriptive nature of our analyses does not deprive them of their value. More and more refined "re-descriptions" of the facts have given us a wealth of insight into what factors play a role in the grammars of languages. One can only hope that this is a step towards a deeper understanding of human language.
We are far from being able to claim that we have arrived at the roots of parameters. This paper is written with the modest objective of contributing some insights that might be useful in the final, but as of yet distant, complete explanation of the true nature of parameters between languages (for a recent discussion on parameters, cf. Baker 2001).
3. The objet position:Torrego (1998)'s and Basilico (1998)'s analyses
In this section I introduce the proposals that serve as the basis for the analysis that will be argued for in this paper. Both of these studies analyze the syntactic structure of the VP.
According to Chomsky (1995a, 1998a), a transitive construction contains a light verb v that may have an extra specifier for objects to raise to. It is in this specifier where objects check accusative Case. In Spanish, the specifier of v has certain peculiarities. When the object is animate, it is marked with the dative preposition a ("to").
(5)
a. Vi la mesa
saw the table
"I saw the table"b. Vi *(a) la niña
saw to the girl
"I saw the girl"
If the animate object is indefinite, the marking is obligatory with some verbs but optional with others. Verbs that are inherently telic obligatorily mark their objects with a; verbs for which the marking is optional behave as telic when the object is marked, and as atelic when it is not (Torrego 1998). Marked accusative on indefinite objects affects the telicity of the sentence, turning statives into activities and activities into accomplishments. Accomplishments and a subclass of stative verbs like odiar ("hate"), admirar ("admire"), respetar ("respect"), in other words, 'active' emotions according to Pesetsky (1995), require accusative marking. (Examples are Torrego's).
(6)
a. Detuvieron *(a) unos emigrantes
arrested to some immigrants
"They arrested some immigrants"b. Conocieron (a) unos emigrantes
met to some immigrants
"They met some immigrants"
Objects marked with a are always interpreted as specific. Thus, the object in (6)a, if marked with the dative preposition, refers to a set of specific immigrants.
Another property these objects display is that they occur with verbs that take an agent or a cause as their subject. The object in (7)b is ungrammatical with the preposition, even if animate, because the subject is not agentive.
(7)
a. Esa familia esconde (a) prisioneros de guerra
that family hides to prisoners of war
"That family hides prisoners of war"
b. Esa montaña esconde (*a) prisioneros de guerra
that mountain hides to prisoners of war
"That mountain hides prisoners of war"
Torrego interprets the difference between the marked and the unmarked objects in relation to the type of Case they check. She establishes a distinction between those objects that are marked and affected by the action, and those that are marked but not affected. She assumes that the former bear inherent Case. That is to say, the Case of 'non-affected' marked objects is structural, but the Case of 'affected' marked objects is inherent.
The Case of morphologically marked objects is inherent or structural depending on the verbs. With accusative verbs that mark their objects, the object takes both kinds of Case. Torrego assumes that dative a-marking occurs in the specifier of v, in Vmax but outside the nuclear scope of the V.
Note that dative marking of objects only affects animates, and within the class of animates, indefinites. Furthermore, contrary to what most traditional grammars of Spanish state, not all animates must be marked with the dative preposition. The marking depends on the agentivity of the subject of the action (recall (7)b). For the moment, it is important to bear in mind that these properties (agency, indefiniteness) are semantic properties that relate to the type of event that a sentence expresses. We will deal with all these issues in subsequent sections.
Basilico focuses on an interesting idea about internal predication within the VP. He claims that the object in English is an inner subject of a predicate formed by the verb and its indirect complements. The idea is that, in a construction with an indirect object, for instance, the direct object acts as a subject of which the verb and the indirect object are predicated. This is exemplified in the following VP configuration for the sentence Mary sent the letter to Jill.
(8)
According to Basilico, there are two positions for objects inside the VP in English 1. In addition, there may be an external position in the inflectional domain for objects to move to (which we will take to be the specifier of v, following Chomsky 1995 and subsequent work).
Objects in English can be generated in two different positions within the VP, which correspond to different predication forms. One position is the specifier of the inner V', and the other is the specifier of a Transitive Phrase placed between the two VPs of a VP shell. The two predication forms are the same as can be found in entire sentences - thetic and categorical.
A thetic sentence is a sentence in which the existence of the whole event is asserted, rather than assuming that there is an argument of which something is said. Contrary to this, in categorical predication sentences, there is an argument that becomes singled out and about which something is predicated.
Concerning the predication structure of objects, if the object occupies the internal specifier of V' position, the VP expresses a thetic predication: that object or an event involving the object is being asserted. If the object occupies the external position, the predication is categorical: the object is singled out and a property of the object is being asserted. The configuration proposed by Basilico for the English VP is the following.
(9)
In the thetic configuration, the object is generated in internal position. In the categorical configuration, the object is generated directly in Transitive Phrase. As it can be seen in the tree, the sister of the verb is the position where indirect objects are located. Thus, the object is the 'subject' of the predicate formed by the verb and the indirect object.
Basilico bases his claims on phenomena of interpretation related to verbal diathesis alternations, such as the creation/transformation, the locative (spray/load) and the dative alternations.
(10)
a. The cook baked three dozen cookies from one pound of dough
b. The cook baked one pound of dough into three dozen cookies
(11)
a. The farmer loaded a bale of hay onto the truck
b. The farmer loaded the truck with a bale of hay
(12)
a. The instructor gave a book to the student
b. The instructor gave the student a book
The (a) cases involve either the presentation of the direct object or of an event in which the direct object is a part without singling out the direct object from the event (i.e., the predication is thetic). It is in these cases that the direct object is speculated to be generated within the inner object position. On the other hand, in the (b) sentences, the verb and its complement combine to form a property that allows for a categorical predication form, and this property is ascribed to the direct object (or indirect object, as in the instructor gave the student a book). The object of these sentences is generated in the outer object position, the Transitive Phrase.
The sentence that involves thetic predication shows ambiguity of scope among quantifiers. This ambiguity is absent in the categorical counterpart.
(13)
a.The seamstress stuffed a pillow with each bunch of feathers
b. The seamstress stuffed a bunch of feathers into every pillow
In the categorical member of the pair (example (13b)), the quantifier a has wide scope over every. The ambiguity in scope of the thetic objects stems from their being generated in the inner VP, where a trace is bound after movement.
Likewise, when a frequency adverb is inserted in the sentences, it is the thetic alternants that show the ambiguity in scope between the adverb and the object. Whereas in (14)b (categorical), there is only one bag which the robbers stuff frequently, (14)a (thetic) has two readings, one in which the robbers stuffed a different bag each time.
(14)
a. During the holdup, the robbers stuffed a bag frequently with a wad of cash
b.During the holdup, the robbers stuffed a wad of cash frequently into a bag
Finally, there-constructions serve to distinguish these two types of object predication. Whereas both alternants can appear in existential constructions, there are differences in the meaning associated with them. Thus, the thetic examples can be read as an assertion of the existence of a certain state of affairs or event. In contrast, the categorical sentences only allow a reading in which what is asserted is the existence of the postcopular noun.
(15)
a. There were statues carved out of wood (thetic)
b. There was wood carved into the statue (categorical)
In sum, taking all these interpretational pieces of evidence together, Basilico argues that the object in the two members of the pairs that he discusses must be generated in different positions within the VP. Several possibilities for object merging arise in this language: an object can be generated in the internal position (specifier of V') and then ascend to Transitive Phrase, or remain in the position it was generated in, being assigned inherent (partitive) Case. An object can be generated in the specifier of Transitive Phrase directly, and then ascend to the external position for objects outside the VP (here, specifier of v, following Chomsky 1995a). Finally, an object may be generated in the specifier of V' and then ascend to the position outside the VP. In this case, Transitive Phrase would not be projected.
3.3. What can be hypothesized from these analyses
The two proposals developed independently point at the possibility that English displays a more complex structure for the VP than Spanish, since a difference in number of potential projections is observed. Torrego does not consider the existence of a Transitive Phrase internal to the VP in Spanish. When applied to the data on resultatives and middles introduced above (among other parameters involving objects), these two accounts become useful. If we assume a difference in configuration between Spanish and English, we can start envisioning a solution to the problem of why English displays constructions like the resultative that are impossible in Spanish. On the other hand, in order to explain the properties of resultatives and the restrictions of middles in English, as well as the lack of restrictions on the verbs or objects in Spanish middles, a thorough understanding of the semantic properties of the functional projections that are posited in Basilico's and Torrego's work is necessary. This paper undertakes such a study.
Torrego's and Basilico's analyses were generated taking into account different sets of data. Whereas Torrego was interested in elucidating the facts concerning dative marking of direct objects in Spanish, Basilico focused on the predication properties of the two members of an alternation. As we saw, the dative marking properties are intimately related to semantic notions pertaining to event types. Likewise, the spray/load alternations that drive Basilico to his conclusion are examples of alternating event types, and the predication structure that he links to the different positions for objects is part of the semantics of the event.
Extensive research has been conducted in the event literature about the role of the object in causing a certain event type (Dowty 1979, 1991; Tenny 1987, 1992, 1994, 1995). Therefore, the data introduced in Section 1 to illustrate the object parameter could be used to confirm whether the two different structures proposed for the VP in English and Spanish by Basilico and Torrego respectively are on the right track. Were this to be the case, the parameter is on the way to be confirmed as a difference in the number and nature of functional projections, and this paper becomes an exploration on the way semantic features are encoded in those projections.
Torrego speaks about inherent Case. Inherent Case is semantic, interpretable Case. In Spanish, the prepositions that mark inherent Cases (like a or genitive de) do not delete after checking, because they are needed for interpretation (Torrego 1998). Under an account of Case that relies on features, the position that assigns inherent Case should contain an interpretable feature from which the projection acquires its semantic content, in the same fashion as structural Case is a feature - albeit uninterpretable - of a functional projection.
The term inherent Case, when applied to direct objects, is used in the literature as a quasi-synonym for partitive Case. However, the objects that receive inherent Case in Torrego's analysis are those that turn an activity into an accomplishment, for instance, and are considered specific. They are closer to the types of objects that Tenny considers measurers of an action and certainly they do not bear any kind of partitive meaning. The semantic Case that these objects receive must be defined through a semantic feature that reflects their semantics. I will call this feature [measure], taking the notion of measurer as defined by Tenny 1987, 1994 (see Sanz (2000) for a more extensive review of this and other Aktionsart notions). In the relevant examples of this paper, inherent Case is associated with the measuring properties of the object. In both Basilico's and Torrego's proposals, I hypothesize that inherent case is associated with the checking of the interpretable feature [measure].
In English there exist clear syntactic consequences of the fact that an object is a measurer of the event. In Spanish, in contrast, as long as the clitic se is included, sentences can be interpreted as accomplishments. Furthermore, a potentially good measuring object is not enough to turn an activity into an accomplishment in this language (see Sanz 2000 for a thorough account of accomplishment constructions in Spanish). The clitic se becomes essential in accounting for the grammaticality of many constructions that are ungrammatical in English, like some middles. The two constructions that exemplify the parameter under discussion here have the properties of accomplishments. They are analyzed in detail in the following paragraphs.
4.1. The resultative construction
A resultative is an XP that is predicated of the direct object of a verb and describes the resulting state of that object after the action has been performed on it. The following sentences are examples of resultatives. (16)a presents a transitive verb. Flat refers to the state the metal is in after the pounding. (16)b shows that an uncausative verb produces a good sentence when modified by a resultative construction. (16)c exemplifies an activity verb with an agent (i.e., an unergative), and the sentence with a resultative is ungrammatical. Examples are from Levin & Rappaport Hovav (1995).
(16)
a. The blacksmith pounded the metal flat
b. The river froze solid
c. *Dora shouted hoarse
According to Levin & Rappaport Hovav, transitive verbs of change of state like in (16)a above and uncausative sentences ((16)b) permit the presence of resultative predicates because they fulfill the syntactic requirement of containing an internal argument. Stative verbs - remain, for instance - and verbs of inherently directed motion like arrive do not enter into this construction 2.
Interestingly enough, syntactic tests of VP constituency (e.g., do so substitution) reveal that resultative phrases are VP internal (Levin & Rappaport Hovav 1995). They are selected constituents of the verb. In (17), did so would have to include the resultative phrase too.
(17)
*Bill fastened the shutters open, and Mary did so shut
Some ungrammatical resultatives are proposed in Sanz (2000). There is no apparent reason why it could not be the case that aging results in someone turning ugly, or darkening a photograph too much in the picture becoming unrecognizable. Nevertheless, it is impossible to express these concepts through the resultative construction in English.
(18)
a. *She darkened the photo unrecognizable
b. *The lady aged ugly
As Sanz (2000) points out, the notion of measure appears to be crucial in understanding these ungrammaticalities. The sentence requires a measuring argument for the resultative construction to be grammatical. (Tenny 1994). The ungrammatical examples in (18) above illustrate this. Neither of the internal arguments in those sentences is a measurer of the event. Darkening is a degree achievement verb, thus atelic. Aging as in (18)b is a process that has no endpoint. Still, if the lady is conceived of as a measure argument, the sentence with the resultative becomes grammatical.
(19)
The lady aged into an ugly monster
In this example, the lady experiences a change of state that ends with her transformation into a different entity, so to speak (she is not a lady anymore, but an ugly monster). In other words, the sentence is interpreted as an accomplishment: there is a process that leads to a final point which is different from all the previous points. When viewed in this way, the resultative construction is grammatical. The object is considered a measurer (i.e., has checked the feature [measure]).
In sum, it is the presence of a measuring object in the sentence that allows for the resultative to appear. This is equivalent to saying that the object must somehow check the interpretable feature [measure]. Furthermore, the resultative is predicated of the measuring object. That is to say, the object and its resultative form a categorical structure, a structure in which the object is singled out and something is stated about that object. Following Basilico, the object must have been generated in the Transitive Phrase. Reversing the order of the reasoning, if the object that is generated in Transitive Phrase is a measurer of the action, then we must assume that it has checked the feature [measure], which we will therefore place in the Transitive Phrase in English.
A look at the cases that Basilico argues to be categorical (and hence to have the object merge in Transitive Phrase) confirms this conclusion: those sentences are all accomplishments, that is to say, they are precisely the cases that involve a measuring object.
In contrast with English, Spanish lacks resultatives altogether.
(20)
a. El río se congeló (*sólido)
the river se froze solid
"The river froze solid"
b. Juan limpió la mesa (*impecable)
John wiped the table impeccable
"John wiped the table (impeccable)"
Let us assume that the Transitive Phrase is not part of the VP configuration in Spanish. Thus, in a transitive construction like (20)b the object merges in the sister of V position and then raises to the specifier of v. It seems like the object cannot be interpreted as a measuring argument, given that the resultative is ungrammatical. It follows that the object has not checked the feature [measure], or that this feature is not in either the projection where the object merges or in the one where it checks Case. As for detransitive constructions like (20)a, the only position for the object is the internal one, thus no position linked to Case is available.
Torrego noted that some objects can check inherent Case in the specifier of v. Those objects that are assigned inherent Case act as measurers in that they cause the sentence to be interpreted as an accomplishment. I attribute semantic Case to the presence of a semantic (i.e., interpretable) feature that I call [measure] because it is associated with accomplishments. This feature is thus encoded in the specifier of v in Spanish (external to the VP, checked by movement). As opposed to this, as we saw above, the feature must be encoded in the Transitive Phrase in English (internal to the VP).
The question is why this feature is absent from the specifier of v in sentence (20)b. Definite objects check structural Case, and thus we must assume that in this sentence the position of specifier of v holds the structural [accusative] Case feature. We know that structural and inherent Cases are not in complementary distribution in Spanish, since affected marked objects check both. The issue here is whether definite objects can only check structural but not inherent Case, or whether inanimates (which are never marked with the dative preposition in Spanish when in object position) can never be measurers.
Both of these alternatives are awkward, given that in Dowty's and Tenny's accounts of incremental themes and measurers, it is precisely inanimates, and in many cases definites, that are used as examples because they are typically the best measurers. In Sanz (2000), however, I argued that only objects of construction verbs are true measurers in Spanish (the object does not have an independent existence until the event has taken place in its entirety). All other types of verbs, even when they have a potentially good measurer, do not result in unambiguous accomplishment sentences. The distinction between a process and an accomplishment interpretation is achieved through se, not through the presence of a measurer. With construction verbs, however, the object measures out the event. This would mean in my account that the object of a construction/destruction verb, even though it may be inanimate and definite, checks the inherent Case associated with the feature [measure].
In fact, and as confirmation of the above conclusion, these verbs in Spanish accept some sort of resultative phrase, although within a limited range.
(21)
a. Construyó una casa torcida
built a house crooked
"He built a house crooked"
b. Rompió el papel en pedazos
broke the paper in pieces
"He tore the paper into pieces"
Both objects in the previous sentences check the feature [measure]. I assume that they are assigned inherent Case as in the examples by Torrego, even though one is definite and both are inanimate. Other transitive verbs in Spanish with a potentially good measurer, however, always remain syntactically atelic unless the clitic se marks the sentence as an accomplishment (Sanz 2000), or unless the object can be marked with the dative preposition as in Torrego (1998). Neither of these is the case for the object in (20)b, which accounts for its ungrammaticality. Clearly, the feature [measure] is not pervasive in transitive constructions in Spanish.
When turning our attention to detransitive constructions, we must bear in mind that what must be proven is that the object is able to check the feature [measure] without checking accusative Case. This will be shown to be the case for English, given the possibility of merging in a position that assigns inherent Case separate from structural Case. However, this will again be proven to be impossible in Spanish, where the only way to have a grammatical middle is the presence of the clitic se.
Middles are perhaps one of the most elusive topics in English syntax. Some of the restrictions are exemplified as follows (some of the examples are from the work by Hale & Keyser (1986, 1987, 1991, 1993)).
(22)
a. These chickens kill easily
b. Greek translates easily
(23)
a. This car drives easily
b. This violin plays well 3
(24)
a. *Planets see easily/ *Planets see better at night
b. *This wall hits easily
c. *John considers a fool easily anywhere he goes
d. *?Houses build fast but airports build slowly
The meaning of a middle is that every time an event of the type conveyed in the sentence is performed, the event is performed in the manner expressed by the adverb (easily, well, etc.). Adverbs like easily usually imply the presence of an agentive argument that is the causer of the action. With this adverb, the meaning of the middle is usually that it is easy to cause the event to happen.
After an extensive review of the behavior and semantics of middles, Sanz (2000) concludes that the interpretation of middles as generalizations over kinds of events originates from the sentences being interpreted as a repetition of events of the same type at different times and with different entities. (i.e., as expressing quantification over events). Given that they are constructed in an imperfective tense (the present), their generic nature can be traced back to the fact that the events themselves are telic.
But how is telicity achieved? It seems that in English there are at least two ways of forming a grammatical middle. The first one involves verbs that take measurers. With verbs expressing actions that occur in stages (translate, wash, cut), the event is interpreted as telic because it is measured out. The combination of telicity with imperfective tense leads to a reading in which there is more than one event of the same type involved. Entities like this bread in the sentence this bread cuts easily could be interpreted as classes of objects (Fagan 1992), because every event of cutting refers to a different piece of bread (i.e., the sentence would mean 'this type of bread is easy to cut'). Alternatively, the sentence can be interpreted as involving a unique piece of bread on which several events of cutting are being performed. In any case, the existence of several events of the same type is what is relevant.
The second type of middle takes verbs that denote an event whose argument is perceived as capable of bringing it about by itself, even if the action is atelic (examples (23)). I will return to these examples momentarily.
Importantly, a middle cannot be formed on the basis of punctual events or stative predicates (see examples (24)). Only eventive actions with measurers can be turned into middles in English (see also Tenny 2000).
I assume, given the discussion above, that the object of constructions based on verbs like cut or translate checks [measure]. If the feature [measure] were associated in this language with a derived position where objects check Case, it would be impossible to check it overtly in a detransitive construction like the middle. The alternative is to check [measure] by Merge rather than by moving to a position outside the VP. I claimed that the feature [measure] is embedded in a position internal to the VP where objects may start the derivation, the Transitive Phrase. The object of a middle generates (i.e., merges) in that position. Consequently, this feature is checked overtly in middle constructions before the argument moves to subject position.
Let us now turn our attention to the examples in (23), which do not contain measurers and are nevertheless grammatical middles. In these sentences, there is someone who causes the event to happen. The entities involved (this car and this violin) are seen as performing the action expressed by the verb and the external causer is interpreted as causing the event, not as performing the action itself (although technically a car or a violin cannot drive or play by themselves, they are perceived as performing the movement or producing the music). In these cases, the sentence does not express a specific instance of an activity of driving or playing. They rather refer to the fact that an external causer brings about events of driving or playing. For instance, this violin plays well refers to events of playing in which somebody causes the violin to play. The sentence can be paraphrased as "every time someone causes this violin to play, the event comes out well". It does not mean that the violin is performing the activity of playing in this particular sentence. If it were, it should be possible to say this violin is playing easily, which is acceptable only under the special figurative use of individual-level predicates like she is being silly 4.
Thus, in Sanz (2000), I argued that the fact that the verbs drive or play express atelic activities is not relevant for the grammaticality of the middle, since the sentence is not about the particular action expressed by the verbs but rather about bringing about events of that type. That is to say, the object of the construction is not the car or the violin, but the whole event of playing the violin or driving the car, which is subordinated to causation. In this sense, the sentences involving external causation can be considered on a par with other accomplishments containing a measurer.
However, it should be noted that speakers' judgments vary with the choice of adverb. Thus, this violin plays well seems more acceptable to some people than this violin plays easily. Interestingly enough, the grammar of Spanish treats these two sentences as different constructions.
(25)
a. Este violín (*se) toca bien
this violin (*se) plays well
"This violin plays well"b. Este violín se toca fácilmente
this violin se plays easily
"This violin plays easily"
Whereas (25)a is a simple atelic sentence, a stative predicate that rejects se, (25)b is a middle that requires the clitic. The only difference being the adverb, we must hypothesize that adverbs like easily imply the presence of an agent (the violin being a sort of "secondary" performer of the action), and that therefore the verb is used as a transitive. A sentence like (25)a, in contrast, is interpreted as a statement about the properties of the violin itself. In other words, whereas (25)b involves a transitive verb, the verb in (25)a is intransitive. I speculate that English speakers who accept this sentence gladly do not view it as a middle implying an agent, but merely as an intransitive construction similar to its Spanish counterpart. On the contrary, their reluctance to accept the sentence this violin plays easily might be due to the fact that the violin is not a measurer (and that in order to have a middle construction in English, one needs, after all, a transitive verb with a measurer).
Furthermore, speakers who accept this violin plays easily as a middle should note a slight difference in grammaticality between that sentence and this symphony plays easily, under a reading in which the symphony is a measurer (Beth Levin, p.c.). The former would be paraphrased as "it is easy to cause this violin to play (music)", and the latter would be interpreted as "it is easy to play this symphony", with the symphony viewed as a measure 5.
How about examples like this car drives well? Again, resorting to Spanish gives us some insight into the real syntax and semantics of this sentence. Contrary to English, the verb conducir ('drive') is exclusively transitive.
(26)
This car drives at a speed of 100 Km/hour(27)
a. *Este coche conduce a una velocidad de 100 Km/h
this car drives at a speed of 100 Km/hour
"This car drives at a speed of 100 Km/hour"b. Este coche corre a una velocidad de 100 Km/hour
this car runs at a speed of 100 Km/hour
"This car drives at a speed of 100 Km/hour"
In order to translate sentence (26) into Spanish, it is necessary to change the verb for a true intransitive one like correr ('run'). However, in English, the verb can act as an intransitive. This makes us speculate that the sentence this car drives well is nothing but an intransitive sentence and not a detransitive (middle) one. The paraphrase would be something like 'this car moves well', instead of 'this car is driven well', which would be the interpretation of a middle.
The above discussion notwithstanding, I admit that the examples still pose an unresolved puzzle for my theory. A further refinement of the hypothesis about the semantic feature [measure] will be needed in the future in order to establish the real nature of these alleged middles 6.
The se0ntences which are ungrammatical as middles in English form good middles in Spanish, as long as se is present to make the sentence telic.
(28)
a. Los planetas *(se) ven mejor por la noche
the planets se see-PL better at night
"Planets are seen better at night"b. Este muro *(se) golpea muy bien
this wall se hits very well
"This wall hits well", "This wall is easy to hit"c. A Juan *(se) le considera tonto por donde va
to John se him considers a fool everywhere goes
"John is considered a fool everywhere he goes"
These examples are perfectly grammatical. All of them include the telic clitic, which is obligatory. If I am correct in linking the semantic feature [measure] to the position where objects check Case in Spanish, middles, being detransitive structures, should not contain measurers. However, the interpretation of a middle in Spanish does not differ from that of a middle in English: a middle is a generalization over events that stems from a telic construction being expressed in an imperfective tense. In other words, it expresses an iteration of events of the same type. Where does the telicity derive from, in the Spanish cases? As I proposed in Sanz (2000), se checks the [telic] feature of Event Phrase, and thus the sentences above are interpreted as telic. The nature of the verb and the relationship between the verb and its argument is irrelevant, because in Spanish there exists an Event Phrase (external to the vP) with both interpretable and uninterpretable features that can be checked by the clitic. It is the clitic that has the capability of checking the event feature of the sentence 7. Since the events above occur in combination with an imperfective tense (the present), a middle (generic) reading results (every time an event of this type is brought about, the event is carried through in the manner expressed by the adverbial phrase).
In sum, Spanish and English differ in that checking the feature [measure] is not a requirement in Spanish for the grammaticality of a middle (or an uncausative). However, in English, if an object has no measure properties and cannot check the feature, the middle is ungrammatical. This indicates that the object of this construction is generated under Transitive Phrase, where [measure] is embedded.
5. Measure and categorical predication
In the previous section I have proposed that the feature [measure] is embedded in the Transitive Phrase in English, and therefore checked automatically by merging of an object into that position. Since this is the position where, according to Basilico, objects involved in categorical predication are generated, we should find a correlation between categorical predication and the presence of a measuring object in a sentence.
As stated above, the categorical cases in Basilico's (1998) pairs of examples are precisely those that involve measuring out of the verb by its object. In the case of middles, if my hypothesis is correct, the object should be involved in categorical predication 8.
The thetic and the categorical alternants that Basilico analyzes in his study yield differences in scope ambiguities of quantifiers. The thetic alternants allow for a scope ambiguity, whereas the categorical alternants do not. With the thetic alternants, the direct object binds a trace within the VP, since it was generated in the inner VP and as a result, both scope possibilities are possible in the thetic member of the pair.
If the analysis of English resultatives proposed here is correct, and the Transitive Phrase is endowed with the feature [measure], we should expect no scope ambiguity in case that the sentence with the resultative contains quantifiers. This prediction is borne out. Even though conceivably both scope options are possible in the following cases, there is only one in which the resultative can be said to express the state of the object after the action has taken place. This is the one in which every has scope over a (that is, each blacksmith pounded a different metal). Under an interpretation in which each blacksmith pounds the same piece of metal, the resultative cannot be said to express the state of the metal after every event of pounding. Thus, ambiguity of scope is not possible.
(29)
a. Every blacksmith pounded a metal flat
b. Every student wiped a table clean
Similarly, with middles, the facts point at the possibility that the analysis is on the right track. The second option for scope (one in which the same expert translates every book) is at best awkward.
(30)
Every book translates easily (if translated) by an expert
5.2. Ambiguity with frequency adverbs
Basilico proves that when a frequency adverb is placed after the verb in English, thetic alternants show scope ambiguity between the adverb and the object, a scope that is absent in the categorical sentences. Let us apply this test to resultatives (note that the adverb must be placed at the end of the sentence, or else we would be breaking the cohesion between the object and the resultative). Ambiguity seems possible in sentence (31), contrary to our prediction, but there seems to be a preferred interpretation in which it is the same metal that the blacksmith pounds.
(31)
The blacksmith pounded the metal flat frequently
In the case of the middle, this test clashes with the presence of a manner adverb in the construction as we have been assuming so far to be consistent with the literature. Changing the adverb easily into the adverbial in an easy way allows for the sentence to be grammatical.
(32)
These books translate frequently in an easy way
As for its interpretation, it seems to me that the sentence can only mean that it is frequently the case that these books translate easily, not that the books are translated frequently. Thus, only a possibility for interpretation arises, consistent with the analysis argued for in this paper.
The there-construction test that Basilico proposes is impossible with the constructions being analyzed here, and therefore we will not attempt to use this test as a confirmation of our analysis.
In sum, the analysis outlined in this paper predicts that the internal arguments in English middles and resultative constructions, being measurers, participate in a categorical type of predication because they are generated in Transitive Phrase. I have linked the Aktionsart behavior of these objects as measurers to the checking of a semantic feature [measure] that is embedded in that functional projection and checked by Merge. The fact that [measure] is checked by Merge in one of the potential positions for object generation accounts for the pervasive effects of measurers in this language (see Sanz 2000 for more constructions).
In contrast, the effects of the semantic feature [measure] are scarce in Spanish, although we can see some overt effects in the form of some indefinite objects that show dative marking. The Aktionsart properties and effects of these marked objects serves as the basis for an analysis in which the feature [measure] is linked to the Case checking position.
In Sanz (2000) a thorough analysis of the clitic se in transitive constructions like the following is undertaken.
(33)
Pedro se leyó un libro
Pedro se read a book
"Pedro read a book (completely)"(34)
Pedro (*se) leyó un libro durante tres horas
"Pedro read a book for three hours"(35)
a. Pedro #(se) leyó un libro en una hora
Pedro se read a book in an hour
"Pedro read a whole book in an hour"b. Pedro #(se) leyó el libro entero ayer
Pedro se read the book whole yesterday
"Pedro read the whole book yesterday"
As a conclusion, it is established that the presence of the clitic means that the sentence is unambiguously telic, and that if there are elements that point at the sentence being telic (like the adverb todo/'all' with consumption verbs, or the adverbial en una hora/'in an hour'), the clitic appears as a marker.
(36)
a. El niño ??(se) lo comió todo
the child se it ate all
"The child ate it all"b. ??(Se) comió el pastel entero
se ate the cake whole
"He ate the whole cake"c. ??(Se) comió todos los pasteles
se ate all the cakes
"He ate all the cakes"d. ?(Se) bebió una cerveza en un minuto
se drank a beer in a minute
"He drank a beer in a minute"e. ? (Se) comió el pastel en una hora
se ate the cake in an hour
"He ate a cake in an hour"
Thus, as long as the clitic is present, the telicity of the sentence is ensured. As we saw for middles, this makes the checking of [measure] unnecesary. But in other cases, as those noted by Torrego and introduced in section 3.1 above, we see that the checking of this semantic feature is performed, since the dative marking is linked to measuring. I assume that the feature, when present, is in the derived position where objects check Case. The fact that the Transitive Phrase position does not exist in the VP configuration in Spanish explains why this language lacks resultatives, restrictions on middles, the three kinds of alternations analyzed by Basilico (1998) and many other constructions like measuring of movement verbs by goal phrases, for instance. The fact that the semantic feature is embedded in an external object position explains that it is impossible to be checked in detransitives, and as such, if detransitives are to be intrepreted as accomplishments at all, they require an alternative way of signaling it. Hence the obligatoriness of the clitic se.
Given that the objects are derived and are merged initially in the sister-of-V position, any facts concerning categorical and thetic predication in Spanish must be determined with different tests than those offered by Basilico for English. This is beyond the focus of this paper and remains open for future research.
Before closing the discussion, I would like to point out that Basilico assumes that objects generated in the inner VP may move to Transitive Phrase to satisfy a strong [nominal] feature associated with Transitive Phrase. However, this would be a problem for my analysis, since by moving to Transitive Phrase the feature [measure] would also be checked and there would not be any difference between the objects originally generated there and those raised to that projection by movement. I believe that this alternative does not occur at all. The sister-of-V position hosts another type of semantic Case, i.e., partitive Case. If it were possible to have the object merge in the internal position and then ascend to Transitive Phrase, the same argument would be checking both partitive and measure semantic Cases in the same sentence, and this would result in a contradiction (measure is associated with telic events, whereas partitive is characteristic of atelic ones). Thus, I assume that if an object generates in the sister-of-V position, the Transitive Phrase is not projected. An object that initially merges in the internal position would only have the option of moving to the specifier of v in order to check structural Case, if needed.
In this paper, I have taken two constructions as examples of a parameter involving objects in English and Spanish. One of them is the resultative construction, which exists in English but not in Spanish. The other is the middle, which exists in both languages but shows different restrictions for grammaticality. In particular, in Spanish, the type of verb or object are virtually irrelevant, as long as the clitic se is in the sentence. However, English imposes severe restrictions to its middle structures: the sentence must have a measureable verb with a measuring argument in order to be possible.
Two previous studies of object positions have been taken as points of departure. Torrego (1998) analyzes objects in Spanish, while in the same year, Basilico establishes a possible configuration for the English VP containing two potential positions for object generation. In Torrego's analysis, the position where objects check accusative Case (specifier of v), is claimed to be able to assign inherent Case as well as structural Case.
The objects that are assigned inherent Case in Torrego's analysis are all measurers in the sense of Tenny (1994). Likewise, the objects that Basilico claims are generated in the Transitive Phrase in English (one of the two positions), all lead to accomplishmenthood. Thus, I have speculated that the objects of these positions (specifier of v for Spanish and Transitive Phrase for English) check some kind of semantic feature that allows them to be measurers. I have called this feature [measure] and, instead of linking the nature of measurers to a thematic role as in Tenny (1994), I have associated this property with the checking of inherent Case.
I have proposed an analysis of the two constructions (resultatives and middles) in light of my hypothesis, concluding that, in Spanish, the feature is checked by movement of an object to the specifier of v. Checking [measure] is not independent of the checking of accusative structural Case. Contrary to this, in English, the semantic feature is checked by Merge of an argument in the Transitive Phrase, which is not a derived position, but a position for object generation. The consequence is that Spanish bans the resultative construction (most inanimate objects check only structural Case) and allows a wide range of verbs and objects to enter into the middle construction (whose telicity is marked independently of the object, since there is no Transitive Phrase nor a specifier of v available). English, on the other hand, permits resultatives (overt objects that check accusative Case generate first in Transitive Phrase, thus checking [measure] automatically by Merge), but the detransitive mechanism is also constrained, applying only to those objects that are merged in Transitive Phrase.
This can be taken to mean that in some languages, there is a larger inventory of functional projections for object generation, and therefore semantic Case and structural Case are separated. The two positions within the VP in English assign inherent Case: the internal-most position assigns partitive, whereas the Transitive Phrase assigns measure. Accusative Case is assigned in the specifier of v, outside the VP.
In Spanish, objects are generated in only one position, which is the sister-of-V position. This position may assign partitive Case. But only objects marked with structural Case would be assigned the inherent Case that results in measuring out. Thus, structural Case and this type of semantic (inherent) Case are encoded in the same functional projection.
This suggests that languages can either collapse interpretable and uninterpretable features in the same projection, or separate the features in different projections. Accordingly, languages will allow or disallow certain constructions or impose more or less rigid constraints on them. The strongest prediction that results is that languages which severely constrain the middle with regards to [measure], for instance, will allow resultatives and other phenomena not covered in this paper, like delimitation of movement verbs by goal phrases. Languages like Spanish, which disallow resultatives, also disallow delimitation by goal phrases, and impose very few restrictions in the verbs that can enter into a middle construction. For this type of language, an alternative device to check the event type of the sentence, independently of type of object, might be necessary.
Abraham, Werner, Samuel David Epstein, Höskuldur Thráinsson & Jan-Wouter Zwart, eds. 1996. Minimal Ideas. Amsterdam & Philadelphia: John Benjamins.
Baker, Mark & Lisa Demena Travis. 1998. "Events, Times and Mohawk Verbal Inflection". Canadian Journal of Linguistics 43:2.149-203.
Baker, Mark. 2001. The Atoms of Language. The mind's hidden rules of grammar. New York: Basic Books.
Basilico, David. 1998. "Object Position and Predication Forms". Natural Language and Linguistic Theory 16:3.541-595.
Bobaljik, Jonathan D. & Colin Phillips, eds. 1993. Papers on Case and Agreement. (= Working Papers in Linguistics, 18.) MIT Department of Linguistics and Philosophy, Cambridge, Ma.
Bowers, John. 1993. "The Syntax of Predication". Linguistic Inquiry 24.591-656.
Butt, Miriam & Wilhelm Geuder, eds. 1998. The Projection of Arguments: lexical and compositional factors. (= Lecture Notes, 83.) Stanford, Ca: Stanford Linguistics Association (Center for the Study of Language and Information).
Carrier, Jill & Janet H. Randall. 1992. "The Argument Structure and Syntactic Structure of Resultatives". Linguistic Inquiry 23.173-234.
Chomsky, Noam.1995a. The Minimalist Program. Cambridge, Ma.: MIT Press.
Chomsky, Noam.1995b. "Formal Features". Talk given at the Workshop Optimality in Syntax. Is the best good enough? Cambridge, Ma.: MIT, 20 May, 1995.
Chomsky, Noam. 1998a. "Minimalist Inquiries". Ms. MIT, Cambridge, Ma.
Chomsky, Noam. 1998b. "Minimalist Inquiries". Talk given at Kanda University of Foreign Languages. Tokyo, 18 November, 1998.
Collins, Chris & Höskuldur Thráinsson. 1996. "VP-Internal Structure and Object Shift in Icelandic". Linguistic Inquiry 27:3.391-444. Davidson, Donald. 1967. "The Logical Form of Action Sentences". Rescher 1967.81-120.
Dowty, David R. 1979. Word Meaning and Montague Grammar. Dordrecht: Reidel.
Dowty, David R. 1989. "On the Semantic Content of the Notion of 'Thematic Role'". Chierchia et al. 1989.69-129.
Dowty, David R. 1991. "Thematic Proto-Roles and Argument Selection". Language 67.547-619.
Fagan, Sarah. 1992. The Syntax and Semantics of Middle Constructions: A study with special reference to German. Cambridge: Cambridge University Press.
Filip, Hana. 1993. Aspect, Situation and Nominal Reference. Doctoral dissertation, University of California at Berkeley, California.
Filip, Hana. 1996. "Quantification, Aspect and Lexicon". Kruijff et al. 1996. 43-56.
Grimshaw, Jane. 1990. Argument Structure, Cambridge, Ma.: MIT Press.
Hale, Ken & Samuel J. Keyser. 1986. Some Transitivity Alternations in English. (= Lexicon Project Working Papers, 7.) Cambridge, Ma.: Center for Cognitive Science, MIT.
Hale, Ken & Samuel J. Keyser. 1987. A View from the Middle. (= MIT Lexicon Project Working Papers, 10.) Cambridge, Ma.: Department of Linguistics and Philosophy, MIT.
Hale, Ken & Samuel J. Keyser. 1991. On the Syntax of Argument Structure. (= Lexicon Project Working Papers 34.) Cambridge, Ma.: Center for Cognitive Science, MIT.
Hale, Ken & Samuel J. Keyser. 1993. "On Argument Structure and the Lexical Expression of Syntactic Relations". Hale & Keyser 1993.53-109.
Hale, Ken & Samuel J. Keyser, eds. 1993. A View from Building 20: Essays in linguistics in honor of Sylvain Bromberger. Cambridge, Ma.: MIT Press.
Johnson, Kyle. 1991. "Object Positions". Natural Language and Linguistic Theory 9:4.577-636.
Keyser, Samuel J. & Teun Roeper. 1984. "On the Middle and Ergative Constructions in English". Linguistic Inquiry 15.381-416.
Koizumi, Masatoshi. 1993. "Object Agreement Phrase and the Split VP Hypothesis". Bobaljik & Phillips 1993.99-148.
Krifka, Manfred. 1989. "Nominal Reference, Temporal Constitution and Quantification in Event Semantics". Bartsch et al. 1989.75-115.
Krifka, Manfred. 1992. "Thematic Relations as Links between Nominal Reference and Temporal Constitution". Sag & Szabolcsi 1992.29-53.
Levin, Beth. 1996. "Lexical Semantics and Syntactic Structure". Lappin 1996.487-507.
Levin, Beth & Malka Rappaport. 1990. "The Lexical Semantics of Verbs of Motion: The perspective from unaccusativity". Ms., Bar Ilan University.
Levin, Beth & Malka Rappaport Hovav. 1995. Unaccusativity: At the syntax-lexical semantics interface. (= Linguistic Inquiry Monographs, 26.) Cambridge, Ma.: MIT Press.
Levin, Beth & Malka Rappaport Hovav. 1999. "Two Structures for Compositionally Derived Events". Proceedings of the 9th Annual Meeting of Semantics and Linguistic Theory. Matthews, Tanya & Devon Strolovitch, eds. University of California. Santa Cruz, Ca.99-223.
Mclure, William T. 1990. "A Lexical Semantic Explanation for Unaccusative Mismatches". Dziwirek et al. 1990.305-318.
Mclure, William T. 1994. Syntactic Projections of the Semantics of Aspect. Doctoral dissertation, Cornell University, Ithaca, N.Y.
Parsons, Terence. 1990. Events in the Semantics of English: A study in subatomic semantics. Cambridge, Ma.: MIT Press.
Pesetsky, David. 1995. Zero Syntax: Experiencers and cascades. Cambridge, Ma.: MIT Press.
Pustejovsky, James. 1988. "The Geometry of Events". Tenny 1988.19-39.
Pustejovsky, James. 1991. "The Syntax of Event Structure". Cognition 41.47-81.
Ramchad, Gillian Catriona. 1998. "Deconstructing the Lexicon". Butt & Geuder 1998.65-96.
Sanz, Montserrat. 1996. Telicity, Objects and the Mapping onto Predicate Types. A cross-linguistic study of the role of syntax in processing. Doctoral dissertation, University of Rochester, N.Y.
Sanz, Montserrat. 2000. Events and Predication. A New Approach to Syntactic Processing in English and Spanish. Current Issues in Linguistic Theory 207. John Benjamins.
Sanz, Montserrat. In preparation. "Yet Another Try at Unifying the Uses of Spanish se". Ms., Kobe City University of Foreign Studies. Kobe, Japan.
Smith, Carlota. 1991. The Parameter of Aspect. (= Studies in Linguistics and Philosophy, 43.) Dordrecht: Kluwer.
Smith, Carlota. 1996. "The Aspectual Situation Types of Navajo". International Journal of American Linguistics 62.227-263.
Tenny, Carol. 1987. Grammaticalizing Aspect and Affectedness. Doctoral dissertation, MIT, Cambridge, Ma.
Tenny, Carol, ed. 1988. Studies in Generative Approaches to Aspect. (= Lexicon Project Working Papers, 24.) Cambridge, Ma.: Department of Linguistics and Philosophy, MIT.
Tenny, Carol. 1988. "The Aspectual Interface Hypothesis: The connection between syntax and lexical semantics". Tenny 1988.1-39.
Tenny, Carol. 1992. "The Aspectual Interface Hypothesis". Sag & Szabolcsi 1992.1-27.
Tenny, Carol. 1994. Aspectual Roles and the Syntax-Semantics Interface. Dordrecht: Kluwer.
Tenny, Carol. 1995. "How Motion Verbs are Special: The interaction of semantic and pragmatic information in aspectual verb meanings". Pragmatics and Cognition 3:1.31-73.
Tenny, Carol. 2000. "Core events and adverbial modification". Tenny & Pustejovsky (eds.). 285-334.
Tenny, Carol & James Pustejovsky, eds. 2000. Events as Grammatical Objects. Stanford, Ca.: Stanford Linguistics Association (Center for the Study of Language and Information).
Thráinsson, Höskuldur. 1996. "On the (Non-) Universality of Functional Categories". Abraham et al. 1996.253-281.
Torrego, Esther. 1998. The Dependencies of Objects. Cambridge, Ma.: MIT Press.
Travis, Lisa Demena. 1991. "Derived Objects, Inner Aspect and the Structure of VP". Paper presented at the 22nd Annual Meeting of the Northeast Linguistics Society. University of Delaware.
Travis, Lisa Demena. 2000. "Event Structure in Syntax". Tenny & Pustejovsky 2000.145-185.
Van Valin, Robert D. 1990. "Semantic Parameters of Split Intransitive". Language 66.221-260.
Van Voorst, Jan. 1988. Event Structure. (= Current Issues in Linguistic Theory, 59.) Amsterdam & Philadelphia: John Benjamins.
Vendler, Zeno. 1967. Linguistics and Philosophy. Ithaca, N.Y: Cornell University Press.
Verkuyl, Henk J. 1972. On the Compositional Nature of the Aspects. Dordrecht: Reidel.
Verkuyl, Henk J. 1989. "Aspectual Classes and Aspectual Composition". Linguistics and Philosophy 2.39-94.
Verkuyl, Henk J.1993. A Theory of Aspectuality: The interaction between temporal and atemporal structure. Cambridge: Cambridge University Press.
Footnotes
1 Basilico is not the only proponent of this idea. Other researchers who have claimed this structure for the VP in English include Travis (1991), Johnson (1991), Bowers (1993), Koizumi (1993) and Collins & Thráinsson (1996).
2 As Levin & Rappaport Hovav point out, (i) cannot mean that Willa became breathless as a result of arriving. Breathless is, rather than a resultative, a depictive phrase.
(i) Willa arrived breathless
These authors propose that verbs of inherently directed motion cannot take a resultative clause because they already specify an achieved endpoint and they may not take a second encoded delimiter specifying a change of state. They contrast these with break, which describes the attainment of a state, not of a location, and can occur with a resultative phrase.
3 I thank an anonymous reviewer from John Benjamins for having brought this example to my attention.
4 I thank Eloise Jelinek (p.c.) for this observation.
5 In any case, middles seem to be exploited in a type of discourse that intends to be marked, like commercial discourse. For this reason, judgments vary widely from speaker to speaker and from verb to verb, which makes the analysis difficult. I will continue to assume that, in spite of these differences, the major factor distinguishing a good and undisputed middle is the fact that the event is measured out.
6 A possibly fruitful path to follow would be to consider the semantic zones for adverbs that Tenny (2000) proposes. As I said, I leave this for future research.
7 In this paper, I am concentrating only
on the interaction between imperfectivity and accomplishmenthood that causes
an iterative interpretation. Therefore, it is only the telic aspect of se
that I focus on. For a detailed analysis of se beyond telicity and
in relation with global properties of events and sentence predication, see Sanz
(in preparation).
8 It also follows that the objects of achievements and stative predicates are generated in the other object position, internal to the V' (since they are never measurers). Hence, these constructions should be thetic. This latter part of the argument will not concern us in this paper.
| ISSN: 1139-8736 |