Publications Pages Publications Pages. Search my Subject Specializations: Select Users without a subscription are not able to see the full content. Lexical Semantics, Syntax, and Event Structure Malka Rappaport Hovav, Edit Doron, and Ivy Sichel Abstract This book focuses on the linguistic representation of temporality in the verbal domain and its interaction with the syntax and semantics of verbs, arguments, and modifiers.
More This book focuses on the linguistic representation of temporality in the verbal domain and its interaction with the syntax and semantics of verbs, arguments, and modifiers. Authors Affiliations are at time of print publication. Print Save Cite Email Share. Show Summary Details. Subscriber Login Email Address. Library Card. View: no detail some detail full detail.
Once admitted into our ontology of entities, events can be shown to have linguistically relevant different internal topological properties cf. They have been central in the se- mantic understanding of pluractionality and distributivity, on a par with plurality in the nominal domain Schein , Lasersohn , Landman Events and objects have been shown to be linguistically commensu- rate, as shown by the systematic interactions found between nominal and temporal reference and the construction of telicity, as discussed in Verkuyl Verkuyl , Filip and Krifka , Krifka the latter ex- ploring a mereological approach based on the lattice theoretic framework of Link Events have also been exploited as discourse entities in theories of discourse representation as in Kamp and Reyle In what follows, I will take as my starting point the premise that events need to be repre- sented in the semantics to deliver an adequate account of generalizations concerning verb meaning within natural language.
In this short article, I will try to lay out a perspective on this question, moti- vated by the idea that syntactic representations and compositional semantic representations track each other in a systematic and predictable way. By this I do not mean merely that the interpretation of structure proceeds by function-argument composition—- most people already believe this.
In fact, the power of the lambda calculus allows the expression of some quite non- transparent relationships between structure and interpretation, due to type shifting. Under the comfortable darkness of complicated higher order func- tions, linguists have been living with highly non-trivial and unconstrained mappings between form and meaning.
Consider sentence 1 below and its Neo-Davidsonian semantic representation. The problem essentially is that the semantic facts seem to require that the event variable be existentially bound low, as low as possible in fact, since it never seems to interact quantificationally with other elements. On a theory in which the event variable is closed off at the sentence level, this is technically awkward to achieve. However, recent work by Champollion pear proposes a formal solution to integrating the event based semantics which is based on low existential closure of the event variable in the verbal denotation, and which is more modular therefore with respect to what one chooses to do later in the semantic composition.
However, since my concern in this paper is with the internal structuring of the verbal denotation and how it tracks the internal syntax of the VP, I will be abstracting away from this complication here, and using standard Davidsonian representations. This would be fine if every verb had to be memorised with its own idiosyncratic list of arguments in a stipulated order.
But if nothing further is said, we could also give a representation for V as in 4 below, which would correctly deliver 1 for the syntactic representation in 5. We can see from the very flexibility of the lambda calculus, that there is no logical necessity for such a generalization. It is simply a fact about human language, and as such we are responsible for accounting for it. Since we know that there are strong generalizations about the ways in which argument structure maps onto syntax, we need a way to state these constraints somewhere in the grammar. This ends up invoking hierarchical structure within lexical repre- sentations, in addition to a mapping rule which requires the preservation of that hierarchical structure.
- Slow Sex: The Art and Craft of the Female Orgasm.
- Meaning and grammar.
- The Lexical Encoding of Idioms *.
I give an example of such a tree structure in 6 below. We have a number of different options here in stating a compositional semantics over this structure. One option is to stipulate that the verb, although it moves from V to v, is only interpreted in one position— the low V position. Then we are left with exactly the same semantic interpretation for the verb that we had before in 3 , as long as we have a convention that allows vacuous merge, which will simply pass up the compositional semantics of the non-vacuous daughter.
However, the whole point of the complex syntactic representation in 6 was to be able to express generalizations about argument structure semantics and syntactic hierarchy, otherwise we could have just as well made do with the traditional syntactic tree. I choose to implement the generalizations in a different architecture, but the nature of the generalizations is actually the same as those found in the lexical tradition. The field has largely converged on an understanding of what the patterns are, although there is disagreement about the model of grammar that underpins it cf.
The compositional semantics for such a tree, with v interpreted as some kind of causing projection, and V as some sort of caused dynamic event, could look as in 7 , one that to my knowledge nobody has so far seriously proposed. This underspecified event description is then the argument of the higher predicate stab which provides the conceptual content of the event. Of course, even this proposal would have to stipulate the order of merge of the two arguments to deliver the correct alignment of thematic roles in practice, and so is subject to the same criticisms and modifications as in the previous, more recognisable trees.
At the risk of belabouring the obvious point, the purpose of this section has been to show that a commitment to events in the semantic represen- tation of verbal meaning, and a Neo-Davidsonian implementation in terms of severed thematic relations massively underdetermines the syntactic rep- resentations that could go along with it. To the extent that natural language has some observed systematicities concern- ing the ways in which events and event descriptions are built up, we need to add these in explicitly, either as modifications of the semantic represen- tations of verbs themselves in some kind of lexical representation, or as a constrained output of semantic composition.
As such, robust generalizations in this domain are an important clue to the deep nature of natural language representations in particular, and to the mapping between syntax and semantics. The structure of the paper is as follows. In the next section, I will briefly outline the linguistic facts that show clearly that the syntax is sensitive to and hierarchizes internal and external arguments differently, and well as the different components of cause and result in event semantics.
None of this data is new and many of the generalizations are uncontroversial. Next, I will lay out a proposal for the decomposition of the verb phrase on which both argument hierarchies and subevental hierarchies can be built. I end by showing that differences in lexicalization of these structures is all that underpins some superficially great differences in surface form.
The paper closes with a discussion of the relationship between syntax, semantics and lexicalization that is required for basic explanatory adequacy in this domain. For many purposes the exact nature of the labels and their definitions are ignored or put off. Unfortunately, finding the precise labels and their definitions is ex- actly the heart of the problem, and reliable entailments over the proposed content of traditional labels have proved elusive see Dowty for an im- portant discussion.
Many authors now favour an extremely reduced, more event oriented set of roles for structurally licensed positions in the clause either in addition to, or completely supplanting traditional theta role labels Grimshaw , Jackendoff , Baker Thus, to mediate the connection to the syntax, it is now widely acknowl- edged that lexical representations need to include event structure templates, or abstract representations of force dynamical interactions, in parallel to other kinds of conceptual information Levin and Rappaport Hovav , Pustejovsky , or action tier Jackendoff Croft In the following subsections, I summarize what I take to be the facts about subject and object selection that have emerged over the course of 50 years of research into argument structure patterns.
The field has made genuine progress in this area, and although many details remain to be worked out, certain robust generalizations have become clear. However, choice of subject is clearly shows correlations to event structure notions across languages, suggesting that hierarchies un- derpinning event participancy feed the choice of syntactic subject. John broke the window.
Intentional Agent b. The strong winds broke the window. Inanimate Cause c. The iron key opened the old rusty lock. Instrument d. The stone hit the floor. To clarify, I think it is important to recognise that there is no objective way of isolating the cause of a particular dynamic change in the world. There are also no doubt constraints on the cognitively natural ways in which human beings construe things as being caused.
Consider now the causative-inchoative alternation, shown in The stick broke. John broke the stick. What is important to note about this alternation is that it is extremely common and pervasive crosslinguistically see Haspelmath for a typo- logical study , and that the addition of an expressed causer is what makes the difference between the transitive and the intransitive version.
No theory of argument structure can ignore this kind of relationship between events, or the idea of causer as a more prominent participant when it comes to Subject selection. One might even argue that Causer or Initiator in a general sense is prototypically the most prominent participant in any event structure.
This is consistent with the general con- sensus from the thematic role literature, where either Causer or Agent sit on top of the thematic hierarchy. Turning to monotransitives, we find evidence for the difference between internal and external arguments. The famous unaccusative-unergative dis- tinction Perlmutter refers to the important grammatical difference in the behaviour of monotransitive verbs, correlated with participant role.
Some accounts propose a purely semantic i. Thus, most of these accounts assume that there is a structural difference between an unaccusative phrase structure and an unergative one, which underpins their different syntactic behaviour. One problem often cited with the notion of unaccusativity is that trans- lations of an unaccusative verb in one language do not always straightfor- wardly yield an unaccusative verb in another language, even where both languages make the distinction clearly Zaenen for the comparison be- tween Italian and Dutch.
However, this only shows that behaviour cannot be predicted directly from the semantics of real world situations, but that facts about situations in the world feed, but underdetermine the way in which events are represented linguistically. Plausibly, only linguistic repre- sentations are symbolic and categorical; the real world is messy and open to different choices of representation in many cases. It seems to be this sort of initiating or facilitating argument that is privileged when it comes to Subject selection, when in competition with an argument that merely undergoes change.
Once again, the generalization goes in the direction from linguistic repre- sentation to entailments. I leave aside for now the question of how individual lexical verbs carry information that allows them to match up to and lexi- calize different event structures possibly in a non one-to-one fashion , and take it up again in section 4. In fact, there is good evidence that Objects hold a semantically priv- ileged position with respect to certain kinds of event entailments.
This notion, though now widely embraced, has been implemented in rather different formal ways in the literature, not all compatible. John loaded the hay on the truck. John loaded the truck with hay. Classic data from Verkuyl , Verkuyl cf. John ate the the sandwich? As Hay et al. In the following three exam- ples, we see that three equally respectable direct Objects: undergoers of locational change 14 a , property change b or material affectedness c see Ramchand and Hay et al.
John pushed the cart. Mary dried the cocoa beans. Michael stretched the rubber band. As one can easily demonstrate, the mere existence of an Undergoer does not necessarily imply telicity, as the English examples in 15 show. The chocolate melted for hours. John melted the chocolate for hours. John pushed the cart for hours. John pushed the cart to the store in an hour. Thus, while being the undergoer of a change, and achieving a defi- nite change of state often go together on a direct Object, the two notions are logically separable.
Ramchand b calls the entailment type for the participant that achieves a change of state the Resultee. The fol- lowing sentences from English show a pure Undergoer and a composite Undergoer- Resultee role respectively. Undergoer; no transition to final state b. Undergoer-Resultee;transition to final state 4 In Ramchand b , participants can accrue entailments by moving through A- positions, so some thematic roles in the traditional view are actual composite from the point of view of the primitives of initiator, undergoes and resultee.
The reader is invited to consult that work for details. The Path in this sense is not a species of Undergoer at all, but comple- mentary to it: in 17 , the path describes the ground that the Undergoer traverses. In 18 , we see examples of Undergoer, Undergoer- Resultee, Path, and even Measure in object position although the latter type of Object is notorious in not showing all the canonical proper- ties of direct Objects in some cases. John rolled the cart. Undergoer b. John rolled the cart over. Undergoer-Resultee c.
John destroyed the cart. Undergoer-Resultee d. John walked the West Highland Way. Path e. John ate the apple. Path f. I have argued here that our current knowledge shows that there is indeed a privileged relationship between the internal argument and the path of change represented by the dynamic event.
Even though there 5 It is often claimed after Verkuyl that Objects are those whose quantizedness has a direct effect on the felicity of the resulting VP. Path vs Resultee, being related to the path of change gives an argument privileged status when it comes to the object relation and accusative case. This special feeding relationship with grammatical objecthood is one which all theories of argument structure effects need to deliver. Across languages, morphology is required to demote the causing argument if another kind of participant is to be promoted to subject.
I am not claiming here that Subject and Object are thematic notions. On the contrary, they are grammatical notions which seem to be fed in a systematic way by argument structure. The nature of this feeding suggests that for the purposes of Subject selection, causers are structurally the most salient, closest to the attracting head or feature; for the purpose of Object selection, which we know to be a structurally lower position, it is the path and path-related notions that are structurally the closest to the attracting head or feature.
As far as lin- guistically relevant distinctions are concerned, it is clear that the different classes have different behaviours, as evidenced by the linguistic diagnostics used to distinguish them in the literature see Dowty Ap- plicatives have also be treated more recently as functional heads in their own right which introduce arguments of certain types in their specifier position. A detailed treatment is beyond this scope of this particular article.
However, the pattern of unmarked alternations vs. Events The state vs. Event distinction is very prominent in English and comes with a robust list of criterial behaviours, revolving around the different ways in which these two categories interact with tense. I illustrate with the present tense here. John likes mangoes. State holds now b. John runs the race. On the other hand, eventive verbs in the simple present tense cannot be given the interpretation that the even- tuality is ongoing at the present time for that, we need the progressive.
Instead they receive habitual or planned future interpretations, or can be used in vivid past tense narrations. Intuitively, while dynamic eventualities denote some eventuality which includes change over time, states denote unchanging situations. It is per- haps not surprising that they interact in very different ways with temporal specification and modification. The argument structure of stative verbs also does not conform to the generalizations stated above for dynamic verbs.
When it comes to spatial relationships, Talmy defines Figure as the entity whose spatial location, or movement through space is at issue, while the Ground is the entity with respect to which that position or motion is defined. Saliency and functional considerations are often relevant in determining which entity in a static eventuality is chosen as the bearer of a property ascription. The bearer of a property ascription Figure, or Holder then contrasts with the non-Subject participants in a static eventuality which provide additional information specifying the property being ascribed.
Accomplishments are often seen as complex, as consisting of both a process portion and a result Pustejovsky , Higginbotham , while activities are pure processes. Here too we find linguistic tests distinguishing the three different aktionsart, although tests for distinguish- ing achievements from accomplishments are rather less sharp see Dowty for details and discussion, and Mittwoch, this volume. In general, then, there is much linguistic evidence for the four natural classes of event shape as laid out in 21 taken from Truswell, this volume.
Let us simply represent these as two primitively different kinds of eventuality and notate them as es and ed in what follows. However, such an analogy underplays the internal complexity of ac- complishments, and underplays too the deep linguistic cut between dynamic activities and states despite the fact that they have the cumulative property in common. There is, further, an interesting difference between individuals and events in this regard. Mereologies of individuals are stated in terms of material part-whole relations which are rather easy to match up with real world in- tuitions of subparts.
On the other hand, deciding when something counts as a subevent of another event is fraught with paradoxes and technical semantic difficulties and is most often thought to require inertial worlds or some such intensional machinery Dowty , Landman because of the notions of intention and causation that link parts of one event to another. I do not propose to solve any of these technical issues here with respect to truth conditions, but merely mention them as a prelude to sidestepping them altogether. There are many people already working on these issues, and on developing precise truth conditions for non-culminating events with notional results.
In this article, I take the position that these relationships are in some sense axiomatic, and fundamental to our cognitive structuring and individuation of eventualities in the world. Thus, I will simply take abstract causational, or force-dynamical glue to be a primitive of subevental combination and ask how far simple recursion of these primitives can take us in building more complex events up from simple ones. Causational or force-dynamical relationships could potentially build ex- tremely complex event chains from these building blocks, indeed complex networks of interacting eventualities.
However, when it comes to meanings that are lexicalized as single monoclausal verbal domains, the situation is interestingly constrained. With respect to the addition of result, the data also suggest that only one such delimitation per event is possible Simpson , Tenny on the unique delimitation condition. Thus, the typology we see can be created by augmenting the dynamic core event with either a causally upstream or causally downstream state, but no further.
I have discussed some of these cases in previous work Ramchand and Svenonius , Ramchand a, Ramchand a , so I simply repeat the well-known examples from English below. John ran. John ran his shoes ragged. There have been arguments in the literature that subevental decomposition of this type is syntactically real and can be diagnosed via certain kinds of adverbial modifiers such as almost and again von Stechow , Beck and Johnson With regard to point i , once monoclausality is controlled for, I do think with Williams pear that there are some irreducible encapsulation effects that arise as a product of word formation, even though in this theory, word formation feeds off the syntax.
As far as point ii is concerned, I would take the data to indicate, contra Siloni, that those decompositions are in fact warranted for acquire and not in fact for dry. Be that as it may, syntactic representation per se is not the most important point I wish to make in this article. Specifically, the causing event, when it can be seen to be explicitly added, always adds morphology or participants that is hierarchically above the core dynamic event; result events are always added below the core dynamic event. Moreover, the Cause event is associated, when it is, with a higher argument, whereas the result predicate either introduces a new internal argument or is constrained to modify it Levin and Rappaport Hovav These points are not new or controversial, but it is worth pointing out that they are in some sense so natural that their remarkableness sometimes escapes attention.
This would be remarkable enough. In fact, the situation is even more interesting, because the two kinds of generalization actually converge. The conjecture built into my own research agenda, as most explicitly laid out in Ramchand b is that we are looking at two faces of the same generalization. To see this, we need to allow two basic versions of thematic role: a dynamic one and a static one. The static one is just the general Holder re- lation from Kratzer or generalized Figure relation from Talmy , which has the definition and simple specifier-complement structure shown in The subject of predication is the Holder, and the complement is the Ground.
This creates the Undergoer relation, and it has the same specifier complement structure that the stative predi- cation above has, with the only difference that the predicational head here is dynamic. The complement of a predicate of change is Path. I propose to limit recursion to structures with a maximum of one dynamic predication per event phase. This is a constraint that comes from our gen- eral cognitive relationship to event perception— independently perceived dynamic change corresponds in interpretation to a distinct event.
The maximally expanded subevental structure for caused changes lead- ing to a result, would look as in 28 , with a stative predication embedding a dynamic one, and the dynamic one in turn embedding a stative one. The- matic roles do not need to be listed separately, nor do their properties need to be memorized or known in advance. Causative embedding will ensure the relative prominence of the different argument positions, and the mini- mal relationships of property-holding both static and dynamic will derive specifc and different entailments for the different positions.
This is just a fancy way of saying initiator. The labels on the tree in 28 should therefore be seen not as labels in a template, they are there for ease of readability. The functional sequence here is actually quite spare, once the effects of hierarchy and predication are factored out. Bounded paths give rise to verb phrases that are classified as accomplishments in the literature, while un- bounded paths give rise to activities.
As I hope to have shown, the limited internal decomposition we do find corresponds in granularity and type to the event decomposition we need to deliver the argument structure generalizations as well. This is true even for languages which do not show any overt morphological evidence for decomposition in their verbal lexical entries. I will take it therefore that we have strong reasons to accept this structure even in cases where the morphology does not spell it out explicitly, since we thereby gain an account of both aktionsart, thematic roles, their correlation and their interaction with other syntactic processes.
The answer depends on what ones assumptions about lexicalization are. And there are many. I will argue in what follows that the only real argument against de- composed syntactic structures is one theory-specific stipulation about lexical attachment which has no independent motivation, i. Those who operate with a fully generative and structured lexicon, could employ a structure such as this within the lexicon as a module. I take this to be in fact the general position found in recent work by Levin and Rappaport Hovav.
However, there are a host of other sorts of rich conceptual meanings and specific features of events that are contributed by specific lexical items. There are at least three different families of approach to the general prob- lem of the integration of structural and conceptual components of meaning. Under one view, these meanings both reside within the lexicon and are com- bined in parallel i.
This therefore involves parallel representation, with the combination of Lexical information in series with Syntax. I take this to be the approach of Beth Levin and Malka Rappaport Hovav in their seminal and decades long contribution to the topic of verbal meaning Levin and Rappaport Hovav , Levin and Rappaport Hovav We could label this view, the Lexicon-Internal Unification approach.
Syntax, Lexical Semantics, and Event Structure (Oxford Studies in Theoretical Linguistics)
The syntacticization of meaning espoused in many constructivist ac- counts, such as DM as described above, introduces a new problem for the unification of structural and conceptual aspects of meaning. The reason for this is that said syntacticization has not gone hand in hand with revision to ancient modes of lexical insertion. So while structural meaning has been decomposed, conceptual content can still only be inserted holistically. In DM this needs to be recaptured by means of multiple contextual allomorphy rules and listed selectional frames at the point of vocabulary insertion.
To my mind, this undermines quite a lot of the attractiveness of the decompositional approach. It is unclear to me however, the extent to which a strict distinction between structural and conceptual aspects of verbal meaning is actually made. Generalizations, when they exist, are abstractions over listed patterns. But this makes the approach more similar to the parallel unification view of Ramchand, described in the next paragraph. If, in response to categorial decom- position, lexical items now come associated with a set of categorial features that indicate which parts of the skeleton they are designed to conceptually specify, we could keep the parallel view.
Essentially, then, each lexical item would constitute a set of cross modular associations, connecting syntactic labels to conceptual content. Conversely, be- cause structural meaning is extremely abstract. The lexical item is thus a chunk of cross modular associations, combining one or more category features with the concep- tual content that fleshes out its description cf. Meaning combination proceeds by the parallel unification of lexical conceptual information and syntactic-structural information at every stage of the derivation.
- (PDF) Event Structure and Verbal Decomposition | Gillian Ramchand - faxarytogy.tk.
- Nets of awareness: Urdu poetry and its critics.
- Dead Reckoning (Southern Vampire Mysteries, Book 11)?
- Algebra VII - Combinatorial Group Theory, Applns to Geometry.
- Event semantics and adverbial modification.
We could call this the Cross Modular Unification approach. This kind of relationship between structure and lexicalization is en- shrined in the Exhaustive Lexicalization Principle stated below, together with the principle of Non-Terminal Lexicalization which expresses the idea that conceptual content can potentially be associated to chunks of struc- ture. If we get this right, it can account for both flexibility and selectional rigidities Non-terminal lexicalization is a way to accommodate structurally decom- 12 For a technical implementation of how lexicalization can be stated over category bun- dles, without either word order movements or fusion operations, see Bye and Svenonius See Fabregas for extensive discussion of its effects in the domain of Spanish directional complements.
I have done this i to emphasize visually the commonalities in the three cases, ii to visually separate the head contributions from the phrasal elements and iii most importantly , to emphasize the fact that these trees are intended to represent hierarchical relations with no implications of linear order.
I assume that linearization for language, is a language-specific and largely autonomous process that I put aside here. Quite systematically, aspect appears outside of the main verb stem, and tense in turn appears outside of that. They then line up sentence finally as V-Asp-T. This is exactly the order you would expect from a head final language with this proposed hierarchical structure.
Note that the Bengali complex predicate construction shown above, and even the English particle verb construction have otherwise posed paradoxes for lexicalist theories of argument structure. On the one hand, they are clearly morphemically compositional, and it can be shown that the compo- nent parts are even independent syntactic units. On the other hand, the combination of lexemes changes the argument structure properties some- thing that lexicalists assume to be in the domain of the lexical module and the constructions are monoclausal by all diagnostics.
The view proposed here accounts for the predicational unity of the complex predicates as well as their resultative semantics. In all of the above examples, it is still possible to conceive of lexical insertion in a more traditional manner under terminal nodes, with head-to- head movement in the syntax, or in the morphology as the need arises. The event structure decomposition I have proposed above bears close similarities to some of the event structure templates proposed by Rappaport- Hovav and Levin They also are concerned to express constraints on the fitting together of lexical items and event structure templates.
In recent work, they have been examining the conjecture that while event schemas involving both manner and result exist in most languages, it is not possi- ble for a single lexical verb to be associated with both at the same time Rappaport-Hovav and Levin If Manner-Result complementarity is a correct generalization then it says something important about the way in which lexically encoded conceptual content can be paired with event structure skeletons of the type shown above. However, it is an idea that emerges most naturally under a system that expects lexical items to insert under a single terminal node.
Interest- ingly, Rappaport-Hovav and Levin concede that the lexicalization constraint as they state it must apply not to whole verbs but to simplex forms more generally. In other words, when morphemes can be seen to combine productively to create verbal lexical items in certain languages, the lexicalization constraint applies to the individual morphemes, not to the verb itself. In fact, I think there are a number of relatively clear cases of verbs in English whose lexical conceptual meaning contributes con- tent to more than one element of an abstract event schema.
Perhaps even clearer, the verbs to lay and to stand require there to be a physical placement of an object on the part of an agent, and they also require a specific orientation of the object in its final result state. These verbs contribute lexical conceptual content to both process and result sub events.
While there seems to be some suggestive evidence in the direction of manner-result complementarity, we need to ask ourselves whether the pat- terns are absolute or merely tendential, because of natural overall limits of usability on the complexity and specificity of lexical items. In fact, by the end of the article, Rappaport-Hovav and Levin modify the manner- result complementarity prohibition to one that rather involves the incom- patibility of scalar vs.
The idea here is that a morpheme cannot conceptually describe both a scalar change and a non-scalar change simultaneously. This places the incompatibility within the semantics of con- ceptual content, and not with a constraint on how monolithic lexical items can be associated with event templates. I will thus continue to assume that lexical items can in principle bear conceptual content that describes more than one subevent in the kinds of decompositions shown above.
This will account for the generalizations map- ping argument structure to syntax for all lexical items, as well as the natural classes of verb aktionsartal types. Thus, the semantics of event structure and event participants is read directly off the structure, and is fleshed out by lexical items.
Lexical semantics - Wikiwand
In the attempt to derive argument structure and aktionsartal general- isation from a very minimal set of primitives and combinatoric principles, the agenda pursued here is very similar to that pursued by Pietroski see also Lohndal, this volume. To summarize my own position again, the semantic combinatoric prin- ciples proposed in the present article can be described as follows, each cor- responding to a different syntactic configuration.
We could get rid of the causational glue for event-event embedding if we allowed ourselves more specific category labels for the functional sequence within the lower verb phrase. We would retain conjunction as the semantic glue at the expense of reifying Cause and Result in the ontology, and indeed with stipulating their position in the hierarchy.
Clearly worse, though, would be to get rid of ii.
- Login using.
- Navigation menu.
- Closed-End Investment Companies: Issues and Answers.
- Citations of: The syntax of event structure - PhilPapers;
Recall that ii is necessary to systematically link up participants with hierarchically ordered event descriptions. Moreoever, there is nothing in the system that makes it necessary that the functional head that introduces the Theme occurs before the functional head that introduces the Agent.
We are essentially back to the situation we were in with the trees shown in section 1—- the semantics of conjunction coupled with the freedom to stipulate thematic role labels does not depend on any particular hierarchical representation in order to work. However, we have seen that hierarchical order is precisely what natu- ral language is showing us in this domain. The hierarchical generalizations and the close correspondence between event decomposition and argument structure go completely unexplained.
Instead, the claim here is that there is a primitive cognitive no- tion of property ascription that natural language symbolic structures hook up to systematically. Because of the natural semantic glue connecting sub events, the particularities of the- matic participancy emerge as a natural consequence of the ways in which sub events relate to each other. In this paper, I have tried to argue that Post-Davidsonian decomposition is an elegant and parsimonious way of accounting for the robust linguistic generalizations that exist concerning the typology of verbal meaning, their arguments and the syntax.
Because of the power of both the syntax and semantic tool boxes, we do not yet have a theory that allows these patterns to fall out naturally. Until we make highly specific and constrained pro- posals of this type about the mapping between syntax and semantics, these generalizations will remain unexplained. Mateu Cuervo and Y. Roberge Eds. Bingley: Emerald.
Related Syntax, lexical semantics, and event structure
Copyright 2019 - All Right Reserved