1 BASICS OF THE THETA SYSTEM 1 2 VERB

The general picture: The θ-system (which corresponds, roughly, to the Lexicon in linguistic theory) is the system enabling the interface between the systems of concepts and the computational system (syntax) and, indirectly (via the syntactic representations), with the semantic inference systems. In the modular view of Fodor and Chomsky, the cognitive systems operate independently of each other, and generally, the information processed in any given system is not legible to the others. But for the interface to be possible, each system should contain also some information that is legible to other systems. In Fodor's framework, these are gathered by a central system responsible for passing on the information. The θ-system can be viewed, then, as the central system of the systems of concepts. As in the traditional view of the Lexicon, I assume the system is computational (-it contains both a storage of lexical items, and rules operating on these items), contra Marantz 1997, 2000, Borer (2004), among others, who reduce the lexical component to a list of roots.

1) The Theta system consists of :

a. Lexical entries, which are coded concepts, with formal features defining the θ-relations of verb-entries.

b. A set of arity operations on lexical entries, which may generate new entries, or just new options of realization.

c. Marking procedures, which 'prepare' a verb entry for syntactic derivations: assign an ACC(usative) feature to the verb in the relevant cases, and determine merging properties of arguments (which arguments merge externally and which internally).

The topic of this lecture is (1a). An overview of the full system can be found in Reinhart (2002).

The θ-features

3. Some history and motivation for θ-features instead of θ-roles^¹

A puzzle of θ-selection (Reinhart 1991, 1996)

2 a) The father/*the spoon/*hunger fed the baby.

b) Max / *the leash / *hunger walked the dog to his plate.

c) The baby/ *the spoon /* hunger ate the soup

d) Lucie / ??The razor /*the heat shaved Max.

e) Lucie / *the snow / *the desire to feel warm dressed Max

3 a) Max opened the window (in order to enter).

b) The key opened the window (*in order to be used).

c) The storm opened the window (*in order to destroy us).

4 a) The painter / the brush / autumn reddened the leaves.

b) Max / the storm / the hammer broke the window.

c) Max / the heat/ the candle melted the ice.

d) Max /exercises /bicycles developed his muscles.

e) Max / the storm / the hammer enlarged the hole in the roof.

The unaccusative problem.

5 a) Transitive: Max broke the window.

b) Unaccusative: The window broke.

The prevailing assumption till the nineties: Unaccusatives are listed as basic entries. Transitive alternates (known as the causative -incohative alternation) were assumed to be the derived form.

Chierchia (1989): All unaccusatives are derived by a lexicon reduction rule from a transitive base entry.

6) V (θ₁, θ₂)---> R_e(V) (θ₂)

Remaining problem: over-generation.

7 a) Lucie ate the soup.

b) *The soup ate.

Where to search the answer: While there is no property shared by the outputs of unaccusative reduction, the transitive inputs share (universally for virtually all entries) the property of allowing agent, cause, or instrument subject, as in (4).

4. Basics of the features system.

Objective: Define the linguistic coding of causal relations. (Jackendoff's (1990) 'actor' tier, but including roles like goal.)^²

8) /c = Cause change; /m = Mental state

agent cause/instrument theme/patient … experiencer

/c: + + - -

/m: + - - +

9) θ-Clusters: [+c+m] [+c-m], [-c-m] [-c+m]

Contextual interpretation of θ-clusters.

Dowty (1991): the meaning of θ-roles is often contextually determined.

Lexicon generalization: (cause and instrument):

10) A [+c-m] cluster is an instrument iff an agent ([+c+m]) role is also realized in the derivation, or contextually inferred (cause otherwise).

5. The solution to the unaccusative problem: The unary [+c] cluster.

4 b) Max / the storm / the hammer broke the window.

Agent/ cause / instrument

11) a) break ([+c], [-c-m])

Contextual interpretation:

b) [+c] spelled out as either [+c+m] or [+c-m]

c) [+c-m] is interpreted as instrument or cause, by (10).

12 [+c] reduction (Expletivization):

V ([+c], α) → V (α)

-With very few exceptions, all([+c],[-c-m]) verbs have an active unaccusative alternation.

-With very few exceptions, all unaccusative verbs have a transitive [+c] alternate in one language or another. (In Hebrew, the transitive alternate for come (ba) is bring (heb'i))

Experiencing verbs alternations

13) a) Something worried Max

b) Max worries

c) puzzle, grieve, delight

14 a) De discussie verbaasde Lucie (The discussion surprised Lucie)

b) Lucie verbaast zich (Lucie was surprised)

15 worry, surprise ([+c], [-c+m])

a) Lucie / the heat wave / the gun pointed at him worried Max.

b) Fred/ Fred's behavior /the discussion surprised Lucie

Fred/ Fred's gedrag /de discussie verbaasde Lucie.

6. Levin and Rappaport's (1995) alternative analysis of unaccusatives.

16 The unaccusatives derived from a transitive entry are those which denote eventualities that are

a. "externally caused" and

b. "can come about spontaneously, without the volitional intervention of an agent" (L&R, p. 102).

17) A one-place verb describes an internally caused eventuality if "some property inherent to the argument of the verb is 'responsible' for bringing about the eventuality "(p.91).

18 a) The diamond glowed. (unergative)

b) The vase broke. (unaccusative)

19) "Inherently caused" unaccusatives, are base generated and listed individually in the lexicon: exist, happen, come, arise, remain, appear, stand (in the 'simple position' reading, as in The statue stood *(in the corner))

20) Virtually all verbs in (19) have a [+c] transitive alternate in Hebrew. E.g.: xolel/hitxolel (bring about/ happen), kiyem/hitkayem (carry out)/ take-place, exist), herim / hitromem (raise /arise) (Reinhart, 2000, section 3.4.1)

7. The full set of feature-clusters.

The two binary features: +/-c = Cause change and +/-m = Mental state define the nine feature clusters in (21). The correspondence of these clusters to the known θ-roles is not always one to one - often they have varying contextual interpretations. I label them by the role that they are most typically related to. Further details are provided in the footnotes.

21 a) [+c+m] - agent

b) [+c-m] - instrument (...)

c) [-c+m] - experiencer

d) [-c-m] - theme / patient

e) [+c] - cause^³

f) [+m] - sentient^⁴

g) [-m] - subject matter /locative source (Typically Oblique)

h) [-c] - goal / benefactor(Typically Dative (or PP))

i) [ ] - Arb(itrary)^⁵

8. Theta assignment and merge.

Derivation of (2)

2) Max washed the child.

9. Interpretation (The Inference interface).

22 Event-semantics representations (as, e.g. in Parsons, 1990)

a) Verb entry: wash [Agent] [Patient ]

b) Max washed the child.

c) e (wash(e) & Agent(e)=Max & Patient(e)=the child))

d) e (wash(e) & [+c+m](e)=Max & [-c-m](e)=the child))

23 [[y do something] CAUSE [x become BROKEN]] (L&R (1995) and others)

Basic relations in the perception of causality (shen 1985)

Enable:

24 a) Max entered the pool and drawned.

b) The relation enable holds when one event is perceived as a necessary condition for the occurrence of the second

Cause:

25 a) The vase fell on the floor and broke.

b) The relation cause holds when the first event is conceived as a (necessary and) sufficient condition for the second

Interpretative properties of the θ-clusters:

26 a) A /+c feature is associated with a participant whose relation to the event denoted by the verb is perceived as providing (by its existence or actions) a sufficient condition for that event taking place.

b) A /+m feature is associated with a participant whose mental state is relevant for the event

c) All selected θ-cluster stand in the enable relation to the event.

References.

Borer, Hagit (2004) The Grammar Machine. In The Unaccusativity Puzzle, ed. by Artemis Alexiadou, Elena Anagnostopoulou, and Martin Everaert. 288-331.Oxford: Oxford University Press.

Chierchia, Gennaro (1989) A Semantics for Unaccusatives and its Syntactic Consequences. Ms., Cornell University, Ithaca, N.Y.

Chierchia, Genaro (1995) 'The Variability of Impersonal Subjects', in E. Bach, A.Kratzer, E. Jelinek and B.H. Partee ( a cura di) Quantification in Natural Languages, Kluwer, Dordrecht and Boston, pp. 107-144.

Dowty, David (1979) Word Meaning and Montague Grammar, Reidel, Dordrecht.

Dowty, David (1991) Thematic Proto-roles and Argument Selection. Language 67, 547–619.

Grimshaw, Jane (1990) Argument Structure. Cambridge: MIT Press

Jackendoff, Ray (1987) 'The status of Thematic relations in linguistic theory', Linguistic Inquiry 18.3, 369-411.

Jackendoff, Ray (1990) Semantic Structures, MIT press, Cambridge, Mass.

Marantz, Alec (1997) “No escape from Syntax: Don't try a morphological analysis in the privacy of your own lexicon”. U. Penn Working Papers in Linguistics 4.2: 201-25.

Marantz, Alec (2000) Reconstructing the Lexical Domain with a Single Generative Engine, ms. MIT.

Marelj, Marijana (2002) 'Middles in Dutch/English Type of Language.' LINK 4: 3-73. Utrecht: University of Utrecht.

Parsons, Terry (1990) Events in the Semantics of English: A Study in Subatomic Semantics, MIT Press, Cambridge, Mass.

Reinhart, Tanya (1996) Syntactic Effects of Lexical Operations: Reflexives and Unaccusatives, OTS Working papers in Linguistics , TL-97-002, University of Utrecht, available also at http://www.let.uu.nl/~tanya.reinhart

Reinhart, Tanya (2000), The Theta System: Syntactic realization of verbal concepts, OTS Working papers, 00.0/TL (available also at http://www.let.uu.nl/~tanya.reinhart)

Reinhart, Tanya (2002) 'The Theta System - an overview' Theoretical Linguistics 28(3)

Shen, Yeshayahu, (1985) The Structure of Action in the Short Narrative Text, Unpublished Ph.D Dissertation, Tel Aviv University

Tanya Reinhart [email protected]

The Theta System [email protected]

LSA 2005 Office hours: Monday 12-2,

(room D862, or lounge, 8^th floor)

2. VERB CLASSES

2. Some classes of two-place verbs:

2.1. [+c] subjects (discussed in class 1).

a. With a [-c-m] (theme)second argument:

1 V([+c], [-c-m]) - break, open...

a) The wind /Max /the key opened the door

b) The storm /Max /the stone broke the window.

c) The heat/ Max / the candle melted the ice.

Universally, all verbs with this feature cluster have an unaccusative alternate, as in The window broke. The converse also holds: All one-place unaccusative verbs (with very few exceptions) have an active alternate with a [+c] role in one language or another^⁶.

b. With a [-c+m] (experiencer)second argument:

2 V ([+c], [-c+m] - worry, amuse, scare, surprise... ,

a) Max / the noise / the gun scared Lucie.

b) Fred/ Fred's behavior /the discussion surprised Lucie

Fred/ Fred's gedrag /de discussie verbaasde Lucie.

Verbs with this feature cluster also have a one place alternate ('Subject Experiencer'). In English, examples like (4a) are rare. (Commonly, the English alternate is in a passive form.) In other languages it is very productive, and with few exceptions, as with the Dutch (4b).

3 a) Max worries

b) Max verbaast zich (Max was surprised).

Unlike the alternates of (2), the experiencer alternates of (3) are unergative. This is argued in detail in Reinhart (2000), but for here suffice it to note that verbs occurring with zich in Dutch allow only an unergative derivation (-The subject merges externally).

Exercise: Check whether subject experiencer alternates like (3) are unaccusative or unergative in your language of choice, using the diagnostics in the background appendix.

2.2. Verbs with two [/+c] clusters.

("Manner verbs" (Levin and Rappaport 1995), "verbs of indirect causation" (Grimshaw 1990, Rapp, 199⁷⁷).

a. With a [+c+m] (agent) subject:

L&R's 'manner verbs' like peel, cut, screw, sow, drill, allow either an agent or an instrument subject (4a,b), but not cause (4c). They also do not have a reduced (unaccusative) entry (5).

4 a) Max peeled the apple (with the knife)

b) The knife peeled the apple.

c) *The heat peeled the apple.

5 a) *The apple peeled.

b) *The hole drilled. (Max drilled a hole.)

With agentive verbs, an instrument is always allowed optionally, but not selected as part of the verb's grid. The verbs of the two-[/+c] class select a [+c-m] cluster. These verb concepts are, thus, associated with a pair of necessary and sufficient conditions.

6) drill/peel ([+c+m], [-c-m], [+c-m])

agent, theme, instrument

7) Optionality generalization: When a verb selects two [/+c] roles, only one is obligatorily realized. (-The other, then, is present only in the semantics).

If only the instrument is realized, as in (4b), it must be the external one.

Haiden 2004^⁸ discusses a large set of verbs of this class in German (mostly the same in English (8,9))..

8) a. The army surrounded the village with a wall.

b. A wall surrounds the village.

c. *The occupation surrounds the village with a wall.

9) a. Hans schmückt den Balkon mit Blumen (German)

H decorates the balcony with flowers

b. Blumen schmücken den Balkon

flowers decorate the balcony

b. With a [+c] subject:

10 a) Max/the hose/ the storm filled the pool with water.

b) The water_[+c-m] filled the pool.

11) fill/stain ([+c], [-c-m], [+c-m])

12) Output of [+c] reduction: The pool_[-c-m]i filled t_i with water_[+c-m].

Note: the label instrument does not always capture the actual meaning of the [+c-m] cluster. This is to be expected in view of the contextual determination of clusters meaning.

2.3. [-] verbs (two-place unaccusatives).

a. Levin and Rappaport's (1995) unaccusative verbs selecting a locative as their second argument, like live, appear (and stand, lie, in their 'simple position' use).

b. Piacere/appeal verbs like escape, elude, occur, belong, lack, miss, suffice.

A typical property of these verbs is that they do not have an alternate with a [+c] role, and they always realize only as unaccusative. (The unaccusative properties of the (b) set are discussed in detail in Pesetsky (1995). E.g., (13a) and (14a) seem similar, but the passivization test in (b) shows that (14a) is unaccusative, i.e. the subject got there by movement. )

13 a. The solution surprised Max

b. Max was surprised by the solution

14 a. The solution escaped Max.

b. *Max was escaped by the solution.

15 a. (first merge:) escaped the solution Max

b) (move:) The solution_i escaped t_i Max.

16) The solution_i appealed t_i to Max.

(Max in (15) is not Accusative, but Dative, as in (16). In English this distinction is not marked morphologically, but it is obvious from e.g. German.)

17) escape/ appeal ([-c-m] [-c])

Recall that [-c] is interpretable as either [-c-m] or [-c+m]. Marelj's FITR will determine that only the second is allowed here. So the [-c] argument ends up interpreted as experiencer.

18 a) The solution_i escaped t_i Max.

b) Max_i escaped t_i the police.

c) *The problem escaped a solution.

19 a) The solution_[-c-m]i escaped t_i Max_[-c].

b) Max_[-c]i escaped the police_[-c-m] t_i.

The Theta System

Background material

Background material: UNACCUSATIVE DIAGNOSTICS

Unergative derivation: Lucie walked (Lucie originates as subject.)

Unaccusative derivation: Lucie_ifell t_i

(Lucie originates in object position and moves.)

Nominalization (applicable in many languages e.g. English, Dutch, etc.):

(1) a. She walks fast - She is a fast walker

b. She fell gracefully -*She is an experienced faller

c. Hij is gevallen – *Hij is een valler

Auxiliary Selection (Dutch &Italian)

(2) a. Jan is gevallen
Jan is fallen (Jan fell)

b. Jan heeft gesprongen
Jan has jumped (Jan jumped)

Impersonal Passives (Dutch)

(3) a. ...er werd gesprongen
…there was jumped

(4) b. ...*er werd gevallen
…there was fallen

Prenominal past participials (Dutch)

(5) a. de dinerende studenten/de vallende studenten
the dining students/de falling students (present participle)

b. de gevallen student
the fallen student (past participle)

c. *de gelopen student
the walked student (past participle)

Word Order - Simple Inversion (Hebrew)

(6) a. *rakdu šloša yeladim ba-mesiba
danced three boys in+the party

b. hit’alfu šloša xayalim ba-hafgana
fainted three soliders in+the demonstration

Possessive Dative (Hebrew)

(7) a. šney sfarim naflu le-dan.
two books fell to Dan
‘Dan’s books fell’

b. *ha-kelev šaxav le-dina
the dog lay to-Dina
‘Dina’s dog lay’

ne -Cliticization (Italian)

(8) a. Sono arrivati molti studenti.
are arrived many students (Many students arrived)

b. Ne sono arrivati molti
of- them are arrived many

(9) a. Hanno corso molti studenti
have run many students (Many students ran)

b. *Ne hanno corso molti
*of-them have run many

Genitive of Negation i.e. Slavic Genitive (Russian)

(10) a. Ne objavilos’ studentov (unaccusative)
NEG showed up students-Gen

b. *Ne tancevalo studentov (unergative)
NEG danced students-Gen

1 The idea has been around, of course, e.g. Jackendoff (1987), though the actual feature system I am using here is rather different than the earlier inspiring proposals.

2 Aspectual properties of verbs are governed by an independent system, which interacts with the θ-feature clusters. Grimshaw (1990), who, like Jackendoff, attempts to capture aspect within the system of thematic roles, assumes a different division of the two role-systems than his. She takes the thematic roles to include agent, experiencer, goal, source, location, and theme, while cause is her major aspectual role. I share her belief that the cause role (or, more precisely, the feature /+c in my system below) plays a crucial role in determining the eventive (telic) nature of predicates, but, I do not think that capturing this requires postulating two distinct θ-roles systems. Rather, aspectual computation should be able to read the θ-features, and combine this information with other semantic properties of verbs and of their complements, which are independent of the thematic properties.

3 The cluster that corresponds to cause semantically is [+c+m]. However, I did not find any verb that selects an argument that is obligatorily only a cause (i.e. cannot be realized also by an instrument or an agent). Hence, it appears that the role cause actually corresponds in natural language to the cluster [+c]. The construal [+c-m], then, is obtained for cause arguments only via the cluster [+c].

4 This cluster has not been identified as an independent θ-role before. In terms of thematic content, candidates for bearing this cluster are arguments of verbs like laugh, cry, and sleep, which require an animate argument, but do not involve necessarily agency or a causal relation with this argument. In the Theta system it is assumed that this is also the θ –role of subjects of psychological verbs like love, know, believe, see, hear, which have been viewed as instances of the experiencer role before. Unlike the standard experiencer, which has varying realizations, [+m] arguments realize always externally (which follows, in the full system from their feature composition). It is not clear that this cluster, when it occurs with psychological verbs, is semantically distinguishable from the standard experiencer cluster [-c+m]. But assuming that the subject of these verbs is [+m] enables maintaining the very simple Accusative rule that we turn to.

5 To be logically defined, the system should contain also this [ ] cluster (unspecified for both /c and /m. Marelj (2004) discovered that it exists indeed, and it is operative in middle-formation. She distinguishes languages where this operation applies in the lexicon and in the syntax. In Lexicon languages (including English), the operation involves a change in the feature composition of one of the clusters. She shows that the resulting [ ] cluster corresponds in its interpretation to the ARB(itrary) variable that Chierchia (1995) identified in impersonal constructions.

6 Note that the claim is not only that one-place unaccusative entries have a transitive alternate, but that this alternate is a [+c] verb. Again, exceptions are hard to find, but one I know is the verb grow. In many languages, the transitive alternate of the unaccusative grow takes agent but no instrument or cause (which suggests that it selects a [+c+m] argument). Possibly the same holds for cook. The verb drown varies across languages on whether it allows cause, or just an agent, in its transitive entry.

7 Rapp, Irene. 1997. Partizipien Und Semantische Struktur: Zu Passivischen Konstruktionen Mit Dem 3. Status.vol. 54: Studien Zur Deuschen Grammatik. Tübingen: Stauffenberg.

8 Haiden, Martin. 2004. Theta Theory, PhD dissertation, Tilburg university.

BASICS OF INTERNATIONAL DISABILITY LAW LECTURER DR ISTVÁN HOFFMAN
BASICS OF LEVELLING USES OF LEVELLING IN THE CONTEXT
BASICS OF MATERIAL TECHNOLOGY (BAEAAE1BNE) – INDIVIDUAL TASKS

Tags: basics of, 4. basics, theta, system, basics