APPENDIX I ABBREVIATIONS AND TRANSLITERATION 1. LIST OF ABBREVIATIONS FOR SENTENCE GLOSSES The following abbreviations have been used for the sample sentence glosses: [NOM [ [ACC[ [GENI [GENII [GEN21 [DATI [LOCJ [LOq [LOCzl [INS I [SGI [PLI [FEMI [MASCJ [NEUTI [ISTPERS·I [3RDPERS·1 [PASTI [INFI [REFLI [SHORT FORM AD]I [LONG FORM AD]I Nominative Accusative Genitive Genitive, Genitive 2 Dative Locative Locative, Locative 2 Instrumental Smgular Plural Feminine Masculine Neuter 1st Person 3rd Person Past tense Infinitive Reflexive Short form adjective Long form adjective 173
Transcript
APPENDIX I
ABBREVIATIONS AND TRANSLITERATION
1. LIST OF ABBREVIATIONS FOR SENTENCE GLOSSES
The following abbreviations have been used for the sample sentence glosses:
Constituent structure (c-structure) and lexical properties are determined and represented independently. The information from these two components is integrated within functional structure (f-structure). Within the LFG model, three levels of representation are simultaneous. There is no syntactic derivation involved.
Productive relationships in language of the kind that inspired transformational models (for example Chomsky's, 1957, Syntactic Structures) are accounted for, instead, by relationships that hold between lexical items. For example, it is assumed that corresponding to the active verb hit, there is a passive verb to be hit by. The relationship between the arguments of active and passive verbs is predictable; they are related by a productive lexical process. Such relationships are expressed as redundancy rules, and they relate actual lexical items.
Functional structure brings together information contributed from the syntax (constituent structure) and from the lexicon. An example will serve to illustrate the interaction of the three components. Consider the constituent structure representation for the following sentence:
(1) Natasha hit Boris.
It may be represented by the following simplified tree:
(2) S
~ NP VP
I I SUB head
N I
phrasal head
Natasha
ofS
~ V NP I I
phrasal OBJ head hit Boris
177
178 APPENDIX III
The annotations are associated with constituents in the Phrase Structure rules.
2. PHRASE STRUCTURE RULES
The labelling of grammatical arguments in c-structure is accomplished by the PS rules themselves; that is, PS rules are annotated with indications of the arguments that correspond to each constituent. PS expansions make use of the X-bar system suggested in Chomsky (1970) and elaborated and refined by Jackendoff (1977) and Bresnan (1977). I assume that the phrases NP, VP, AP, and PP may be decomposed into syntactic distinctive features [± VI, [± NI. I further assume that Quantifier Phrases, or QPs, are distinguished by a third feature. The feature decomposition allows for generalizations about the similar behavior of different categories to be captured. For example, only categories which are [+NI may bear case (that is, NP and AP). Furtherrflore, the X-bar theory assumes the existence of super categories ... , where the head of the phrase X" is of the form xn - I. (The reader is referred to Jackendoff for the other conventions and assumptions of this theory, and the motivation behind them.) In the annotated PS rules of LFG, nodes are identified by means of arrows: t and L The t refers to the immediately dominating node, and ~ to the node itself. So, the PS rules look like:
(3) S ---+ NP (t SUB) = ~
VP t = ~
This is interpreted to mean that the subject of S is NP. The second equation, under the VP, t = ~, identifies the VP as the head of S, and will result in the sharing of all information between the Sand VP nodes. Similarly, in the equation
(4) VP -+ V NP t = ~ (t OBJ) = ~
the V is identified as the head of VP, all information is shared between the two nodes; while the NP is marked as the object of the VP. Likewise, in the following expansion
(5) NP -+ N t = ~
the N is marked as the head of NP. The notation t = ~ is the unmarked assignment to the head of a phrase.
So, the tree generated by these two equations for the sentence given in (1) is in fact (6). The nodes are numbered to make it easier to refer to them in the discussion to come.
LEXICAL FUNCTIONAL GRAMMAR 179
(6)
Np2
I (l SUB) = i
I t = i
~ y5 NP6
t = i
I I I
t = i (t OBJ) = i
I I Natasha hit Boris
This information may be transferred into the format of the functional structure representation directly.l F -structure representation is derived from c-structure, by use of the information provided by the PS annotations. The features of each lexical item are transmitted to the node which immediately dominates the word in c-structure, and thereby are also contributed to f-structure. What follows is a step-by-step illustration of how information from c-structure and from the lexicon is incorporated into f-structure. Since nodes 1 and 3 are set equal by the equation beneath the YP node, they are also set equal in f-structure.
(7) 1 ~ 3[
The subject of 1(= 3) is 2(= 4).
(8) 1I The information about the lexical item under node 4 - Natasha - comes from the lexical entry for that word, and will be included within the node labeled 2 = 4 in (8). This will include information about the meaning of the word (however that is to be represented), including the fact that Natasha is a Noun that is feminine, 3rd person singular, and that it expresses a female first name. The PRED (short for Predicate) value is the approximation to meaning (here represented as 'Natasha'). The other grammaticali features of the word are contributed by the equations beneath the PRED specification. The lexical entry will then be:
(9) N atasha, N (t PRE D) = 'Natasha' (tNUM) =-PL (t GEND) = + FEM
The meaning of some lexical items includes information about how arguments of that word are interpreted. For example, the verb hit in
180 APPENDIX III
English requires a subject, which is interpreted as the agent of the action, and an object, which is associated with the patient role. (The lexical entry is illustrated in the next section.) Grammatical functions - such as SUB and OBJ - are the interface between constituency and predicate argument structure, since these functions are mentioned both in PS expansions and in the lexical entries. The PS annotations specifying grammatical functions permit the construction of functional structure, where structural and lexical information is merged.
So, returning the example under discussion, the information from the lexical entry for Natasha is incorporated into the functional structure, as shown:
(10) 1 ~3lSUB 2~4lPRED NUM GEND PERS
'Natasha'] ] -PL +FEM 3RD
in similar fashion, the functional structure is compleed by incorporating the information from c-structure and from the lexicon, as shown in Table 1.
\=J=5 SUB
PRED TENSE
081
TABLE I Functional structure of (1 )
2=4
[
PRED NUM GEND PERS
'hit (SUB, OB1)' PAST
(, [PRED NUM GEND PERS
3, LEXICAL ENTRIES
'Natasha' -PL +FEM 3RD
'Boris' -PL -FEM 3RD
]
]
For some lexical items, the value of the predicate (the PRED feature) -that is, the meaning of the expression - is a function of the values of its arguments. For example, the verb kill takes as arguments a killer and a killee,2 Such an item is called a lexical form and is listed with its predicate argument structure, a list of logical arguments and indication of corresponding grammatical functions.'
(11 ) kill (SUB, AGENT
OBJ) THEME
LEXICAL FUNCTIONAL GRAMMAR 181
All lexical entries are referred to as semantic forms. Lexical entries that have a predicate argument structure are referred to as lexical forms. As already mentioned, kill is a lexical form, since inherent to its meaning is the existence of a killer and a killee. This is not so of tree, for example, which is a self-contained semantic form.
Possible arguments include agent, theme,4 etc., each of which holds a fixed position in predicate argument structure (argument 1, argument 2, etc.). In the lexical form, each argument is associated with a grammatical function [SUB (subject), OBJ (object), etc.l, as illustrated in (11). Thus, the lexical entry contains a pairing of logical arguments and grammatical functions. The principle of Function-Argument Biuniqueness requires, basically, that this pairing be one-to-one. Each argument must be assigned a unique grammatical function (even if the assignment is O),5 and each grammatical function that is included in the pairing must be associated with a unique argument.6
Lexical entries for Russian nouns also include information about case. In the theory presented here, the terminal elements of the constituent structure are fully formed words. Inflection is not accomplished by syntactic derivation; rather, all inflected forms are produced by lexical rules.7 Morphological regularity is captured by lexical redundancy rules.
4. LEXICAL REDUNDANCY RULES
As opposed to transformational grammars (such as those of Chomsky, 1965 or 1981) that have set up different levels of constituent structure, each derived from the previous level by application of movement rules, LFG postulates a single level of constituent structure, and excludes the possibility of syntactic derivation or syntactic movement (except for restricted scrambling rules). This constituent structure is the input to the phonological component.
Syntactic relationships that have, in the traditional transformational grammars, been attributed to syntactic movement, are, instead, accounted for in terms of lexical relatedness. As mentioned earlier, corresponding to the active verb:
(12) 'hit' (SUB, AGENT
OBI), THEME
there is a passive verb:
'to be hit' (0, AGENT
SUB). THEME
The relation between the active and passive forms is expressed by the fact that the OBJ of the active is associated with the same logical argument (the second argument) as the SUB of the passive form. This relation is captured by the following lexical redundancy rule: 8
182 APPENDIX III
(13) The Passive in English Functional change: (SUB) -- 0/ (OBLIQUE AGENT)
(OBl) -- (SUB)
Morphological change:
Redundancy rules are not derivational rules. They are intended to relate actually existing lexical items; thus, they are useful in organizing the information stored in the lexicon, but they are not normally referred to during on-line processingY Online processing makes reference only to actually occurring lexical items.
This requires that lexical rules not be allowed to be formulated so as to accept as input abstract forms - non-existing lexical entries from which real entries are to be derived. This makes some strong predictions. Constraints that must hold of derived lexical items appear also to hold of their precursor forms. I 0
Returning now to the principle of Function-Argument Biuniqueness, it becomes apparent that certain types of lexical redundancy rules are excluded in principle: for example, a hypothetical passivization rule that transformed OBl into SUB without deleting or changing the function of the original SUB. Such a rule would produce an inadmissible pairing of arguments to functions
(14) a. No-such-verb
(t PRED) = 'SSSSS (SUB, OBl)' arg.l arg.2
b. No-such-passivized-verb
(t PRED) = 'SSSSS (SUB, SUB)' argo 1 arg.2
since the same grammatical function (SUB) is associated with two different arguments.
It is within functional structure that lexical and constituent information is pooled. It is there, for example, that the arguments of a predicate may be interpreted. This is because the predicate argument structure involves a mapping from arguments to grammatical functions, and the grammatical functions are assigned values in functional structure (where the arguments of each lexical form may be evaluated). More specifically, the arguments of a predicate must be contained within the same clause nucleus (that is, the minimal f-structure containing a PRED whose value includes a lexical form - a pairing of arguments and grammatical functions. I I For illustration, consider the following,
(15)
LEXICAL FUNCTIONAL GRAMMAR
S
~ NP I
(1 SUB) = ~
I Natasha
VP I
1 = ~
I speaks
183
Since VP is the head of S, it receives the unmarked assignment: 1 = L Thus, the information from the VP node and from the S node will merge in functional structure. The equation beneath the NP node will be interpreted as 'mother's (that is, S's [and, therefore, also VP's]) SUB = daughter (that is, the lexical item 'Natasha')'. In other words: 'the subject of speaks is Natasha'. In the lexicon Natasha is given a meaning, or a predicate, as is the verb to speak. Speak is also assigned a predicate argument structure: it takes an agentive subject.
(16) speaks, Verb (1 PRED) = 'to speak (SUB)'
AGENT
Moreover, the word !Jpeaks contains additional information about agreement. For example, it will only occur with a singular subject. In English, this is expressed within the lexical entry by a constraint equation which checks information in functional structure, but does not itself contribute function-values. Constraint equations, written with the symbol [= cJ provide a filtering mechanism, since the functional structure will be well-formed only if the constraint equation is satisfied.
(17) speaks, Verb (i PRED) = 'to speak (SUB)'
AGENT
(i NUM) =c -PL
In this case, the subject Natasha is a singular NP, and so the sentence is fine.
The last equation in (18) contributes the value [-PLJ to the function NUM. Such an equation is referred to as a constituting equation rather than a constraint equation.
As was already mentioned, the association of constituents to the arguments of lexical items is accomplished at the level of functional
184 APPENDIX III
structure. According to the algorithm in Kaplan and Bresnan (1982), the following functional structure representation would be assigned to the above sentence:
(19)
[
SUB [PRED GEND NUM
PRED 'speak (SUB)'
'Natasha' ] ] -MASC -PL
The constraint equation given in (17) is satisfied, since the subject Natasha is marked as singular in the lexicon (as shown in (18)).
5. FUNCTIONAL WELL-FORMEDNESS
Certain common-sense conditions on the well-formedness of functional structures must hold for the sentence to be acceptable:
I. COHERENCE
Coherence is a well-formedness condition that requires that every semantic form contained within the f-structure be the PRED value of a grammatical function mentioned (either in the predicate argument structure or a constituting equation). (See definition in Kaplan and Bresnan, 1982). Thus, the sentence
(20) Natasha speaks green.
is ill-formed (in any but a very poetic style, where green is analogous to the truth in the sentence 'Natasha speaks the truth'). This is because green does not represent any argument of the predicate speaks, nor is it interpretable as an adjunct.
II. CONSISTENCY
The principle of consistency (also called functional uniqueness), requires that, in an f-structure, a particular function name (such as NUM) must have a unique value. l2 A representation that included:
(21 ) NUM NUM
+PL -PL
would be contradictory and inconsistent.
III. COMPLETENESS
An f-structure is complete if and only if it contains values for the
LEXICAL FUNCTIONAL GRAMMAR 185
grammatical functons that are subcategorized by the Predicate.13 Thus, the sentence
(22) *Speaks.
is incomplete, since there is no value provided for the subject argument of speak.
In Russian, however, such sentences are grammatical. This difference between Russian and English may be attributed to the difference between constraint equations, and constituting equations - which contribute information. In Russian, then, agreement with subject would be expressed by constituting equations in the lexical entry, such as for govorit 'speaks':
(t SUB PERS) = 3RD
Thus, this information would be included in the f-structure for the sentence: Govorit.
(23)
rSUB [ PER
PRED 'speak (SUB)'
3rd
In English, however, a constraint equation such as the following for says
(t SUB NUM) =c -PL
could not be satisfied if the subject is not lexically filled. Notice that the Russian constituting equations are equally effective in
ensuring agreement of features. If gender and number information is provided from two sources - the lexical item itself, and from verbal morphology, for example - consistency assures that a singular noun will not be used with a plural verb form. However, if the subject is not lexically filled, then, in Russian, the constituting equation included with the verb will provide information about the grammatical features of that subject, and thus makes the sentence interpretable.
IV. SEMANTIC COHERENCE
Semantic forms with semantic content (that is, excluding dummy elements such as English it and there and idiom chunks that do not have inherent meaning) would have to be linked to a logical argument of another lexical form within their minimal clause nucleus in order for them to be coherently interpreted.
As has been seen in preceding chapters, these simple principles are quite powerful in constraining the grammar.
186 APPENDIX III
6. POSSIBLE RULES
There are two additional principles which constrain LFG. They relate to possible syntactic rules.
I. THE LEXICAL INTEGRITY HYPOTHESIS
The Lexical Integrity Hypothesis, which originated with Chomsky (1970) and was taken up later by Bresnan, prohibits syntactic rules from moving any element into or out of lexical categories such as N, A, or V. (See also Mohanan, 1981, and Simpson, 1983, for discussion.) This type of constraint eliminates, for example, the possibility of a rule like AffixHopping (Chomsky, 1957), since words necessarily emerge from the lexicon with all their affixes in place. Although it is no longer clear what the status of rules like affix-hopping is in recent Chomskyan analyses, such rules expressed real generalizations. In LFG, these generalizations are captured by means of productive lexical processes. For example, the rule that forms plurals in English attaches the plural affix along with the feature [t NUM = +PLj. Likewise, other information conveyed by verbal suffixes is incorporated in the derived forms, thus allowing for apparently discontinuous syntactic dependencies between, for example, the auxiliary have and the suffix -en. The suffix contributes the relevant information to the lexical category as a whole (for example been), which is adjacent to have (as in 'have been'). Thus, Affix-Hopping, like Passive, provides an example of a relationship which had been used to motivate transformations, but which finds a natural treatment in terms of LFG.1 4
II. THE PRINCIPLE OF DIRECT SYNTACTIC ENCODING
This principle (see Bresnan, 1982b) requires that "every non-lexical rule of grammar preserve function-assignments." This excludes the possibility of syntactic movement (or other) rules that would alter grammatical functions.
7. THEORY OF CONTROL AND COMPLEMENTATION
Several of the grammatical functions that are postulated to belong to the universal set of grammatical functions have been discussed. In addition to functions like SUB and OBJ, this set also includes predicate complements and adjuncts.
7.1. Complements versus Adjuncts
These functions are illustrated in the following sentences from Bresnan (1979):
LEXICAL FUNCTIONAL GRAMMAR 187
COMPLEMENTS
(24) a. John didn't sound ashamed of himself
b. Fred struck me as a fool.
c. Jogging keeps Susan in a bad mood.
d. I'll have your brother working again.
j\DJUNCTS
(25) a. John looked down, ashamed of himself
b. Louise enjoyed sports as a girl.
c. Susan ate her lunch in a bad mood.
d. I found the money walking our dog.
See Bresnan (1979) for additional examples illustrating the difference between complements and adjuncts. The major difference is that:
COMPLEMENTS are grammatical arguments, and are required for functional completeness.
ADJUNCTS provide additional information, but are not subcategorized for by particular lexical items.
If a verb requires a complement, then the complement must be present for the sentence to be complete. Adjuncts are never required for grammaticality. Removing the italicized phrase from sentences (24) either makes the sentences ungrammatical or changes the primary meaning of the main predicate. Removing the italicized phrases from (25) does not affect grammaticality, nor does it alter the meaning of what remains. Compare the following (where complements are indicated by the term XCOMP):
(26) put, Verb (t PRED) = 'put (SUB, OBJ, XCOMP)'
(27) *Natasha put the book.
(28) Natasha put the book in the living room.
(29) read, Verb (t PRED) = 'read (SUB, OBJ)'
(30) Boris read the book.
(31) Boris read the book in the living room.
There are other differences as well. Adjuncts have greater mobility, in that they can be found in a variety of positions, while complements occur in a single fixed position. Adjuncts may also be set off by pauses, unlike complements. Compare: 15
188 APPENDIX III
ADJUNCTS
(32) a. In a bad mood, Susan ate her lunch.
b. Susan ate her lunch, in a bad mood.
COMPLEMENTS
(33) a. *In a bad mood, jogging keeps Susan.
b. *Jogging keeps Susan, in a bad mood.
Some predicates optionally include a complement in the predicate argument structure, giving rise to functional ambiguity. Consider the following sentence:
(34) Natasha keeps a cat around the house.
The ambiguity arises because keep can be used in two different senses, each associated with a different argument structure.
keep" V (t PRED) = 'keep (SUB, OBJ)'
keep2, V (t PRED) = 'keep (SUB, 081, XCOMP)'
On one reading of (34), in the house is indispensable in defining where the cat was kept. On the other reading, it merely indentifies the place in which the action occurs. The ambiguity becomes immediately apparent when the word order is changed in such a way as to permit only the adjunct reading.
(35) Around the house, Natasha keeps a cat.
The contrast is even more striking when the OBJ is not something that can be kept in the sense of keep,.
(36) Natasha keeps Boris around the house.
Here, only the complement reading is possible, and the word order may not be changed: 1(,
(37) * Around the house, Natasha keeps Boris.
Adjuncts and complements are also distinguished by extraction. Compare the following two examples (again from Bresnan, 1979):
(38) a. COMPLEMENT:
Who did you say John sounded ashamed of?
b. ADJUNCT:
*Who did you say John looked down, ashamed of?
LEXICAL FUNCTIONAL GRAMMAR 189
Unlike adjuncts, complements may be questioned since they represent an argument of the main predicate.
7.2. Open Complements
Since predicate complements are arguments of the verbs with which they occur, they are listed in predicate argument structure. The symbol XCOMP is used to designate the set of ACOMPs (adjectival complements), NCOMPs (nominal complements), VCOMPs (verbal complements) and PCOMPs (prepositional complements).l7 This is one area where the distinction between grammatical functions and constituency is extremely useful. NCOMPs are distinguished from OBJs, although they share the same constituency. The same is true about PCOMPs and POBJs (prepositionalobjects).
Complements predicate something of another element in the sentence, either of the subject or the object. l8 This relation of predication is captured by the assignment to the complement of a subject that is identical with some other argument. Notice that different verbs place different requirements on the interpretation of the XCOMPs subject:
(39) a. Boris; struck Natasha [[ L as a fool].
b. Boris regarded Natashaj [[ ]i as a fool].
This relation is expressed by a control equation which is included in the lexical entry of each lexical form that takes an XCOMP. So, for example, a verb of sulifect control, such as strike as, includes the following control equation:
(t SUB) = (t XCOMP SUB)
which expresses the relation indicated by the coindexing in (39a) in that the f-structure subject of the XCOMP is identical in every respect with the SUB of the sentence. This may be represented in the f-structure either by coindexing the identical f-structures, or by drawing an arrow from the controllee to the controller.19
Logically enough, object control is expressed by a control equation of the form: (t OBJ) = (t XCOMP SUB). Since object control is the unmarked relation if an object is present, the control equations may be filled in redundantly: object control if there is an object present, unless a subject control equation is specifically required by the lexical item; and subject control otherwise.
The f-structure representation for the following sentence will illustrate the way f-structure is constructed:
190 APPENDIX III
(40) Natasha keeps Boris busy.
S
~ NP VP
I ~ N V NP AP
I I I I Natasha keeps N A
LFG Boris busy.
TABLE II F-Structure representation of (40)
SUB [ PRED 'Natasha' ] NUM -PL GEND +FEM PERS 3RD
PRED 'keep (SUB, OBJ, XCOMP)' TENSE -PAST
OBJ [ PRED
'Boris'
J NUM -PL GEND -FEM PERS 3RD
XCOMP [ SUB I IJ PRED 'busy (SUB)'
The predication relation that holds between complements and their subjects is represented functionally. The subject and complement form a clause nucleus, which is a minimal f-structure containing a lexical form (that is, a PRED value including a list of arguments). This makes the correct predictions for English reflexive pronouns, which require an antecedent within their minimal clause nucleus. In Bresnan (1982a), it is argued that f-structure is the relevant level for the determination of anaphoric relations. Where c-structure and f-structure are significantly different, it is the f-structure representation that determines interpretation of anaphors. Thus, for example, sentences which 'share the same constituent structure, but differ in that one contains a prepositional comple-
LEXICAL FUNCTIONAL GRAMMAR 191
ment while the other contains a prepositional object, differ with respect to reflexivization. The clause containing a complement contains an additional clause nucleus. Consider the following two examples from Bresnan (1979).
(41) a. Susan informed John about the house.
b. Susan kept John about the house.
Although they are identical in constituency, their functional structures differ, as is represented schematically in (42) (where clause nuclei are indicated by eN):
(42) a. [Susan informed John about the house]CN
b. [Susan kept John; [[ ]; about the house] CN ] CN
Therefore, since reflexivization is possible in English within the minimal clause nucleus, we find the predicted contrast:
(43) a. Susan informed John about herself. [Susan; informed John about herself;]cN
b. *Susan kept John about herself. *Susan kept John; II L about herselflcN
This provides a striking illustration of the importance of separating c-structure and f-structure representation.
The distinction between constituent structure and functional structure is important also in situations where the syntactic configuration and argument structure do not coincide. Such is the case of so-called raising constructions. Transformational accounts are required to alter constituency so that subcategorization could be represented by the constituency on one level, while the surface constituency would be represented on another. In LFG, there is a more straightforward account. Consider consider.
(44) Natasha considers Boris boring.
Natasha is the object of consider (as argued in Bresnan, 1982b). In case marking languages such as Russian, Natasha receives the case marking that is normal for objects. However, it does not represent a logical argument of the verb consider (and is not part of its predicate argument structure): consider does not exert any subcategorization requirements over Natasha. 20 Rather, it represents the logical subject argument of to be (although this relation is not apparent in c-structure). Such cases are represented like the previous one: as a case of grammatical control.
(45) consider, V (t PRED) = 'consider (SUB, XCOMP) (OBJ)'
The placement of the object indicates that it is not associated with a
192 APPENDIX III
thematic argument. However, by the redundancy rule discussed (at the beginning of this subsection), the lexical entry will be provided with a control equation:
(t OBJ) = (t XCOMP SUB)
The functional structure representation would be as follows:
TABLE III F-Structure representation of (43)
SUB
PRED TENSE
OB1
[ ~~~ PERS
'consider (SUB, XCOMP) (OB1)' -PAST
[ ~~~ PERS
XCOMP [SUB PRED
'Natasha' -PL +FEM 3RD
'Boris' -PL -FEM 3RD
I 'boring (SUB)'
]
J Ii]
The two subjects are absolutely identical in functional structure, as indicated by the coindexing. They share all features. This type of control relation is referred to as grammatical or functional control.
XCOMPs are called open complements because the subject argument is open, and must be provided by a relation of grammatical control. In the above sentence, the open complement is adjectival, but it might also be verbal, nominal, or prepositional.21
(46) a. Boris considered Natasha to be unusual.
b. Boris considered Natasha an unusual person.
c. Everyone considered Kennedy out of the running.
The fact that consider takes a non-thematic argument would lead us to expect that idiom chunks and the dummy there would be possible in object position. As Bresnan (1982a and 1982b) points out, this is correct:
(47) Consider your goose cooked.
As Bresnan observes, this can have the idiomatic interpretation. The same is true for other control verbs involving extragrammatical arguments:
(48) Close tabs seem to have been kept on Boris and Natasha.
LEXICAL FUNCTIONAL GRAMMAR 193
One important property of this representation is that control relations are preserved by lexical rules. Lexical rules apply not only to predicate argument structure, but to all equations within the lexical entry. A rule that changes OBJ changes every occurrence of OBJ in the lexical representation in the same way. Consider again consider, which has the lexical form
(49) (t PRED) = 'consider (SUB, XCOMPj (OBJ)'
to which the control equation is added redundantly:
(t OBJ) = (t XCOMP SUB)
Now, the passive of consider with have the lexical form
(50) (1 PRED) = 'consider (0, XCOMPj (SUB),
to which will be added the equation:
(t SUB) = (t XCOMP SUB)
The same argument is controlling the XCOMP in both the active and passive forms.
(51) a. They consider Natasha unusual.
b. Natasha is considered unusual.
In LFG, there is a sharp distinction between constituency and lexical subcategorization. Many transformational approaches are founded on the principle that they are essentially the same.22 Sentences like (51) provide examples of where the match-up is less than perfect. In a transformational framework, constituency must be adjusted in the course of the derivation to account for these mismatches. However, LFG does not allow such derivation, and captures within functional structure the relation (of identity) that holds between the constituent-structure object of the main clause of (51a), for example, and the understood subject of the complement phrase.
Another example of the distinction between constituency and grammatical functions is provided by passivization. While in c-structure, both OBJ and NCOMP may have the same constituency, only an OBJ may passivize.23
(52) a. Boris insulted a linguist.
b. A linguist was insulted by Boris.
(53) a. Boris became a linguist.
b. * A linguist was become by Boris.
One crucial element of this model is the independence of lexical and
194 APPENDIX IIJ
constituent representation.24 This makes it possible to determine and represent generalizations about constituent structure and subcategorization independently, without encoding subcategorization in constituent structure.25
To summarize, then, predicate complements have an open (subject) argument that is grammatically controlled by the main predicate's subject or object.
7.3. Open Adjuncts
The consider example involved lexically induced functional control. The XCOMP was subcategorized for by the lexical item. Adjuncts occur with a much freer distribution. They may have an open subject argument as well, but this argument is not lexically controlled. Consider the following example:
(54) Exhausted after a long day at the office, Natasha took a nap.
Here, the phrase 'exhausted after a long day at the office' may appear with any verb:
(55) Exhausted after a long day at the office, Natasha VERBed.
Its occurrence is in no way conditioned by the choice of a particular lexical item, unlike the appearance of the XCOMP of the verb consider. Such adjunct phrases will be assigned the function XADJ analogous to XCOMP, where the X indicates an open argument. Consider the following sentence:
(56) Insulted, Natasha stormed out of the room.
To permit the subject of the XADJ to be interpreted, the assignment in PS expansions of the function adjunct:
(i XADJ) = t
will be accompanied by the equation
U SUB) = i G
where G is a grammatical function (SUB, OBJ, etc.). The set of acceptable adjunct controllers may vary across languages.
So, open adjuncts are like predicate complements in that they also involve grammatical control of their subjects. They differ from complements in that the control relation is configurationally induced for adjuncts, but lexically induced for complements. That is, complements are subcategorized for by lexical items, while adjuncts are not. As shown in chapter 5, this distinction is also reflected in case marking differences. (In
LEXICAL FUNCTIONAL GRAMMAR 195
Russian, adjuncts agree in case with their antecedents, while complements normally bear Instrumental case.)
7.4. Closed Complements
Many of the verbs that subcategorize for open complements may also take closed complements, in which the argument structure is self-contained. Such complements are designated by the term COMP, and are distinguished from the open complement XCOMP.26 For example, the verb consider may also take a COMP:
(57) a. Natasha considered [that the caviar was substandard].
b. Natasha considered [that Boris was a nogood-nik].
In terms of constituency, 'that the caviar was substandard' is an S'. However, in terms of grammatical functions, it represents a complement of consider. (No control relation is required, though, since it is a closed complement.)
To summarize, then, open complements require grammatical controllers; closed complements do not. Not all closed complements, however, fully specify the reference of their arguments. Consider the following example:
(58) Reading mystery stories is fun.
Although there is no overt subject of reading present in c-structure, there is a subject in functional structure, which can serve as a reflexive antecedent.
(59) Reading yourself mystery stories is fun.
However, the subject of reading is not grammatically controlled: it couldn't be, since there is no possible controller! It will be assumed that a subject is provided in functional structure,27 this subject being a functional anaphor, having as Predicate: PRO. The reference of the PRO may be determined by other sentential elements. In this case, there is said to be anaphoric control. In this example from Bresnan, the PRO's interpretation is clear:
(60) I had to speak to John recently about scratching himself in public.
The interpretation of PRO obeys many of the same properties as the interpretation of overt pronouns. The preceding sentence will be interpreted in much the same way as the following one.
(61) I had to speak to John recently about his scratching himself in public.
196 APPENDIX III
Bresnan (1979 and 1982a) discusses some of the properties of anaphoric control, and its relation to the interpretation of overt pronouns. One similarity is that both PRO and real pronouns may have split antecedents. Compare:
(62) a. Tom told Mary that they should not scratch each other.
[they = Tom and Mary]
b. Tom spoke to Mary about scratching each other in public.
[about Tom and Mary scratching each other]
This would be impossible with grammatical control, where the controller is uniquely determined by the control equation. Thus, we have seen that there are two distinct types of control, or referential dependency. Anaphoric control involves the identity of reference of PRO and some other NP in the sentence.n Grammatical control, in contrast, involves identity of f-structure of the subject of an open function (either an adjunct or complement) and another grammatical function. Grammatical control may be induced lexically, by lexical forms which contain a control equation, or configurationally, in the case of adjuncts, by a control equation introduced in PS.
7.5. Closed Adjuncts
Just as there were open and closed complements (XCOMPs and COMPs), so there are open and closed adjuncts (XADJs and ADJs). Adjunct phrases need not modify a particular phrase; they may provide additional information about the event as a whole, and in such cases, their subjects are not grammatically controlled. Consider the following example:
(63) Natasha kissed Boris in the garden.
In the garden does not have a functionally controlled subject. There are general principles governing the distribution and interpreta
tion of these open and closed functions. The reader is referred to Bresnan (1982a) for details. In summary, though, control is a relation of referential dependency. Grammatical control involves identity of f-structure, while anaphoric control involves only identity of reference of an anaphor and some other element of the sentence. In the case of anaphoric control, no antecedent is required for well-formedness. However, grammatical control requires a controller, since the controlled subject (of the open complement or adjunct) has no independent identity.
7.6. The Constituency of Complements
The control equation for XCOMPs is provided only in f-structure,
LEXICAL FUNCTIONAL GRAMMAR 197
through the control equation from the lexicon. At the level of c-structure, there is no subject of the XCOMP present at all. This representation finds support from phonological evidence. Contraction, for example, distinguishes between those elements that are transformationally analyzed as "WH-traces", and those subjects which in LFG are analyzed as cases of grammatical control. As discussed in Postal and Pullum (1978 and 1979) (as well as in Chomsky and Lasnik, 1978, Jaeggli, 1980, and elsewhere), there is a contrast between the following examples: 29
(64) a. Where do you want [PRO to go)?
b. Where do you wanna go?
(65) a. Who do you want [t WH to visit us?]
b. *Who do you wanna visit us?
In LFG, the null subject of (65) is structurally present, while that of (64) is present only in f-structure. This contrast would be explained if null elements in c-structure were phonologically detectable, while those in f-structure were not. This is an automatic consequence of c-structure being the input to the phonological component.30
Similar evidence against the assumption that PRO is structurally present is not available. Nor is there evidence to the contrary. Out of a desire to reduce the use of null structure to those cases where it is clearly required (as in the case of long-distance dependencies involving so-called "WH-movement"), Bresnan and Kaplan have assumed that PRO, like grammatically controlled subjects, is present only in f-structure. I will, however, make the opposite choice and assume that PRO in Russian is an NP, which is assigned (t PREO) = PRO in the phrase structure expansion.:!l This accounts in a natural way for the alternation found between overt NPs and PRO. More importantly, however, this facilitates the account of structural case marking by allowing case marked subjects to be structurally present.32 This is, however, an assumption that is not standard within LFG.
NOTES
1 See Kaplan and Bresnan (1982) for detailed discussion of the algorithm. See Marantz (1984) for discussion of logical arguments.
] See Bresnan (1980) for more detailed discussion of predicate argument structures. 4 Although Ilhere are some problems in defining these notions precisely, as is discussed in Marantz (1984), for example. 5 Allowing assignment of 0 is not a way of relaxing the requirement that all arguments be assigned functions. Arguments that are associated with the function 0 are interpreted in a specific way (as bound arguments Isee Halvorsen's, 1983, theory of semantic representation of LFGJ).
198 APPENDIX III
(, Extra-grammatical (non-thematic) functions are permitted, however, but only in a very restricted set of circumstances. 7 See Lieber (1980) for a discussion of the nature of the lexicon under this assumption. Kiparsky (1982), Lapointe (1980), Mohanan (1982), and Selkirk (1981) also argue for insertion of fully inflected forms. 8 See Bresnan (1982b and 1977a) for discussion, and for arguments that Passive should be done by a lexical rather than a transformational rule. Y Related lexical items are individually represented in the lexicon. It might be useful to think of the redundancy rules as relating a large, central core of active vocabulary, while these same rules may be used productively to create and interpret peripheral lexical items. This periphery is constantly expanding, though, especially for the language learner. III See, for example, the discussion of Object Control in chapter 5. II As defined in Bresnan (1982a). The clause nucleus (which is basically the LFG version of a "simplex sentence') is also an important domain for anaphoric relations and other phenomena. The reader is referred to Bresnan for discussion. 12 As defined in Bresnan (1979) and Kaplan and Bresnan (1980). 1.1 See definition in Bresnan (1979) and Kaplan and Bresnan (1982). 14 Affix-hopping is discussed in Bresnan (1979). See also Falk's (1984) analysis of the English auxiliary system. I; The observations and sentences in this section are due to Bresnan (1979). 16 Such a construction is in fact possible, but only with the interpretation of keePI' This reading of keep does not usually include human objects. However, a sentence like the following is possible (as pointed out by Ken Hale [personal communication]):
Around the house, Natasha keeps Boris and several other bodyguards.
17 The use of A, N, Y, and P is merely for clarity of exposition. It is not necessary to encode category in functional designations, as is pointed out in Bresnan (1982a). 18 This is a universal constraint on complements. See Bresnan for other principles governing complementation. I" The latter option is preferred in recent work by Bresnan. The former is used here. 211 See detailed argumentation in Bresnan (1982b). 21 The third sentence is from Maling (1983), where there is an interesting discussion of what types of PPs make good complements. 22 Consider, for example, Chomsky's (1981) Projection Principle, which holds that sub categorization is projected to all levels of syntactic representation. 23 See also Grimshaw (1982a, b) for discussion. 24 In this respect, the LFG model differs from that of Chomsky (1981), for example, in that the "Projection Principle" (requiring constituent structure to be the reflection of subcategorization) is not assumed to hold. " The need to state subcategorization restrictions in terms of grammatical functions such as SUB (subject), OB] (object), and the like, rather than syntactic constituents, is argued for in Grimshaw (1979 and 1982a). 26 Elsewhere, COMP may be written as SCOMP. The same closed complement is represented by both notations. The former is used here. 27 This subject is introduced in the phrase structure expansion by the phrase structure annotation: U PRED) = PRO. 2~ If no such relation is established, then the PRO may be interpreted in accordance with discourse context, as in "Reading mystery stories is fun'. This would not involve anaphoric control. 2" The notation used here is that of Chomsky's Government-Binding framework, not of LFG.
LEXICAL FUNCTIONAL GRAMMAR 199
30 In Chomsky (1981 p. 318, fn.), this difference is explained in terms of the case marking on the different types of null elements. He suggests that "non-Case-marked trace is 'invisible' to rules of the PF [Phonologicalj-component, a special case, perhaps, of the more general property ... that elements must have appropriate features to be 'visible' in the interpretive components." These facts fall out of the LFG representation, with no further explanation required. 31 This allows S' to be expanded as
NP YP W PRED) = PRO)
rather than as
(NP) YP (1 SUB PRED) = PRO
J2 Alternatively, one could assume a PS expansion of S to (NP) YP, where, by convention a node would be automatically generated, but optionally labelled. Nothing in the current analysis would be inconsistent with such a convention. This alternative would be more in keeping with lhe approach of Kaplan and Bresnan (1982).
REFERENCES
Academy of Sciences of the U.S.S.R.: 1960a, Grammatika russkogo jazyka, Nauka. Moscow.
-- 1980, Russkaja grammatika Tom II Sintaksis, Nauka, Moscow. Andrews, Avery: 1971, 'Case Agreement of Predicate Modifiers in Ancient Greek',
Linguistic Inquiry 2, pp. 127-152. -- 1980, 'The VP Complement Analysis In Modern Icelandic', Montreal Working
Papers in Linguistics, May, 1976. -- 1982a, 'Long Distance Agreement in Modern Icelandic', in Jacobsen and Pullum
(eds.), 1982. -- 1982b, 'The Representation of Case in Modern Icelandic', in Bresnan (ed.), 1982. Aoun, Joseph: 1985, A Grammar of Anaphora, MIT Press, Cambridge. Atlas, David J.: 1977, 'Negation, Ambiguity, and Presupposition', Linguistics and Philoso
phy I, pp. 321-336. Babby, Leonard H.: 1973a, 'The Deep Structure of Adjectives and Participles in Russian',
Language 49, pp. 349-360. -- 1973b, 'A Note on Agreement in Russian', Glossa 7, pp. 253-263. -- 1975, A Transformational Grammar of Russian Adjectives, Mouton, The Hague. -- 1978, 'Participles in Russian: Attribution, Predication, and Voice', in Comrie (ed.),
1978. -- 1 980a, Existential Sentences and Negation in Russian, Karoma, Ann Arbor. -- 1980b, 'The Syntax of Surface Case Marking', Cornell Working Papers in Linguistics
I,pp.I-32. -- 1984, 'Prepositional Quantifiers and the Direct Case Condition in Russian', in Flier
and Brecht (eds.), 1985. -- 1986, 'The Locus of Case Assignment and the Direction of Precolation: Case
Theory and Russian', in Brecht and Levine (eds.), 1986. Baker, Mark: 1982, 'Objects, Themes, and Lexical Rules in Italian', in Levin et al. (eds.),
1983. Barnetova, Vilma, H. Belicova-Kfiikova, O. Leska, Z. Skoumalova, and V. Strakova: 1979,
Russkaja Grammatika 2: Sintaksis, Academia Praha, Prague. Bellelti, Adriana: 1988, 'The Case of Un accusatives', Linguistic Inquiry 19, pp. 1-34. An
earlier version appears as Lexicon Project Working Papers 8, MIT, 1986. Birnbaum, Henrik: 1964, Studies on Predication in Russian, i: Predicative Case, Short
Form Adjectives, and Predicatives, The Rand Corporation, Santa Monica. -- 1965, Studies on Predication in Russian 11: On the Predicative Use of the Russian
Infinitive, The Rand Corporation, Santa Monica. Borer, Hagit: 1982, Parametric Variation in Clitic Constructions, Ph.D. dissertation, MIT. -- 1984, Parametric Syntax: Case Studies in Semitic and Romance Languages, Foris,
Dordrecht. Borras, F. M. and R. F. Christian: 1971, Russian Syntax: Aspects of Modern Russian Syntax
and Vocabulary, 2nd edition, Clarendon Press, Oxford. Brecht, Richard D., and Catherine V. Chvany (eds.): 1974, Slavic Transformational Syntax,
Michigan Slavic Publications, University of Michigan, Ann Arbor.
200
REFERENCES 201
Brecht, Richard D., and J. S. Levine (eds.): 1986, Case in Slavic, Slavica Publishers, Columbus.
Bresnan, Joan W.: 1977, 'Transformations and Categories in Syntax', in Butts and Hintikka (eds.), 1977.
-- 1978, 'A Realistic Transformational Grammar', in Halle et al. (eds.), 1978. -- 1979, 'A Theory of Grammatical Representation', duplicated lecture notes, MIT. -- 1980, 'Polyadicity', in Hoekstra et al. (eds.), 1980. In revised form in Bresnan (ed.),
1982. -- 1982a, 'Control and Complementation in English', in Bresnan (ed.), 1982. Also,
Linguistic Inquiry 13, pp. 343-434. -- 1982b, 'The Passive in Lexical Theory', in Bresnan (ed.), 1982, -- (ed.): 1982, The Mental Representation of Grammatical Relations, MIT Press,
Cambridge. Brooks, Maria Z.: 1975, Polish Reference Grammar, Mouton, The Hague. Burzio, Luigi: 1981, Intransitive Verbs and Italian Auxiliaries, Ph.D. dissertation, MIT. -- 1986, Italian Syntax: A Government-Binding Approach, D. Reidel, Dordrecht. Butts, Robert E. and Jaakko Hintikka (eds.): 1977, Basic Problems in Methodology and
Linguistics, Part 3 of the Proceedings of the Fifth International Congress of Logic, Methodology and Philosophy of Science, London, Ontario, Canada 1975, D. Reidel, Dordrecht.
Carlson, Lauri 1979, 'Grammatical Case in Finnish', Manuscript, MIT. -- 1981, 'Aspect and Quantification', in Tedeschi and Zaenen (eds.), Tense and Aspect
(Syntax and Semantics 14), Academic Press, New York. Chafe, Wallace L.: 1976, 'Givenness, Contrastiveness, Definiteness, Subjects, Topics, and
Points of View', in Li (ed.), 1976. Chomsky, Noam: 1957, Syntactic Structures, Mouton, The Hague. -- 1965, Aspects of the Theory of Syntax, MIT Press, Cambridge. -- 1970, 'Remarks on Nominalization', in Jacobs and Rosenbaum (eds.), 1970. -- 1980, 'On Binding', Linguistic Inquiry 11, pp. 1-46. -- 1981, Lectures on Government and Binding, Foris, Dordrecht. -- and Howard Lasnik: 1977, 'Filters and Control', Linguistic Inquiry 8, pp. 425-504. -- and Howard Lasnik: 1978, 'A Remark on Contraction', Linguistic Inquiry 9, pp.
268-274. Christian, Reginald F.: 1963, 'Some Consequences of the Lack of Definite and Indefinite
Article in Russian', Slavic and East European Journal 5, pp. 1-11. Chvany, Catherine V.: 1974, 'The Grammar of Dolzen: Lexical Entries as a Function of
Theory', in Brecht and Chvany (eds.), 1974, pp. 78-122. A Russian translation appears in Novoe v zarubeznoj lingvistike, 15 Progress, Moscow.
-- 1975, On the Syntax of Be-Sentences in Russian, Siavica Publishers, Cambridge. -- 1983, 'On 'Definiteness' in Bulgarian, English, and Russian', In Flier (eds.), American
Contributions to the IX International Congress of Slavists, Kiev, Siavica Publishers, Cambridge.
-- and R. D. Brecht (eds.): 1980, Morphosyntax in Slavic, Slavic a Publishers, Cambridge. -- 1986, 'Jakobson's Fourth and Fifth Dimensions (On Reconciling the Cube Model of
Case Meanings with the Two-Dimensional Matrices for Case Forms)" in Brecht and Levine (eds.), 1986.
Cole, Peter, and Jerold M. Sadock (eds.): 1977, Grammatical Relations, (Syntax and Semantics 8), Academic Press, New York.
Comrie, Bernard: 1974, The Second Dative: A Transformational Approach', in Brecht and Chvany (eds.), 1974, pp. 123-150.
-- 1977, 'In Defense of Spontaneous Demotion: The Impersonal Passive', in Cole and Sadock (eds.), 1977.
202 REFERENCES
1978, 'Genitive-Accusatives in Slavic: The Rules and their Motivation', in Comrie (ed,),1978,
-- (ed.): 1978, Classification of Grammatical Categories, Linguistics Research, Urbana. Corbett, Greville G.: 1978a, 'The Agreement Hierarchy', Journal of Linguistics 15, pp.
202-224. -- 1978b, 'Apposition involving dva, tri, cetyre in Russian: A Solution to Worth's
Riddle', Quinquereme I, pp. 258-264. -- 1978c, 'Numerous Squishes and Squishy Numerals in Slavic', in Comrie (ed.), 1978. -- 1979, Predicate Agreement in Russian, (Birmingham Slavonic Monographs 17),
Department of Russian Language and Literature, University of Birmingham, Birmingham. -- 1980, 'Animacy in Russian and Other Slavonic Languages: Where Syntax and
Semantics Fail to Match', in Chvany and Brecht (eds.), 1980. -- 1981 a, 'Agreement with Quantified Subjects in Russian: A Fictitious Linguistic
Change', Russian Linguistics 5, pp. 287-290. -- 1981b, 'Resolution Rules for Predicate Agreement in the Slavonic Languages', Ms.,
University of Melbourne. -- 1983, Hierarchies, Targets, and Controllers: Agreement Patterns in Slavic, Pennsylva
nia State University Press, University Park, Pennsylvania. Crockett, Dina B.: 1976, Agreement in Contemporary Standard Russian, Slavica Publishers,
Ann Arbor. -- 1977, The Scope of Denial in Russian Negative Sentences', Lingua 43, pp. 229-
245. Dahl, Osten, and Fred Karlsson: 1976, 'Verbal Aspects and Object Marking: A Comparison
Between Finnish and Russian', International Review of Slavic Linguistics 1, pp. 1-30. Davison, R. M.: 1967, The Use of the Genitive in Negative Constructions, The University
Press, Cambridge. Deribas, L. A.: 1956, 'Prjamoe dopolnenie pri perexodnom glagole s otricaniem', Russkij
jazyk v skole 17, pp. 21-25. Falk, Y. N.: 1980, 'The English Auxiliary System: A Lexical-Functional Analysis', Ms.,
MIT. -- 1984, 'The English Auxiliary System', Language 60, pp. 483-509. Fassi Fehri, Abdel Kader: 1982, Linguistique arable: forme et interpretation, Faculte des
Lettres et Sciences Humaines, Rabat. Flier, Michael S. and Richard D. Brecht (eds.): 1985, Issues in Russian Morphosyntax,
Slavica Publishers, Columbus. Fodor, Janet D.: 1970, The Linguistic Description of Opaque Contexts, PhD. dissertation,
MIT. Fodor, Jerry and Jerrold Katz: 1964, The Structure of Language, Prentice-Hall, Englewood
Cliffs, N.J. Ford, Marilyn, loan W. Bresnan, and Ronald M. Kaplan: 1982, 'A Competence-Based
Theory of Syntactic Closure', in Bresnan (ed.), 1982. Franks, Stephen: 1986, 'Case and the Structure of NP', in Brecht, R. D. and J. S. Levine
(eds.), Case in Slavic, Slavica Publishers, Columbus. Freidin, Robert, and Leonard H. Babby: 1982, 'On the Interaction of Lexical and Struc
tural Properties: Case Structure in Russian', Ms .. Department of Linguistics, McGill University.
Givan, Talmy: 1973, 'Opacity and Reference in Language: an Inquiry into the Role of Modalities', in Kimball (ed.), 1973.
-- 1975, 'Cause and Control: On the Semantics of Interpersonal Manipulation', in Kimball (ed.), 1975.
Grimshaw, Jane: 1979, 'Complement Selection and the Lexicon', Linguistic Inquiry /0, pp. 279-326.
-- 1982a, 'Subcategorization and Grammatical Relations', in Zaenen (ed.), 1982. -- 1982b, 'On the Lexical Representation of Romance Reflexive Clitics', in Bresnan
(ed.), 1982.
REFERENCES 203
Gvozdev, Aleksandr N.: 1955, Ocerki po stilistike russkogo jazyka, Gos, Ucebno-Pedagog. Izd-vo, Moscow.
Halle, Morris, Joan Bresnan, and George A Miller (eds.): 1978, Linguistic Theory and Psychological Reality, MIT Press, Cambridge.
Halvorsen, Per-Kristian: 1983, 'Semantics for Lexical-Functional Grammar', Linguistic Inquiry 14, pp. 567-615.
Hoekstra, Teun, Harry van der Hulst, and Michael Moortgat (eds.): 1980, Lexical Grammar, Foris, Dordrecht.
Ioup, Georgette: 1975, 'Some Universals for Quantifier Scope', in Kimball (ed.), 1975. -- 1977, 'Specificity and the Interpretation of Quantifiers', Linguistics and Philosophy
1, pp. 233--245. Isacenko, A V.: 1966, '0 grammaticeskom porjadke slov', Voprosy Jazykoznanija 15, pp.
27-34. Jackendoff, Ray: 1972, Semantic Interpretation in Generative Grammar, MIT Press,
Cambridge. -- 1977, X-Syntax: A Study of Phrase Structure, MIT Press, Cambridge. Jacobs, Roderick A, and Peter S. Rosenbaum (eds.): 1970, Readings in English Transfor
mational Grammar, Ginn, Waltham. Jacobson, Pauline, and Geoffrey Pullum (eds.): 1982, The Nature of Syntactic Representa
tion, D. Reidel, Dordrecht. Jaeggli, Osvaldo A: 1980a, On Some Phonologically-Null Elements in Syntax, PhD.
dissertation, MIT. -- 1980b, 'Remarks on To Contraction', Linguistic Inquiry 11, pp. 239-245. -- 1982, Topics in Romance Syntax, Foris, Dordrecht. Jakobson, Roman: 1936, 'Beitrag zur allgemeinen Kasuslehre', Travaux du Cercle Linguis
tique de Prague, pp. 240-299. Reprinted in Jakobson (1971). -- 1958, 'Morfologiceskie nabljudenija nad slavjanskim skloneniem', American Con
tributions to' the IVth International Congress ofSlavists, Mouton, the Hague. -- 1963, 'Les embrayeurs, les categories verbales et Ie verbe russe', in Essais de
tinguistique generate, Editions de Minuit, Paris. -- 1971, Selected Writings, Mouton, The Hague. -- and S. Kawamoto (eds.): 1969, Studies in General and Oriental Linguistics, TEC
Co., Tokyo. Janko-Trinickaja, N. A: 1962, Vozvratnye glagoly v sovremennom russkom jazyke, Nauka,
Moscow. Jespersen, Otto: 1917, 'Negation in English and Other Languages', reprinted in Selected
Writings of Otto Jespersen, pp. 3-152, G. Allen & Unwin Ltd., London. Kaplan, Ronald, and Joan Bresnan: 1982, 'Lexical-Functional Grammar: A Formal System
for Grammatical Representation', in Bresnan (ed.), 1982. Karcevski, Serge: 1927, Systeme du verbe russe: Essai de linguistique synchronique,
Legiografie, Prague. Kayne, Richard S.: 1975, French Syntax: the Transformational Cycle, MIT Press, Cam-
bridge. -- 1981, 'ECP Extensions', Linguistic 1nquiry 12, pp. 93-134. Keenan, Edward L.: 1976, 'Towards a Universal Definition of 'Subject", in Li (ed.), 1976. Keil, Rolf-Dietrich: 1970, 'Zur Wahl des Objectkasus bei verneinten Verb en im modernen
Russisch', ZeitschriJt fur slavische Philologie XXV, pp. 109-133. Keyser, Samuel J., and Thomas Roeper: 1984, 'On the Middle and Ergative Constructions
in English', Linguistic Inquiry 15, pp. 381-416. Ms., MIT. -- 1984, 'On the middle and ergative constructions in English', Linguistic Inquiry 15,
pp.381-416. Kimball, John P. (ed.): 1973, .Syntax and Semantics 2, Academic Press, New York. -- 1975, Syntax and Semantics 4, Academic Press, New York.
204 REFERENCES
Kiparsky, Paul: 1982, 'Lexical Morphology and Phonology', Ms., MIT. Klenin, Emily: 1974, Russian Reflexive Pronouns and the Semantic Roles and NOlin
Phrases in Sentences, Ph.D. dissertation, Princeton University. -- 1978, 'Quantification, Partitivity, and the Genitive of Negation in Russian', in
Comrie (ed.), 1978. -- 1980, 'Individuation: An Historical Case Study" in Chvany and Brecht (eds.), 1980. -- 1983, Animacy in Russian: A New Interpretation, Slavica Publishers, Columbus. Klima, Edward S.: 1964, 'Negation in English', in Fodor and Katz (eds.), 1964. Kout, J.: 1960, 'Prjamoe dopolnenie v otricatel"nyx predlozenijax v russkom jazyke',
Russkij jazyk v skole 1Y60, pp. 27-3 2. Kroch, Anthony: 1979, The Semantics of Scope in English, Garland, New York. Kuno, Susumo: 1970, 'Some Properties of Non-Referential Noun Phrases', in lakobson
and Kawamoto (eds.), 1970. Kurytowicz Jerzy: 1946, 'Ergativnost' i Stadial"nost'v Jazyke', Reprinted in Kury-lowicz
(1960), pp. 95-103. -- 1960, Esquisses Linguistiques, Polish Academy of Sciences, Krakow. Ladusaw, William A.: 1980, Polarity Sensitivity as Inherent Scope Relations, Ph.D. disserta
tion, University of Iowa. Distributed by the Indiana University Linguistcs Club, Bloomington, Indiana.
Lapointe, Steven: 1985, A Theory of Grammatical Agreement, Garland Pub., New York. Lasnik, Howard: 1972, Analyses of Negation in English, unpublished Ph.D. dissertation,
MIT. Levin, Beth: 1983, On the Nature ofErgativity, unpublished Ph.D. dissertation, MIT. Levin. Lori: 1981, 'Lexical Representation of Quirky Case in Icelandic', Ms., MIT. -- 1985, Operations on Lexical Forms: Unacc/.lsative Rules in Germanic Languages,
Ph.D. dissertation, MIT. --, Maika Rappaport, and Annie Zaenen (eds.): 1983, Papers in Lexical-Functional
Grammar, Indiana University Linguistics Club, Bloomington, Indiana. --, and J. Simpson: 1981, 'Quirky Case and Lexical Representations of Icelandic
Verbs', Chicago Linguistic Society 17, pp. 185-196. Li, Charles N. (ed.): 1976, Subject and Topic, Academic Press, New York. Lieber, Rochelle: 1980, On The Organization of the Lexicon. Ph.D. sertation. MIT.
Distributed by the Indiana University Linguistics Club, Bloomington, Indiana. Linebarger, Marcia c.: 1980, The Grammar of Negative Polarity, Ph.D. dissertation,
MIT. Lobanova, N. A.: 1975, 'Semantic Differentiation among Negative Sentence Variants',
Slavic and East European JournallY, pp. 200-204. Lunt, Horace G.: 1958, fi1l1damentals of Russian, W. W. Norton, New York. Maling, Joan: 1983, 'Transitive Adjectives: A Case of Categorial Reanalysis', in F. Heny
and B. Richards (eds.), 1983, Linguistic Categories: Auxiliaries and Related Puzzles, D. Reidel, Dordrecht.
Marantz, Alec: 1984, On the Nature of Grammatical Relations, MIT Press, Cambridge. Marantz, Alec, and Timothy Stowell (eds.): 1982, Papers in ,\:vntax, (MIT Working Papers
in Linguistics 4), MIT, Cambridge. Matthews, William. K.: 1967, Russian Historical Grammar, Athlone Press, London. May, Robert: 1977, Ihe Grammar of Quantification, Ph.D. dissertation, MIT. Distributed
by the Indiana University Linguistics Club, Bloomington, Indiana. Mel'cuk, Igor A.: 1980, '0 PadeZe Cislovogo Vyrazenija', Russian Linguistics 5, pp.
55-74. -- 1984, Poverxnostnyj sintaksis russkix c'islovyx vyraienij, Wiener Slawistischer
Almanach, Vienna. Michelsen, Lew R.: 19611, 'Impersonal Sentences in Russian', in American Contriblltions to
the VIth International Congress ofSlavists, Mouton, the Hague.
REFERENCES 205
Milner, Jean-Claude: 1978, De la syntaxe a l'interpretation, Editions du Seuil, Paris. Milsark, G. L.: 1977, 'Toward an explanation of certain peculiarities of the existential
construction in English', Linguistic Analysis 3, pp. 1-30. -- 1979, Existential Sentences in English, Garland, New York. Mohanan, K. P.: 1982, Lexical Phonology, Ph.D. dissertation, MIT. Distributed by the
Indiana University Linguistics Club, Bloomington, Indiana. -- 1983, 'Move NP or Lexical Rules? Evidence from Malayalam Causativization', in
Levin et al. (eds.), 1983. -- 1986, The Theory of Lexical Phonology, D. Reidel, Dordrecht. Morison, Walter A.: 1959, Studies in Russian Forms and Uses - the Present Gerund and
Active Participle, Faber, London. Mustajoki, A.: 1985, Padeia dopolnenija v russkix otricatel'nyx predloienijax 1: izyskanija
novyx metodov v izucenii staroj problemy, University Press, Helsinki. Naylor, Kenneth E.: 1977, 'A Note on the 'Quantification Form' in Russian', Folia Slavica
1, pp. 89-95. Nekrasov, N. P.: 1905, '0 zamenitel'nyx paddax: roditel'nom i vinitel'nom v sovremennom
russkomjazyke', Izvestija otdelenija russkogo jazyka i slovesnosti 10, pp. 31-65. Nichols, Johanna: 1981, Predicate Nominals: A Partial Surface Syntax of Russian, Uni
versity of California Press, Berkeley. Ostler, Nicholas D. M.: 1979, Case-Linking: A Theory of Case and Verb Diathesis Applied
to Classical Sanskrit, Ph.D. dissertation, MIT. Distributed by the Indiana University Linguistics Club, Bloomington, Indiana.
Paduceva, Elena Y.: 1974, 0 semantike sintaksisa, Nauka, Moscow. Panov, Mixail Y.: 1968, Russkij jazyk i sovetskoe ocestvo, Nauka, Moscow. Partee, Barbara H. (ed.): 1976, Montague Grammar, Academic Press, New York. Patton, Thomas E.: 1978, '011 Strawson's Substitute for Scope', Linguistics and Philosophy
2, pp. 291-304. Perlmutter, David M.: 1978a, 'Impersonal Passives and the Unaccusative Hypothesis', in J.
Jaeger et al. (eds.), Berkeley Linguistics Society 4,157-189. -- 1978b, 'The Unaccusative Hypothesis and Multiattachment: Italian Evidence', paper
presented to the Harvard Linguistics Circle, May 9, 1978. -- and C. G. Rosen (eds.): 1984, Studies in Relational Grammar 2, University of
Chicago Press, Chicago. Pesetsky, David: 1981a, 'Russian Quantification and Categorial Selection', Manuscript,
MIT. -- 1981 b, 'Russian Quantifiers and Category Blindness', paper presented at the North
Eastern Linguistics Society meeting, MIT, November 6, 1981. -- 1981c, 'Russian Quantifiers, Subcategorization and Trace Theory', paper presented
at the Linguistics Society of America meeting, New York City, December 28,1981. -- 1982, Paths and Categories, Ph.D. dissertation, MIT. Peskovskij, A. M.: 1956, Russkij sintaksis v naucnom osvescenii, 7th edition, Gos. Ucebno
Pedagog. Izd-vo, Moscow. Pinker, Steven: 1982, 'A Theory of Acquisition of Lexical-Interpretive Grammars', in
Bresnan (ed.), 1982. Popova, Z. D.: 1973, 'Yinite1'nyj padd pri glagolax s otricaniem', Russkij jazyk za rubeiom
1973, pp. 67-69. Postal, Paul M., and Geoffrey K. Pullum: 1978, 'Traces and the Description .of English
Complementizer Contraction', Linguistic Inquiry 9, pp. 1-29. Pul'kina, I. M., and E. B. Zakhava-Nekrasova: (nd), Russian: A Practical Grammar with
Exercises, Progress, Moscow. Pullum, Geoffrey K., and Paul. M. Postal: 1979, 'On an Inadequate Defense of 'Trace
Theory', Linguistic Inquiry 10, pp. 689-706. Quicoli, A. Carlos: 1972, 'Remarks on Case Agreement in Ancient Greek', Ms., MIT.
206 REFERENCES
Quine, Willard Van Orman: 1960, Word and Object, MIT Press, Cambridge. Rappaport, Gilbert c.: 1978, Two Types of Subject Control in Russian', paper presented
at the Linguistics Society of America meeting, Boston, MA, December, 1978. -- 1979, Detachment and Adverbial Participle Clauses in Russian, PhD. dissertation,
University of California, Los Angeles. -- 1980, 'Deixis and Detachment in the Adverbial Participles of Russian', in Chvany
and Brecht (eds.), 1980, pp. 273-300. -- 1984, Grammatical Function and Syntactic Structure: The Adverbial participle of
Russian, Slavic Publishers, Columbus. -- 1986, 'On a Persistent Problem of Russian Syntax: Sentences of the Type Mne
negde spat", in Russian Linguistics 10, pp. 1-31. ' Rappaport, Maika: 1983, 'On the Derivation of Derived Nominals', in Levin et al. (eds.),
1983. Ravic, R. D.: 1971, '0 vyhore padeZa prjamogo dopolnenija pri perexodnyx glagolax s
otricaniem v russkom jazyke', in Fonetika, Fonologia, Grammatika k semidesjatiletiju A. A. Reformatskogo, Nauka, Moscow.
Reformatskij, Aleksandr A: 1955, Vvedenie v jazykovedenie, Gos. Ucebno-Pedagog. Izd-vo, Moscow.
Restan, Per A: 1960, The ohjective case in negative clauses in Russian: the genitive or the accusative?', Scando-Slavica VI, Copenhagen.
Revzin, Isaak I.: 1978, Struktura jazyka kak modelirujuscej sistemy, Nauka, Moscow. Revizina, O. G., and I. I. Revzin: J 973, 'Vyrazenie soglasovatel'nymi sredstvami znacenija
opredelennosti v slavjanskix jazykax,' Kuznecovskie Ctenija, Institut slavjanovedenija i balkanistiki, Moscow.
-- 1974, 'Ob odnom slucae svjazi soglasovanija s opredelennosfju v slavjanskix jazykax', Sovetskoe slavjanovedeni 1974.
Rochette, Anne: 1982, 'French Infinitival Complements', in Marantz and Stowell (eds.), 1982.
Rosen, Carol G.: 1981, The Relational Structure of Reflexive Clauses: Evidence from Italian, Ph.D. dissertation. Harvard University.
-- 1984, 'The Interface between Semantic Roles and Initial Grammatical Relations', in Perlmutter and Rosen (cds), 1984.
Ruzicka, Rudolf: 1963, '0 transformacionnom opicanii tak nazyvaemyx bezlicnyx predlozenij v sovremennom russkom literaturnom jazyke,' Vopro.IY lazykoznanija XII, pp. 23-31.
Safarewiczowa, H.: 1959 and 1960. Forma dopetnienia hlizsz ego w rosyjskim zdaniu zaprzeczonym, CZt;sc I, Slavia Orientalis 8 pt. 4 77-109, cZt;SC II, ihid .. 9. pt. 1,69-137.
Saxmatov, A A.: 1963, Sintaksis russkogo jazyka, Mouton, The Hague. Schachter, Paul: 1977, 'Reference-Related and Role-Related Properties of Subjects', in
Cole and Sadock (eds.), 1977. Schein, Barry: 1980a, 'On Non-finite Complements in Russian', Manuscript MIT. -- 1982, 'Small Clauses and Predication', Manuscript MIT. Scholz, Friedrich: 1973, Russian Impersonal Expressions Used with Reference to a Person,
Mouton, The Hague. Selkirk, Elizabeth: 1982, The Syntax of Words, MIT Press, Cambridge. Sells, Peter: 1985, Lectures on Contemporary Syntactic Theories, CSLI, Stanford Univer-
sity, Stanford, CA Simpson, Jane: 198'2, The Role of Pro in Case-Agreement', Ms., MIT. -- 1983, Aspects of Warlpiri Morphology and Syntax, PhD. dissertation, MIT. Sorensen, Hans Christian: 1957, Studies on Case in Russian, Rosenkilde og Bagger,
Copenhagen. Stankiewicz, Edward: 1962, The Singular-Plural Opposition in Slavic Languages', Interna
tional lournal of Slavic Linguistics and Poetics V, pp. 1-15.
REFERENCES 207
Stowell, Tim: 1978, 'What Was There Before There was There', in D. Farkas et al. (eds.), Proceedings of the Fourteenth Regional Meeting of the Chicago Linguistics Society, University of Chicago, Chicago, pp. 457-471.
-- 1981, Origins of Phrase Structure, Ph.D. dissertation, MIT. -- 1983, 'Subjects Across Categories', The Linguistic Review 2, pp. 285-312. Strawson, P. F.: 1974, Subject and Predicate in Logic and Grammar, Methuen, London. Svedova, N. Ju. (ed.): 1970, Grammatika sovremennogo russkogo literaturnogo jazyka,
Nauka, Moscow. Timberlake, Alan: 1975, 'Hierarchies in the Genitive of Negation', Slavic and East
European Journal 19, pp.123-l38. -- 1976, 'Subject Properties in the North Russian Passive', in Li (ed.), 1976. -- 1980,'Oblique Control of Russian Reflexivization', in Chvany and Brecht (eds.),
1980. -- 1986, 'The Semantics of Case in Russian Predicate Complements', Russian Linguis
ticslO,pp.137-166. Uglitsky, Zinaida: 1956, 'Accusative and Genitive with Transitive Verbs preceded by a
Negative in Contemporary Russian', The Slavonic Review, pp. 377-387. Vaillant, Andre: 1977, Grammaire comparee des langues slaves, tome V, la Syntaxe,
Editions Klincksieck, Paris. Vanek, Anthony L.: 1977, Aspects of Subject - Verb Agreement, Linguistic Research,
Edmonton. Veyrene, Charles Jacques: 1970, Histoire de la langue russe, Presses Universitaires de
France, Paris. Vinogradov, Viktor V.: 1947, Russkij jazyk, Gos. Ucebno-Pedagog. Izd-vo, Moscow. Ward, Dennis: 1965, The Russian Language Today: System and Anomaly, Hutchinson
University Library, London. Weatherly, J. 8., and Michael K. Launer: 1977, 'A Transformational Study of Some
Russian Infinitive Impersonal Sentences', International Review of Slavic Linguistics 2, pp.215-250.
Williams, Edwin: 1980, 'Predication', Linguistic Inquiry 11, pp. 203-238. Worth, Dean S.: 1959, 'Grammatical and Lexical quantification in the Syntax of the
Russian Numeral', International Journal of Slavic Linguistics and Poetics 1, pp. 177-131.
-- 1984, 'Russian Gen2, Loc2 Revisited', International Journal of Slavic Linguistics and Poetics 29, pp. 295-306.
Yokoyama, Olga T.: 1979, Studies in Russian Functional Syntax, Ph.D. dissertation, Harvard University.
-- 1980, 'Studies in Russian Functional Syntax', in Kuno (ed.) Harvard Studies in Syntax and Semantics 3, Harvard University, Cambridge, MA, pp. 451-778.
-- 1984, 'A Diversified Approach to Russian Word Order', in Flier and Brecht (eds.), 1985.
Zaenen, Annie (ed.): 1982, Subjects and Other Subjects: Proceedings of the Harvard Conference on the Representation of Grammatical Relations, December, 1981, Indiana University Linguistics Club, Bloomington, Indiana.
Zec, Draga: 1985, 'Objects in Serbo-Croatian', Berkeley Linguistics Society 11.
case agreement, see agreement, case alternations xiii, xv, 2, 10, 30
apparent nominative/genitive 65ff, H3, H7-n.24, II H. 159
morphological 19-22,60 syntactic xi
object position 10, 30, 32ff, 60, 151
subject position 123, 129, 150. 151
210
with prepositions 12 assignment xiii, xiv
by prepositions 12, 120-n.13 direct case condition xiv, 87 -n.31,
120-n.13,159 inherent 7-9, 61, 165ff, 16H, 169,
172-n.l0 lexical, see case, assignment, inherent mechanisms I, 13,9H-101, I 59ff quirky, see case, assignment, inherent structural 1, 5-10. 14, 15. I 65ff,
166, 168, 172-n.9 in copular constructions 14 to post-verbal position xiii, 7, H, 9.
1O,61,16H to pre-verbal position 26-n.3, 172-
n.9 to grammatical functions 1,9, 10, 15.
166,167 endings, see inflection, case endings features 60
functions, grammatical xiii, xiv, 1, 180 ACOMP, see complements, adjectival ADJ, see adjuncts DOM, see domain NCOMP, see complements, noun OBJ, see objects PCOMP, see complements, prepositional PRED, see predicate
212 INDEX OF SUBJECTS
SUB, see subjects VOMP, see complements, verbal XCOMP, see complements
de-argumented forms 153-n.20 perfectivelimperfective 153-n.19 transitive 74, 75
well-formedness, functional 184ff, see coherence; consistency; completeness
word order xiii, 1, 72-73, 121-n.33 basic 7,10 neutral 66, 72, 73, 86-n.15, 109
X-bar theory 7, I 19-n.9, 168, 178
STUDIES IN NATURAL LANGUAGE AND
LINGUISTIC THEORY
Managing Editors JOAN MALING and LUIGI RIZZI
Luigi Burzio, Italian Syntax: A Government-Binding Approach. xiii + 468 pp.1986.
William D. Davies, Choctaw Verb Agreement and Universal Grammar. xi + 202 pp., 1986.
Katalin E. Kiss, Configurationality in Hungarian. 268 pp., 1987. Douglas Pulleyblank, Tone in Lexical Phonology. xii + 249 pp., 1986. Lars Hellan and Kirsti Koch Christensen, Topics in Scandinavian Syntax.
vii + 273 pp., 1986. K. P. Mohanan, The Theory of Lexical Phonology. xii + 219 pp., 1986. Judith L. Aissen, Tzotzil Clause Structure. xxiii + 290 pp., 1987. Takao Gunji, Japanese Phrase Structure Grammar: A Unification-Based
