- 1 This paper attempts to cover a great deal of data in a limited space; accordingly, a number of pro (...)
1Within the framework of Government & Binding theory, much of the current work on the syntax of sentential negation addresses two essential questions; its status in X-bar theory and its base-generated position with respect to other categories in the sentence. This research has to a large degree concentrated on the effect of negation on syntactic movement. In this way, if it appears to interfere with verb-raising or clitic climbing, it is accorded the status of a head of a phrase; if it interferes with adjunct extraction, it is accorded the status of a Specifier of a phrase. While its status as a potential barrier will play a role in this paper, I will propose that it also has a somewhat more positive role to play. Following work by Kitagawa (1986), Pollock (1989), and many others, what follows will assume that sentential negation generates the functional projection Negation Phrase or NegP. I will claim that the Specifier of NegP in Literary Welsh, Colloquial Welsh, and Russian provides an additional position in which Noun Phrases can be Case-licensed.1
2This paper represents work in progress and as such, certain of the ideas presented here are somewhat tentative in nature. Nevertheless, it is hoped that the paradigms under discussion and the theoretical questions they raise will be of interest to the reader.
- 2 The gaps marking the site of subject and object relativization, shown as [e], reflect the basic VS (...)
3The first paradigm to be discussed is taken from Literary Welsh (LW). LW is traditionally described as having two distinct relativization strategies called the direct and indirect strategies, each of which is associated with relativizing out of certain positions within the sentence. In the direct strategy, the particle a precedes the relative clause. As shown in (1)-(2), an instance of subject relativization, there is a gap in the relativization site, and the verb does not exhibit agreement with the relativized NP.2
(1) y dynionia ddarllenodd [e]i y llyfr
the men PT read-3sg the book
the men that read the book
(2) *y dynionia ddarllenasant [e]i y llyfr
the men PT read-3pl the book
4In the indirect strategy there is still a gap in the relativization site which, as in the direct strategy, cannot be filled by a pronominal, but we find agreement appearing on the head governing the relativization site. In this case, the particle that precedes the relative clause is yr. This is illustrated in (3)-(4), where a periphrastic direct object is relativized.
(3) y llongiy gwnaeth Sion ei werthu [e]i
the boat PT did Sion 3sm-sell
the boat that Sion sold
(4) *y llongiy gwnaeth Sion gwerthu [e]i
the boat C did Sion sell
5An interesting effect is induced by negating a relative clause. In those cases which normally require the direct strategy, i.e. subject relativization, negating the relative clause appears to force the indirect pattern; agreement with the relativized NP subject is obligatory. This is shown in (5)-(6).
(5) y dynionina ddarllenasant [e]i y llyfr
the men NEG read-3pl-past the book
the men that didn’t read the book
(6) *y dynionina ddarllenodd [e]i y llyfr
the men NEG read-past the book
6The second paradigm to be discussed comes from Russian, where negating a clause appears to alter the Case-assigning properties of the sentences. If we take as an example a regular transitive verb, the direct object normally surfaces with Accusative Case. When the sentence is negative, the direct object may surface either with Accusative or Genitive Case.
(7) ja vizu knigu
I-NOM see book-ACC
I see the book
(8) ja ne vizu knigi
I-NOM NEG see book-GEN
(9) ja ne vizu knigu
I-NOM NEG see book-ACC
An interesting wrinkle involves the interpretation of these sentences; If an NP that could surface with Genitive Case appears instead with Accusative Case, that NP tends to be interpreted as definite. In this way, an ambiguity with respect to definiteness that is present in affirmative sentences in Russian can be removed in a negated sentence, as shown in (11).
(10) Ja ne vizu knigi
I-NOM NEG see book-GEN
I don’t see a/the book
(11) ja ne vizu knigu
I-NOM NEG see book-ACC
I don’t see the book
7The final set of data to be discussed comes from negation in Colloquial Welsh, which differs in certain interesting ways from Literary Welsh. First, it requires an additional negative marker, similar to French pas in distribution and placement, and second, a preposition appears preceding the direct object in negated sentences. This preposition is not found in the corresponding LW sentences, as shown in (13).
(12) Phalodd Sion ddim *(o) ’r ardd.
NEG-dug-3sg Sion NEGP the garden
Sion didn’t dig the garden
(13) Ni phalodd Sion yr ardd.
NEG dug-3sg Sion the garden
Sion didn’t dig the garden.
Where this set of facts interact with those of Russian is that the preposition does not surface in CW when the direct object is indefinite.
(14) Rosim i ddim lliw ynddo ariod.
NEG-put-1sg I NEG color in-3sgm ever
I didn’t put coloring in it ever.
8These facts show sentential negation interacting in interesting ways with Case and agreement, and that this interaction appears to be sensitive to the notion of (in)definiteness. Below, some background on the theoretical framework I will be assuming is given.
9In what follows I will assume a Government & Binding (GB) approach to syntax, as articulated in Chomsky (1986) and elsewhere. I will make crucial reference to the distinction between levels of the grammar at which certain operations take place; D-structure, S-structure, and Logical Form.
10Regarding phrase structure, I will assume that all lexical heads project a phrase, also termed a maximal projection. Phrase structures obey a binary branching requirement (Kayne 1984). Abstracting away from the directionality of branching, a maximal projection will have the structure given in (15), with X the head.
11Following Larson (1987), I will assume throughout that transitive verbs (verbs which subcategorize for both an internal and an external argument) generate a two-tiered VP, the verb being generated in the head of the lower VP, and the head of the higher VP being empty. The head of the higher VP will be referred to as delta, and for ease of exposition I will refer to this VP as deltaP in order to distinguish it from the lower VP. I will adopt Koopman and Sportiche’s (1988) VP-internal subject hypothesis, in which the surface subject is generated within deltaP. Further, following a version of Baker’s (1988) Unity of Thematic Assignment Hypothesis (UTAH), NPs bearing an agentive theta role will be generated in Specifier of deltaP, and patients and themes will be generated within VP. This is shown in (16).
- 3 Sportiche (1990), Johnson (1990), Mahajan (1990), and others.
12In keeping with much recent work, this analysis will assume that Case is checked in a licit Specifier-head configuration, either at S-structure or at LF.3 This is formalized as a Licensing Condition on Chains (LCC), given in (17). The definition of a licit Specifier-head configuration is given in (18).
(17) LCC: A chain is Case-checked at S-structure or at LF.
(18) Licit Spec/head Configuration:
- 4 The restriction that either the head or the specifier must dominate the head of a chain accounts f (...)
13A licit Spec/head configuration is one in which a member of the chain of the Case-assignee is in a Spec/head configuration with a morphologically complete member of the chain of a Case-licensing head. Either the head or the Specifier must dominate the head of a chain.4
14I will adopt the view, advocated in Sportiche (1990) and Mahajan (1990), that agreement with an NP is the surface reflex of Case-checking in a licit Spec/head configuration established at S-structure. Armed with these assumptions, we can now address the paradigms involving negation, Case, agreement, and definiteness in Literary Welsh, Russian, and Colloquial Welsh.
15A negative sentence in standard Literary Welsh includes a negative particle which immediately precedes the tensed verb.
(19) Rhedodd Sion i ffwyrdd.
ran Sion away
Sion ran away.
(20) Ni redodd Sion i ffwyrdd.
NEG ran Sion away
Sion didn’t run away.
- 5 Harlow 1981, Sproat 1985 and others.
16On the assumption that the S-structure position of the tensed verb is in T, where it has moved to support tense morphology,5 and if the negative marker itself has not undergone movement, then the most straightforward initial assumption is that NegP is generated above TP. This is illustrated in (21).
- 6 The 3sg. agreement marking that surfaces on the tensed verb when the subject is non-pronominal is (...)
- 7 An asterisk outside the parantheses means that the sentence is ungrammatical if the item is not pr (...)
17In non-relativized structures in Welsh, the agreement paradigm is straightforward. Agreement cannot surface if the argument in question is a non-pronominal NP but is obligatory if the argument is pronominal. Agreement can appear on a verb, a preposition, a noun in a possessive construction, and a verb that is not inflected for tense (referred to as a verb-noun in the literature). A pronominal that triggers agreement then surfaces optionally. The basic agreement paradigm is illustrated in (22)-(24).6,7
(22) Subject of tensed verb:
Gwelodd (*gwelsant) y dynion y ci.
saw-3sg (saw-3pl) the men the dog
The men saw the dog.
Gwelsant (*gwelodd) (hwy) y ci.
saw-3pl (saw-3sg) (they) the dog
They saw the dog.
- 8 See Rouveret 1991 for an interesting analysis of PPs as incorporating an additional functional pro (...)
(23) Object of preposition:8
Mae Wyn yn son amdano (ef).
is-3sg Wyn PROG speak about-3sgm (him)
Wyn is speaking about him.
Mae Wyn yn son am (*amdano) Sion.
is-3sg Wyn PROG speak about (about-3sgm) SionWyn is speaking about Sion
(24) Object of Verb-noun (periphrastic construction)
Mae Sion yn *(ei) ddarllen (ef).
is-3s Sion PROG 3sgm read (it)
Sion is reading it.
Mae Sion yn (*ei) ddarllen y llyfr.
is Sion PROG (3sg) read the book
Sion is reading the book.
18Given that agreement is seen as a reflex of Case-checking at S-structure, the fact that non-pronominal NPs do not trigger agreement in Welsh can be accounted for by claiming that they are not Case-checked until LF. Pronominal NPs, on the other hand, have to be Case checked at S-structure. I will maintain that they do not move to the Specifier position until LF; but at S-structure they may form an A-chain with a null pleonastic in Specifier position. This chain is Case-checked at S-structure, its head triggering agreement.
19The chain that is formed can be compared to the chain between there and the argument NP in existential sentences like the one below in English, which exhibits agreement on the verb although the NP the three men does not raise to replace the expletive in subject position until LF.
(25) [There]iwere [three men]i in the room
*[There]iwas [three men]i in the room.
20The pronominal NP can only be dropped under identification by agreement morphology on the governing head (Rizzi 1986). The S-Structure configurations of a pronominal prepositional object and a pronominal subject are illustrated in (26). The agreement that is triggered on T and P results from co-indexing between the Case-licensing heads P and T, and the element in their Specifier.
21One of the more problematic facts for previous analyses of Welsh to account for is the inability of pronominal direct object to trigger agreement on a tensed verb. This is shown in the contrast between (27), where the direct object is in a regular VSO sentence, and (28), where the object is in an AUX-SVO, or periphrastic sentence, following a verb-noun.
(27) Welodd Wyn *(ef).
saw-past Wyn himWyn saw him.
(28) Mae Wyn wedi ei weld (ef).
be-3s Wyn PERF 3sm see (him)
Wyn has seen him.
22Under the analysis proposed here, whereby the notion licit Spec-head configuration is defined in terms of morphological completeness, this distinction is not unexpected. The verb in a tensed sentence is arguably not morphologically complete until it has incorporated with the tense morphology generated under T. After raising to T, the V is in a licit Spec/head configuration with Spec/TP, but this position is filled by the chain formed by the subject. Accordingly, a direct object pronominal will never be in a licit Spec-head configuration with the chain of its Case-licensing head, if that head is a verb in T. This is illustrated in (29).
23On the other hand, the lower verb in a periphrastic construction given in (30) does not have to raise to T; it is morphologically complete without [tense] features. Thus, the null pleonastic co-indexed with the object in the Specifier of the lower VP is in a licit Spec/head configuration, and agreement surfaces on the untensed verb.
24Note that while this analysis rules out Case-licensing the pronominal direct object of a [+tense] verb via Spec/head co-indexation at S-structure, the manner in which these NPs manage to satisfy the Case-licensing requirement has not been specified. Following a suggestion made in Sportiche (1990), I will suggest that an argument can be Case-licensed exceptionally under government rather than by Spec/head co-indexation if the following conditions hold.
25First, the capability is sensitive to movement through a position in the structure; which activates the Case-licensing mechanism. In Welsh, the position at which the V may assign Case under government is delta, the head of the higher VP in a Larsonian shell. The member of the verb-chain in delta can assign Case to an NP in the Specifier of the functional projection it immediately dominates. The LF representation of a VSO sentence in Welsh is given in (31).
26We can use this treatment of agreement and Case to propose a revision to the analysis of relativization in Welsh, and the role that sentential negation has to play in this paradigm.
- 9 Awbery 1977, Harlow 1981, Rouveret 1990.
27Previous analyses have treated the gap in the relativized position in the indirect strategy as different in kind from that found in the direct strategy. They make the quite reasonable assumption that indirect strategy is used when the direct strategy is blocked, i.e. when the relativized position is in a sense too far away. In this way, the indirect strategy is seen as a kind of resumptive pronoun strategy which involves no movement.9
28We will consider a quite different analysis of these facts. The unacceptability of an in situ pronominal in the relativization site co-occurring with agreement on the governing head, which distinguishes the indirect strategy from non-relativized contexts, will be taken as evidence that the element in the relativized position is a trace. I propose that both the direct and indirect patterns are derived by operator movement to Spec/CP, and that the surface differences between them can be derived under the analysis of Case and agreement described above.
29Consider first the salient fact that subjects and direct objects of tensed verbs pattern together in exhibiting the direct pattern of relativization. There is one feature which subjects and VSO direct objects in Welsh share that distinguishes them from all other arguments: the heads that are responsible for Case-licensing these NPs, V and T, are in the same position at S-structure. In a simple sentence, they are both under T. In a relative clause, I assume that the [V+T] complex raises to C, wherever possible.
30Both the differences in agreement and the different pre-sentential particles will be argued to follow from the LCC- the chain created by A-bar movement to Spec/CP must trigger agreement, and subjects and direct objects of tensed Verbs both trigger agreement on the head of CP. Other Case-licensing heads (i.e. prepositions, [-finite] verbs, and noun heads of possessive NPs) remain within their projection, and agreement is triggered lower in the structure.
31By extending the LCC to chains created by A-bar movement, the possibility exists that more than one licit Spec/head configuration may exist. I will maintain that a condition operates such that agreement is always realized at the lowest licit Spec/head configuration. This is formulated as the Condition on Agreement Realization, under (32):
(32) Condition on Agreement Realization (CAR):
Agreement is established at the lowest licit Spec/head configuration.
32The distinct complementizers; y in the indirect pattern and a in the direct pattern, will be derived as follows. First, I take the default head of CP to be y; the particle found in the indirect strategy as well as preceding regular embedded clauses. If C is in an agreement configuration at S-structure, (that is, if an A’ chain formed by A-bar-movement is Case-checked in Spec/CP) then C surfaces as a. The particle a is in this way analyzed as the realization of [C+agreement]. The form agreement takes is highly impoverished, and does not vary according to number, person or gender. However, it is sufficient to satisfy the requirements of the LCC. The configuration triggering an agreeing C is illustrated in (33).
33Before showing how this can provide an explanation for the role played by sentential negation in the paradigm, I will propose an analysis for relativization resulting in the indirect and direct patterns, to show that the distinction between a movement and a non-movement strategy is not required to account for their differences.
34As an illustration of relativization resulting in the indirect strategy, consider relativization out of a prepositional phrase. Under the assumptions specified above, the object of the preposition wh-moves to Spec/CP. Movement must proceed through its Case-licensing position; in this case, the Specifier of PP. The A’ chain formed by wh-movement includes a trace in Spec/PP. Since the head of PP does not move, its position determines the lowest potential Spec/head configuration for agreement. Thus, in accordance with CAR, agreement is realized on P. Since agreement is realized only once on a chain, the form of the complementizer is the non-agreeing C, y. This is illustrated in the tree under (35), where subject agreement is realized on T, agreement with the relativized prepositional object is realized on P, and no agreement is realized on C.
(34) y dyn y canodd ef amdano (*ef)
the man C spoke-3sg he about-3sm (*him)
the man that he spoke about
35The agreement pattern, complementizer selection, and unacceptability of in-situ pronominal exhibited in the indirect strategy of relativization are thus accounted for.
- 10 Due to limitations of space I will not discuss subject relativization here. See de Freitas & Noona (...)
36Now consider the direct pattern of relativization, which is found with subjects and with direct objects of tensed verbs.10 Under direct-object relativization, at S-structure the direct object operator is in Spec/CP. The Case-licensing head for the direct object is V, which has raised to C. Since the Case-licensing head for the direct object is in C, the lowest potential Spec/head configuration is again between the complex head dominated by C and its Spec position. Agreement with the relativized direct object is realized on C, the head of the complex head, which then surfaces as a. Subject agreement is realized on T.
(36) y dyn a welodd Sion [e]
the man C+agr saw-3sg Sion
The man that Sion saw.
37In this way, the agreement patterns and complementizer choice associated with the direct and indirect strategies reflect the distinct S-structure positions where Case-licensing takes place. Next, consider how this analysis accounts for sentential negation apparently forcing the indirect pattern of agreement.
38Consider first negated subject relatives, which show a fairly straightforward pattern. Compare subject relativization out of an affirmative sentence, as illustrated under (38), where agreement is not permitted, with the negated case under (39), where person/number agreement is obligatory. In neither case can a pronominal surface in the relativization site.
(38) y dynion a ddarllenodd (*ddarllenasant) [e] y llyfr
the men C+Agr read-3sg (*read-3pl) the book
the men who read the book
(39) y dynion na ddarllenasant (*ddarllenodd) [e] y llyfr
the men Neg read-3pl (*read-3sg) the book
the men who didn’t read the book
39I have suggested that negation heads its own projection and is generated above TP. I will maintain that NegP blocks verb movement into C via Travis’s (1984) Head Movement Constraint (HMC). In a negated relative clause, the Case-licensing head for the relativized subject (T) remains in T, so the lowest Case Spec/head configuration for the subject is Spec/TP. As above, although movement to Spec/CP still takes place (as shown by the inability of a subject pronominal to co-occur with subject agreement on the verb), agreement is realized on T, surfacing as morphological person/number agreement.
40I will assume that the negative marker found in negated relatives is a synthetic form, created by the head of NegP raising to C. Neg-raising to C can occur whether or not raising of the V+T complex to Neg takes place. The former is a possibility made available in work by Rizzi & Roberts (1989), who suggest that the structure given in (iii) below is acceptable, following head movement from Y to X, as it is the head of the X0 created by syntactic raising that has raised further, leaving behind a trace.
41If we accept that this is what is happening in Welsh, na(d) can be viewed as the morphological merger of the negative marker and the complementizer y.
- 11 This hypothesis receives independent support from negated relatives involving periphrastic sentenc (...)
42In this way, there are two potential derivations in negated relative clauses. In one, the [V+T] complex remains in T, and Neg raises independently to C. In the second, the [V+T] complex adjoins to Neg, but Neg itself then raises to C. In either derivation, the merger of C and Neg results in the form na(d). In either derivation, the tensed verb cannot raise all the way to C. It cannot skip over NEG and raise to C without violating the head movement constraint. If it adjoins to NEG, then the [Neg-V+T] complex is still blocked from combining with C due to morphological constraints (Welsh allows at most one element to prefix to the tensed V); in other words, a tensed V may raise to NEG or to C, but not both.11
43While it is generally theoretically undesirable for an analysis to allow two possible derivations under the same set of initial conditions, in the following section I will argue that these two potential landing sites for the tensed verb (under T or under NEG) trigger distinct agreement facts for a VSO direct object.
44When a direct object is relativized out of a negated clause, one option can be referred to as the ’no agreement’ option. The direct object pronominal is obligatory.
(42) y llyfr na ddarllenais i *(ef)
the book Neg read-1sg I (it)
the book which I didn’t read
45I have suggested that the intervening negation prevents the verb from moving into C. While this presents no problem for the relativized subject, for which agreement is simply established at TP as opposed to CP, it does present a problem for the object; the same problem confronted by a VSO direct object in non-relativized structures. The Verb in T provides no licit Spec/head configuration for the direct object.
46This inability of a relativized direct object to satisfy the LCC in negated relative clauses forces a true resumptive strategy.; no operator movement takes place, and instead a base-generated operator in Spec/CP A-bar binds the object pronoun in its base-position. In this way, the asymmetry between subjects and VSO direct objects in negated relatives is captured: The former reflect agreement triggered lower in the structure but still involve wh-movement, but the latter cannot satisfy the LCC and remain in situ.
47In the alternative option of negated object relatives, we find an object agreement marker on the negative marker itself. Thus, (43) exists as a variant of (42), with the agreement marker s surfacing on Neg. In these cases, the object pronoun is optional.
(43) y dyn nas gwelais i (ef)
the man Neg-3sg saw-1sg I (him)
the man who I didn’t see
- 12 The claim that this agreement is triggered in Spec/NegP rather than Spec/CP or some other category (...)
48Since agreement is taken here to be a reflex of a licit Spec/head configuration at S-structure, agreement on the Negative element leads us to posit that a licit spec-head configuration can be established between the head of NegP and an element in its Specifier position; that NegP introduces a potential Case-licensing configuration in which the direct object can satisfy the LCC.12
49Note, however, that this construction is unlike the pattern of relativization found with relativized subjects in negated clauses, which trigger agreement in Spec/TP. With direct objects, a pronominal is free to occur in direct object position, doubling agreement on NEG. This is taken to be possible only where movement has not occurred, i.e. in unrelativized structures. This suggests that these structures involve agreement with an A-chain, rather than agreement with an A-bar chain formed by wh-movement.
50By deriving object agreement on Neg by formation of a A-chain, we would expect agreement on Neg in unrelativized structures as well. This is substantiated in (44)-(45).
(44) Nis gwelodd Wyn (ef).
Neg-3sg saw Wyn himWyn did not see him.
(45) Gwn nas gwel Wyn (ef).
know-1sg neg-3s see Wyn himI know that Wyn will not see him.
- 13 Thus, movement of the VSO direct object in a negated relative clause to spec/CP is not possible, i (...)
51This shows that the possibility of a VSO direct object triggering agreement on Neg is not dependent on whether or not it enters into an A-bar relation; It is simply a Case-licensing option made available by the projection of sentential negation.13
52However, this paradigm, specific to VSO direct objects in negated sentences, differs from agreement in non-relativized contexts in one important respect. In non-relativized contexts, agreement is required wherever it is possible. In negated VSO direct object relatives, however, agreement may but need not surface on the negative particle. Naturally, if there is no agreement, then the pronominal is obligatory, just as in the non-relativized contexts.
53This is a rather surprising optionality in the agreement system of Welsh. I will suggest that the apparent optionality reflects whether or not the tensed verb has raised to Neg in negative clauses; if it does, then the Spec of NegP can (and therefore does) provide a Case-licensing position for the direct object. If the verb remains in T, no such position is available for the direct object, and an in situ pronoun is required. In this way, it is not agreement but rather V-raising which encodes an optional aspect.
54In this section, I have proposed that the head of NegP can be a barrier for head-raising, and the Specifier of NegP can provide a Case-position for NPs in Literary Welsh. In what follows, this analysis is extended to the phenomenon of the Genitive of Negation in Russian.
- 14 This can be seen as a Relativized UTAH (Larson 1990:601): Identical thematic relationships are rep (...)
55As above, I will assume a version of UTAH (Baker 1988), such that agents are always generated in a position structurally higher than the position where patients/themes are generated.14 However, in Russian, the base position for NegP will be assumed to be below deltaP. In this way, NegP is base-generated (crucially) below the agent but above the patient/theme argument. Furthermore, unlike in Welsh, Russian Case-checking at S-structure requires actual movement of the NP into Specifier position, and is not satisfied by the creation of a chain headed by a null pleonastic. Thus, the unmarked Russian word order is SVO, with the Nominative subject (in Spec/TP) preceding the verb in T.
56Structural Case-licensing will be analyzed as for Welsh; agreement will be treated as the reflex of S-structure Case-checking. Since only the NP marked with Nominative Case triggers agreement, I will assume that only Nominative Case-checking is an S-Structure phenomenon; other Cases are checked at LF. Also as in Welsh, it will be assumed that Spec/VP in a [+tense] sentence is not a licit Spec-head configuration for Case-checking, because the head is not a morphologically complete member of the verb-chain until it has combined with T. Accordingly, in a [+tense] sentence, Accusative Case is assigned at S-structure via government by the trace of the verb chain in Delta. A difference between the Russian and Welsh Case-system is that in Russian there is no distinction between the levels of the grammar at which Case-licensing of pronominal and non-pronominal arguments takes place; all NPs must be in their Case position at S-structure.
57In the preceding section it was argued that Spec/NegP is a potential Case-licensing position for NPs if the V raises into the head of NegP. In this section, we will refine somewhat the analysis of what features are required for a head to Case-license an element in its Specifier. It will be argued that the head of NegP is specified [+Case], but that it requires the feature [+V] in order to be a Case-licensing head. This will account for the ability of an NP to surface with the Genitive of Negation even when the argument of a [-Case] unaccusative or passive verb.
58The basic Russian word order is S(ubject) V(erb) O(bject). Phrases can be scrambled fairly freely, obscuring the underlying SVO order, but the scrambled sentences are typically focussed or topicalized constructions, and the resulting obligatory stress makes the derived nature of these sentences clear. Phrases are head-initial. The minimal initial hypothesis, in keeping with the assumptions adopted in this thesis, is that Russian phrase structure consists of head-initial phrases, with specifiers occurring to the left (i.e. preceding the head and its complements).
- 15 Travis 1991 proposes that a functional category (AspP) is generated inside a Larsonian VP.
59Since we base our analysis of the Genitive of Negation on Case-licensing options in negative contexts, we must establish the base position of NegP in Russian. Sentential negation in Russian, represented by the element ne, immediately precedes the first verbal element in a clause. These facts might be taken to suggest that NegP in Russian is generated above TP, as suggested by Zanuttini (1990,1991) for Italian, or even higher, above AgrSP, as suggested by Holmberg et al (1990) for Finnish. However, I will maintain that NegP is generated as a functional category between the two VP tiers in Russian.15
60Generation of NegP in different positions is an option made explicit in Ouhalla (1990) and Pearce (1991), where a parameter is proposed which allows NegP in a given language to be generated above either TP or VP. A NegP that is generated below TP is proposed for French by Pollock (1989), for Italian by Belletti (1990), and for West Flemish by Haegeman (1991).
- 16 Baker 1988.
- 17 Russian differs from Literary Welsh in this respect, where raising to NEG is a case of adjunction, (...)
- 18 This is consistent with the fact that no lexical material may intervene between the negative marke (...)
61Given that NegP is generated below the S-structure position of the verb in a tensed sentence (under T), its S-structure position preceding the verb results from syntactic affixation.16 The head of NegP in Russian is analyzed as a morphologically dependent item, like Tense morphology.17 To use Moritz’s (1989) terminology, it is lexically specified as dominating an empty slot to its right, into which the V must raise. This head-to-head movement, being an instance of substitution, results in a Neg head that dominates both a Neg and V element. Following V-raising to Neg, further raising (i.e. to delta or to T to support tense morphology) necessarily carries the negative marker along with the verb.18
62Before addressing the assignment of Genitive Case, we must determine how Accusative Case is assigned, as the relationship between these two Case options is crucial to the analysis. As in the preceding chapter, in a [+tense] sentence, Accusative Case is assigned under government, by the member of the verb chain in delta to an NP in a Specifier position. Hence, in an affirmative sentence, the patient NP raises to Spec/VP and is assigned Case under government.
63Now, consider how this system is affected by the presence of NegP which intervenes between delta and the Spec/VP. The V raises through Negation, then into delta, then on to T. The member of the V-chain in delta is a trace of the [Neg+V] complex, and can potentially assign Case under government to an NP in the Specifier of the functional projection it immediately dominates, which in this case is Spec/NegP.
64However, what we find in such sentences is Genitive Case, not Accusative Case. This is compatible with two possible explanations: First, it is conceivable that a V which has combined with NEG assigns Genitive instead of Accusative Case under government, perhaps through some kind of feature percolation à la Di Sciullo and Williams (1987). Second, it is possible that the NP in Spec/NegP is Case-licensed through Spec/head co-indexation; that while Spec/VP is not a licit Spec-head configuration in a [+tense] sentence (the verb at that point in the derivation not meeting the morphological completeness requirement of the definition of licit Spec-head configuration), the combination in the syntactic component of V and NEG renders Spec/NegP a licit configuration.
65Based on data from unaccusative and passive verbs, I will argue that the latter is the case. As I will show in the following section, these verbs are not specified as [+Case] (which motivates NP-raising to NOM position in affirmative sentences). However, the head of NegP provides the needed [+Case] features which, combined with the [+verbal] features of the V, makes Spec/NegP a licit Spec-head configuration..
66First, consider the ramifications of the first proposal; analyzing the Genitive of Negation as an altered version of the verb’s ability to assign Accusative Case under government. We would therefore expect that only sentences whose verbs are Case-assigners should exhibit the Genitive of Negation. Such a proposal appears to be immediately contradicted by the occurrence of Genitive of Negation on the D-structure objects of passive and unaccusative verbs; verbs which do not project an external argument. The relation between a verb’s internal argument structure and its ability to assign Accusative Case is described as a kind of mutual dependency in Burzio’s Generalization (BG), given below.19
T= Theta marking of an external argument, and A= accusative Case.
67The effect of BG can be illustrated with the behaviour of passive verbs. It has been claimed that passivization suppresses the verb’s ability to assign a theta-role to the external argument of a transitive verb. The internal argument (THEME / PATIENT) of the verb surfaces with the Case normally assigned to the external argument; Nominative Case, and, in English, appears in subject position. Under standard GB analyses, it is claimed that this follows from the passive verb’s failure to assign Accusative Case, and that this forces the argument generated in complement position to raise to the position where Nominative Case can be assigned.
68BG isolates exactly those verbs in Russian that have been claimed not to assign Accusative Case; verbs whose sole argument surfaces with Nominative Case in an affirmative sentence. However, they are also the verbs that allow the Genitive of Negation on their complements. We will therefore conclude that the ability to Case-license an NP with the Genitive of Negation is independent of a verb’s Case-assigning potential.
- 20 This can be compared with Yafei Li’s (1990) analysis of Case-licensing as affected by syntactic af (...)
69We must establish what makes a head capable of Case-licensing an NP in its specifier. Let us say that such Case-licensing requires positive values for the features [V(erb)] and [Case].20 This is the feature specification for a regular Case-assigning verb. A passive or unaccusative verb does not have the specification [+Case], but retains the specification [+V].
70I will propose that the head of NegP is inherently specified as [+Case]. Alone, it cannot Case-license an NP in its specifier; as we saw in Literary Welsh, the Specifier of NegP becomes a potential agreement configuration only if the verb moves into it. In combination with the [+V] features of a verb, the head of NegP is specified as [+Case] and [+V], hence its Specifier becomes a valid Case position for Spec-head-co-indexation. Below, we will show how this analysis accounts for the structural restrictions on the Genitive of Negation.
71Consider a negated transitive sentence. The NP that would surface with Accusative Case in the affirmative equivalent can surface with Genitive, but the NP that would otherwise surface with Nominative cannot. Given that NegP is generated below the D-structure position of the agentive NP, this constraint is expected The agent is generated in Spec/DeltaP, higher then the base position of NegP. As such, its only potential Case position is Spec/TP, where it surfaces with Nominative. Genitive Case on such an argument would require lowering to Spec/NegP, which would result in a trace in Spec/DeltaP which violates the Empty Category Principle (ECP: Chomsky 1986). Only NPs generated below NegP can be Case-licensed in Spec/NegP.
72Objects of prepositions cannot surface with the Genitive of Negation. It has been claimed (Babby 1980 and others) that prepositions assign semantic Case to their complements, and as such cannot raise to a position where they would receive structural Case. Support for the claim that prepositions assign semantic Case comes from prepositions that express motion; whereas prepositions that express motion toward require accusative Case, prepositions that express motion awayfrom generally require Genitive Case. It has been suggested (Neidle 1988:12) that possibly prepositions assign a partially specified feature bundle, and that the second feature value is determined by the directionality of motion.
- 21 Note that the word order within PPs demonstrates that the prepositional complement does not raise (...)
- 22 This explanation can be extended to the lack of Genitive of Negation on the complements of the cla (...)
73Under our analysis, an explanation for the fact that the Genitive of Negation does not appear on prepositional complements does not require reference to a structural/semantic Case distinction. The lack of the Genitive of Negation is expected because the complement of P is Case-licensed by the preposition itself, within PP. 21 In this way, we do not find the Genitive of Negation on the object of a preposition for the same reason that we do not find the Case associated with the verb on the object of a preposition in other languages; the P determines the lowest possible Case-configuration for its complement.22 It is only with the internal arguments of [+tense] verbs that Spec/NegP is the lowest potential Case configuration. This is illustrated in the LF structure given in (49).
- 23 Chvany 1975:156, Babby 1980:105ff, Neidle 1988:53ff.
74A major syntactic difference between sentential and constituent negation is that only the former generates a NegP. Constituent negation is not the head of a phrase NegP, but a lexical item adjoined to a phrase. By positing such a close relationship between the presence of a NegP and the possibility of the Genitive of Negation, this analysis predicts that the Genitive of Negation should not be available under constituent negation. This is upheld by the data. The observation that the Genitive of Negation is incompatible with constituent negation has been made by many researchers.23
75The distinction between sentential and constituent negation may not be evident from surface word order alone. Hence, pragmatically, one can choose to negate only the verb in a sentence, rather than the entire VP, by contrastively stressing the verb. This can be seen in the English sentence; “I didn’t see the film.” Sentences of this type can be distinguished from sentential negation in that they allow violations of the downward entailment requirements of sentential negation. In this way, while the truth of the sentence “I didn’t eat a green vegetable” logically entails the truth of the sentence “I didn’t eat a zucchini”, it is permissible to state: “I didn’t eat a green vegetable, I ate a zucchini”. The latter are therefore instances of constituent negation, not sentential negation, and our analysis predicts that they should not permit the Genitive of Negation. This prediction is correct; such sentences will not allow the Genitive of Negation on their direct object, even if the word order is identical to the unstressed (sentential negation) version where Genitive of Negation is permitted. The appearance of the Genitive of Negation forces a reading where negation has sentential scope.
(50) on ne prosmatrivaet sts’ju a citaet
he-NOM NEG looks over article-ACC (*GEN] but reads
He does not look over the article, but reads it.
76Neidle supports this distinction between sentential and constituent negation and the possibility of the Genitive of Negation with reference to Academy of Sciences of the USSR Grammar (1980, vol. 2:417) where it is stated that an Accusative object is required when one verb is negated in contrast to another.24 On the assumption that sentential negation differs from constituent negation in generating a NegP, this supports the dependency of the Genitive of Negation on the functional projection NegP.
- 25 Neidle (1988:86 fn. 14) mentions Peskovskij 1956:366-367, Pesetsky 1981:8, Babby 1980:14ff, Karcev (...)
77It has been noted by several researchers that sentences involving the Genitive of Negation most naturally contain these phrases post-verbally, even if they are the negated equivalents of sentences where the same argument would occur in S-initial position, bearing Nominative Case.25 Thus, the examples given below show the unmarked word order of a phrase bearing Nominative Case and a phrase bearing the Genitive of negation in a sentence with an unaccusative verb.
(51) ni odna devuska ne prisla
not one-NOM girl-NOM NEG came-fem-sg
Not one girl came.
(52) ne prislo ni odnoj devuski
NEG came-3rs-neut.sg. not one-GEN girl-GEN
Not one girl came.
78This is exactly what is expected under this analysis. We have maintained that Nominative Case is checked in Spec/TP at S-structure (triggering agreement). Since we have maintained that Genitive Case reflects LF Case-licensing in Spec/NegP, a position lower than the S-structure position of the verb in T, the Genitive NP can only be S-initial if preposed by scrambling, creating a marked structure which requires some kind of emphatic stress.
79In the intransitive sentences we have discussed, arguments of unaccusative and passive verbs may surface with either Nominative Case or Genitive Case. Subject agreement is analyzed as the reflex of Case-licensing in Spec/TP. Given that the Genitive of Negation is checked in Spec/NegP, not Spec/TP, Genitive NPs should not trigger agreement on the verb, even if they are superficially in “subject” position, preceding the verb. These facts are born out; Genitive NPs, even when the sole NP in an intransitive S, do not trigger agreement on the verb.
(53) griby zdes’ rastut
mushroom-NOM-PL here grow-3PLMushrooms grow here.
(54) gribov zdes’ ne rastet
mushrooms-GEN-PL here NEG grow-3SGMushrooms don’t grow here.
80This analysis rests on the following claims about the syntax of Russian. First, I have proposed that sentential negation in Russian (unlike constituent negation) generates a NegP, which subcategorizes for VP. I have proposed that the Genitive of Negation is related to Case-licensing in Spec/NegP, which is made possible by the combination of the [+V] feature of the verb and the [+Case] feature of the head of NegP. The ability to assign Genitive Case is therefore restricted by structural constraints; it is only available to internal arguments of the verb. With these claims, we have developed an analysis that accounts for the following facts concerning the Genitive of Negation (GEN) in Russian without additional stipulations required.
1. GEN on an NP otherwise assigned Accusative in negated transitive sentences
2. GEN on an NP otherwise assigned Nominative in unaccusative and passive sentences
3. Ungrammaticality of GEN on an agentive NP
4. Lack of subject agreement with a GEN NP
5. Unmarked position of a GEN NP following the verb
6. Ungrammaticality of GEN on complements of prepositions
81In the following section, I propose an analysis for the apparent definiteness effects exhibited by these constructions. The initial analysis will be driven by the facts described by Neidle (1988), which is supported by my informant work. I will however address Pesetsky’s (1982) findings as well.
82Phrases marked with Genitive under negation have the option of surfacing with Accusative Case or Nominative Case, depending on the verb. However, these alternatives are associated with distinct interpretations; interpretations which differ depending on whether the alternation is between Genitive and Accusative Case or between Genitive and Nominative Case; in other words, whether the NP in question is the D-structure object in a transitive or an intransitive sentence. Whereas with the D-structure objects of intransitive verbs the non-Genitive option (Nominative) leads to an interpretation that is ambiguous with respect to definiteness, with the D-structure objects of transitive verbs the non-Genitive option (Accusative) forces the definite reading. We have already proposed an account for how the D-structure objects of transitive and intransitive verbs pattern together in their ability to bear the Genitive of Negation. Below, we will propose an account for the fact that there is this asymmetry between these arguments when it comes to interpretation. .
83While the pattern discussed (and analyzed) in Pesetsky (1982) can be illustrated as a uni-directional implicature, with Genitive Case entailing a [-definite] interpretation, the data discussed in Neidle (1988) and supported by my own informant work states that the use of Genitive permits the NP to be interpreted as either [+definite] or [-definite], and the Accusative option is unambiguously interpreted as [+definite]. This is schematized in (55).
(55) Genitive --> [-definite] or [+definite]
Accusative --> [+definite]
84Strong support for Neidle’s claim that the Genitive option is not restricted to indefinite NPs comes from the acceptability of Genitive Case on pronouns, demonstrative NPs, and proper names, none of which can be described as indefinite.
(56) on ne vidal etoj strany
he-NOM NEG saw this-GEN country-GEN
He did not see this country.
(57) ja ne vidal masi
I NEG saw Masa-GEN
I didn’t see Masa.
85Nevertheless, given that the construction is apparently in transition in modern Russian, it is possible that for younger speakers the Genitive of Negation is a more marked option, and correspondingly subject to greater restrictions. Below, I will propose an analysis for Neidle’s paradigm, and then propose an explanation for the greater restrictions discussed in Pesetsky (1982).
86In the analysis I will propose here, I will assume that NPs are lexically marked with a value for the feature [definite]. The actual name that we give this feature is not crucial to the analysis; I will refer to it as [definite] for ease of exposition and (with the caveat that the semantics of the notion (in)definite is still a matter of spirited debate) because it appears to approximate the common property of the NPs in question. Following work by Heim (1982), I will assume that the syntactic component is sensitive to this feature. In what follows, I will argue that the distinction between definite and indefinite NPs drives the exhibited pattern of interpretation and Case-assignment in negated sentences in Russian.
87Given that it will be the specification for definiteness carried by the NP itself that will limit the Case-marking options of NPs in negated sentences, rather than the Case-marking that limits interpretation, the paradigm can be clarified by reversing the direction of entailment of Neidle’s observations. This results in the following pattern.
(58) [+definite] --> Accusative or Genitive
[-definite] --> Genitive
88According to the analysis developed up to this point, in an affirmative [+tense] sentence Accusative Case is assigned in Spec/VP, under government by the trace of the verb-chain in Delta. We have maintained that the presence of the functional category NegP immediately dominated by delta essentially blocks Case-assignment under government while introducing an additional Case position where Case-licensing via Spec-head co-indexation is available. Without further refinement, this predicts that the Genitive of Negation would be obligatory on the direct object in a negated sentence. Indeed, this captures the situation historically in Russian, where any NP that met the structural requirements for the Genitive of Negation was obligatorily Genitive, regardless of whether it was definite or indefinite. In contemporary Russian, however, we see that Accusative Case on the direct object is still an option.
- 26 For Russian, one could follow the linguists referred to above in assuming that FP is generated out (...)
- 27 Alternatively, FP could be related to a DP-type analysis of arguments as in Abney (1987). We could (...)
- 28 It could be argued that this structure is similar to that of Pollock (1989), where AgrP (similarly (...)
89I will argue that there is an additional position in which an NP can be Case-licensed via Spec-head co-indexation, a position where Case is realized as Accusative Case. In much of the recent literature, such an additional Case position for the direct object has been proposed. In Chomsky (1989) and Johnson (1990), this projection is generated outside of the Larsonian VP. An alternative possibility is that this projection, which we will simply refer to as Functional Projection (FP), is generated within DeltaP. I will adopt this second option.26 Let us assume here that FP is generated as complement of Neg.27 The D-structure position of FP in a negated [+tense] transitive sentence is given in (59).28
90The verb raises through the head of FP, NegP, and DeltaP, to T. The chain of the raised verb creates two potential Case-positions for the direct object at LF; Spec/FP and Spec/NegP. The properties of FP are considered in the following section.
- 29 Attributing restrictions as to which NPs raise to this projection is not without precedent in the (...)
- 30 If we extend the proposal made for Literary Welsh that agreement is realized as low as possible to (...)
91First, I will maintain that this position is an available Case-licenser only for [+definite] NPs.29 Since Case-licensing in this position is restricted to definite NPs, and is the only means for the direct object to receive Accusative Case in a negated Russian sentence, then the corresponding interpretations of these sentences are as expected. Accusative NPs that are D-structure objects of negated transitive verbs must be interpreted as unambiguously definite; if they were indefinite they would not have the option of being Case-licensed in this position. Further, since no such restrictions in terms of definiteness are attributed to Spec/NegP, the fact that a phrase in the Genitive of Negation is not disambiguated with respect to definiteness is also expected.30
92Thus, this proposal accounts quite naturally for the asymmetry between the internal arguments of intransitive and transitive verbs with respect to interpretation; the observation that only an Accusative marked internal argument in a negated sentence is unambiguously interpreted as definite. In both Pesetsky’s (1982) and Neidle’s (1988) analyses, the fact that the non-Genitive option is interpreted differently depending on whether it is marked Accusative or Nominative is problematic. The fact that the Accusative option is unambiguously definite is consistent with this analysis, since it is only this position that is restricted to definite NPs.
93The first fact to note about the FP option is that it is unavailable for the arguments of passive and unaccusative verbs, which show only a Nominative-Genitive alternation in negative sentences.
(60)*ne suscestvuet takuju stranu
NEG exists-3sg such-ACC country-ACC
There does not exist such a country.
(61)ne suscestvuet takoj strany
NEG exists-3sg such-GEN country-GEN
There does not exist such a country.
- 31 Recall that a verb in a [+tense] sentence, even if specified as [+Case], cannot license an NP in i (...)
94We have suggested that for an element to license an NP in its Specifier it requires positive values for the features [Case] and [V]. These verbs are not [+Case]. Above, we argued that the head of NegP can license an NP in its Specifier because it is inherently specified as [+Case], and therefore does not rely on this feature from the verb. The lack of Case-licensing in Spec/FP in sentences with unaccusative or passive verbs suggests that the head of FP is not specified for the feature [Case]. As such, it is entirely dependent on inheriting this feature from a lexical item that raises into it..31 In an affirmative sentence, this forces the internal argument of such a verb to raise to Spec/TP to be licensed with Nominative Case.
95Thus, the head of FP is distinct in this respect from the head of NegP, which is inherently specified as [+Case] and can license an NP in its Specifier if a V moves into it, regardless of the verb’s specifications for the feature [Case].
96In the following section, I present an analysis of the data presented by Pesetsky (1982), where the Genitive of Negation unambiguously denotes indefiniteness. As above, the basic claim on which the analysis rests is that the Genitive of Negation reflects Case-checking in Spec/NegP. The additional claim will be that only NPs which are negatively specified for the feature [definite] have this option. We will derive this by an additional restriction on Case-licensing in Spec/NegP; only an NP which has copied its referential index onto the head of NegP can be Case-licensed by Spec/head co-indexation at LF.
97In Heim’s (1982) dissertation, all NPs are assigned a referential index. Heim treats indefinite NPs as variables that require a C-commanding operator. Quantifier indexing operates to copy the referential index of every indefinite NP as a selection index onto the lowest c-commanding operator. For Heim, this occurs after S-structure, at the level of the grammar where the interpretive component operates. Crucially, only indefinite NPs copy their indexes onto an operator. In what follows, I will suggest that index-copying is a necessary (though not sufficient) condition for LF Case-checking in Spec/NegP. It is the requirements of this condition that rule out the Genitive of Negation on [+definite] NPs, and force the corresponding [-definite] reading of NPs in the Genitive of Negation.
- 32 See Progovac 1988 for this treatment of negative operators.
98Note that under the assumption that the negative operator adjoins to TP at LF, C-commanding all NPs in its binding domain,32 any indefinite NP in a negative sentence can copy its index onto the negative operator. However, index-copying is not a sufficient condition for the Genitive of Negation. As we have seen in this chapter, not all [-definite] NPs can surface with Genitive of Negation. This follows from the fact that Case-assignment is subject to stricter structural constraints than co-indexation; specifically, an NP bearing the Genitive of Negation must raise to Spec/NegP at LF, and there are strict structural constraints on which NPs may undergo this movement.
99Armed with these assumptions about the relationship between indefinite NPs, index-copying, and Case-licensing in Spec/NegP, the restriction of the Genitive of Negation to indefinite NPs can be handled as follows.
100As in the preceding analysis, an agent NP cannot be Case-licensed in Spec/NegP because it would have to lower to do so. Instead, it raises to the Specifier of TP and is Case-checked there at S-structure, reflected in person-number agreement on the tensed verb. An indefinite D-structure object has two options for LF Case-checking; raising to Spec/FP or to Spec/NegP. By assumption, only an indefinite NP copies its index onto the negative operator, and only an NP that has copied its index onto the negative operator can be Case-checked in Spec/NegP.
101Furthermore, given that Spec/FP is limited to [+definite] NPs, indefinite NPs of transitive verbs have only one option, which corresponds to Genitive Case; Case-licensing in Spec/NegP. If definite, no index-copying takes place, Case-licensing in Spec/NegP is not permitted and the NP must be Case-licensed in Spec/FP.
102Since Genitive Case-assignment in Spec/NegP is only possible if index-copying between the operator NEG and an NP has taken place, and since this co-indexation is restricted to [-definite] NPs, the genitive object is unambiguously interpreted as indefinite, as desired. Likewise, since Spec/FP is restricted to [+definite] NPs, an accusative NP is unambiguously interpreted as definite.
103The derivation in such sentences proceeds exactly as with transitive sentences, except that the option of Case-licensing in Spec/FP is ruled out for verbs that are not specified [+Case], and an additional option, Case-licensing in Spec/TP, is available. An indefinite D-structure object copies its referential index onto the c-commanding negative operator. This permits Genitive Case-checking in Spec/NegP. If [+definite], index copying does not take place and the NP must raise to Spec/TP at S-structure, where it surfaces with Nominative Case and triggers agreement on the tensed verb. An agentive argument of an intransitive verb, whether definite or indefinite, cannot lower to Spec/NegP. Accordingly, it raises to Spec/TP and surfaces with Nominative.
104Given that Spec/TP is not restricted to [+definite] NPs, this allows two Case positions for an indefinite object of an unaccusative or passive verb; Nominative Case-checking in Spec/TP and Genitive Case checking in Spec/NegP. This predicts that the Nominative option allows both the definite and indefinite interpretation, while the Genitive option is unambiguously interpreted as indefinite for arguments of both transitive and intransitive verbs, which is consistent with Pesetsky’s data. Further, we account for the unambiguously definite reading associated with the Accusative option. These facts are summarized below.
(62)UNACCUSATIVE AND PASSIVE VERBS
GEN - [-definite]
NOM - [+definite] or [-definite]
GEN - [-definite]
ACC - [+definite]
105This analysis, which posits indefiniteness as a necessary initial condition for the Genitive of Negation, is plausible and perhaps captures best the language of younger speakers for whom the Genitive of Negation is a more marked option. A weakness in the analysis is that the main supporting claim, that Case-licensing in Spec/NegP at LF requires index-copying, is otherwise unmotivated. I will tentatively suggest that this must relate to an additional restriction on the ability of the head of NegP to Case-license an element in its Specifier; a feature-matching requirement that can only be satisfied by referential index-copying. In the following section, we examine the interaction between sentential negation and Case in Colloquial Welsh (CW).
106CW is sufficiently different from LW with respect to negation to require a brief introduction. First, the negative particle that we find preceding the tensed verb in Literary Welsh no longer surfaces as an independent lexical item. Its continued presence, however, can be seen in the effect it has on the tensed verb. When preceding a vowel-initial verb, it appears in reduced form as a prefix (d-). When preceding a consonant-initial word, it tends to be dropped altogether, but its presence can be noted in the form of consonant mutation on the verb. The verb form in (63a), palodd (dug-3sg), obligatorily appears as phalodd in a negated sentence in LW, under the influence of the preceding negative particle. Even though the negative particle itself is dropped in CW, the consonant undergoes the same initial consonant mutations as though it were overt. Second, as shown in the example, negation in CW requires an additional, sentence medial negative marker, dim.
(63)(a) LW/CW: Palodd Sion yr ardd.
dug-3sg Sion the garden
Sion dug the garden.
(b) Neg-LW: Niphalodd (*palodd) Sion yr ardd.
NEG dug-3sg Sion the garden
Sion did not dig the garden.
(c) Neg-CW: Phalodd (*palodd) Sion ddimo ’r ardd.
NEG-dug-3sg Sion NEGof the garden
Sion didn’t dig the garden.
- 33 The possibility of analyzing dim as generated in the specifier of NegP is considered and rejected (...)
107The negative marker dim follows the tensed verb and the subject, but precedes the uninflected verb and all other verbal elements in the VP, including the aspectual markers bod wedi and bod yn. The sentence below illustrates the position of dim with respect to the elements in VP. I will assume here that the negative marker dim is generated as an adverbial adjoined to the lower VP.33
(64)Mae Sion [wedi bod yn canu].
is Sion [PERF be PROG sing]
Sion has been singing.
(65)’Dyw Sion [[ddim]wedi bod yn canu].NEG-is Sion [[NEG]PERF be PROG sing]
Sion has not been singing.
108The third difference between negative sentences in CW and LW, illustrated by the sentences in (66)-(67), is that in CW we find that the negative counterparts of simple transitive sentences require that a preposition be inserted before the direct object. The same sentence in LW requires no such preposition.
(66)Phalodd Sion ddim *(o) ’r ardd.
NEG-dug-3sg Sion NEGP the garden
Sion didn’t dig the garden.
(67)Ni phalodd Sion yr ardd.
NEG dug-3sg Sion the garden
Sion didn’t dig the garden.
109In CW, this preposition is obligatory for direct objects in VSO structures; in other words, direct objects of [+tense] verbs. In periphrastic sentences, however, negation does not appear to affect the assignment of Case to the direct object.
(68)Oedd Sion yn palu’r ardd.
was-3sg Sion PROG dig the gardenSion was digging the garden.
(69)’Doedd Sion ddim yn palu ’r ardd.NEG-was Sion NEG PROG dig the gardenSion wasn’t digging the garden.
110Accordingly, preposition insertion in negated sentences distinguishes direct objects in periphrastic constructions from direct objects in simple VSO constructions. Recall that these arguments are also distinguished with respect to Case in LW; direct objects in simple constructions are exceptionally Case-licensed under government, but periphrastic direct objects are Case-licensed in a Spec/head configuration with the untensed verb.
111While it appears at least initially intuitive that it is the secondary negative marker dim that is interfering with Case-assignment to the direct object, this analysis will take the position that it is not the presence of dim but rather the lower base-position of NegP in CW that is problematic for Case-licensing under government.
- 34 The claim that the head of NegP is an affix suggests that it should no longer create a barrier for (...)
112As in Russian, I will assume that both T and NEG are affixal in CW. In LW, where the head of NegP is an independent lexical item, movement into NEG creates an adjunction structure. In CW, movement into NEG is substitution into an empty slot.34 Also as for Russian, NegP will be generated below Delta-P. In tensed sentences in CW, the V raises via head-to-head movement through NEG to T, both instances of substitution. The secondary negative element, dim, does not undergo raising, and its S-structure position reflects its base position.
- 35 Evans 1964, Williams 1980:153.
113Recall that the base position for NegP is argued to be above TP in standard literary Welsh. While arguments for viewing the base-position as a locus of parametric variation are documented in the literature, it may seem initially implausible that CW differs from LW in this respect. Consider, though, the historical development of sentential negation in CW. Whereas LW reflects a ’frozen’ form of Welsh, representing the language at the time when the Bible was initially translated into Welsh, CW has evolved naturally from that point. The parallels between CW and contemporary French mentioned above extend to the historical development of the post-verbal negative markers. The post-verbal marker pas in French initially carried no independent negative force. It eventually took on negative content through its frequent association (as an emphatic particle) with pre-verbal ne. The Welsh post-verbal negative marker dim also originally carried no independent negative content; in Middle Welsh, it meant thing. It was used as an intensifying particle in negative constructions, and gradually acquired negative content through its frequent association with the pre-sentential negative particle ni(d).35
- 36 For a discussion of this process, see Horn 1989:452-459, Payne 1985:224, and Jespersen 1917, 1924/ (...)
- 37 They are commonly expressions denoting a small or negligible quantity (i.e. “Not a thin dime”)Horn (...)
114This reflects a quite general process cross-linguistically, noted by Jespersen (1917). This process, known as Jespersen’s Cycle, refers to a cyclical pattern where the negative marker associated with the tensed verb is gradually weakened into a verbal pro-clitic, then reinforced by Negative Polarity Item (NPI) minimizers, and ultimately replaced by its reinforcement.36,37 I will tentatively conclude here that the base position of NegP in CW reflects a historical development from LW; where the medial marker has replaced the initial marker in its role as sentence negator, and its increased salience is reflected in the lower base position of NegP.
115Next, I will try to show that these hypotheses about the differences between NegP in CW and LW provide an explanation for the differences involving the interaction between sentential negation and Case-licensing; specifically, the required preposition found in negated transitive sentences in CW.
116Recall that in LW it was claimed that the direct object of a tensed verb is Case-licensed under government, by the member of the V-chain in delta. The fact that Case-is assigned under government rather than by Spec/head co-indexation is reflected in the lack of agreement marking. As shown above, in LW, negation does not interfere with Case-licensing of the direct object, whether pronominal or non-pronominal. What we find in CW is strikingly different. If the clause is negated then the preposition o must occur immediately preceding the VSO direct object, whether pronominal or non-pronominal.
(71)Phalodd Sion ddimo’r ardd.
NEG-dug-3sg Sion NEGP-the garden
Sion didn’t dig the garden
(72)Phalodd Sion ddim ohonofo.NEG-dug-3sg Sion NEG P-3sgitSion didn’t dig it.
- 38 The fact that FP is base-generated in distinct positions in Russian and Colloquial Welsh is an una (...)
117This data suggests that the ability of the verb to assign Case under government is blocked by the presence of an intervening NegP, as in Russian, and a direct object NP can only be Case-licensed within FP. Accordingly, we will assume that the Case-licensing head in these negative constructions in CW is FP, the head of which is [o]. Given that the preposition, if overt, follows the head of FP, and our assumption that NPs in Welsh do not raise to their Case positions until LF, this entails that the NP is generated inside of FP.38 The S-structure for a negated transitive sentence in Colloquial Welsh is given in (73).
118However, in at least one dialect of Colloquial Welsh, the preposition is not required for all direct objects in negated sentences. These facts are discussed below.
- 39 I will not address here the interesting facts discussed by Awbery (1990) concerning the relative p (...)
119According to Awbery (1990), in Pembrokeshire Welsh there is a three-way distinction regarding the presence of sentential negation and the need for the inserted preposition. Indefinite NPs never appear with a preposition, pronouns always occur with a preposition, and definite NPs may or may not appear with a preposition. She summarizes the situation for object NPs as shown in (74).39
120If we analyze the lack of preposition as indicating Case-licensing in Spec/NegP, and the presence of the preposition as indicating Case-licensing in Spec/FP (where the head of FP is the lexical item [o]), the distribution of NPs parallels to a remarkable degree the Russian data described by Pesetsky (1982). In that dialect, the Genitive of Negation (Case-licensing in Spec/NegP) is restricted to [-definite] NPs, and the Accusative option (Case-licensing in Spec/FP) is restricted to [+definite] NPs. However, these data suggest an even stronger relationship between the specification for definiteness and an NP’s Case-licensing: pronominals cannot be Case-licensed in Spec/NegP, [-definite] NPs must be Case-licensed in Spec/NegP, and [+definite] NPs can be Case-licensed in either position.
- 40 A possibility I am currently exploring is that this can be related to the operation of the NEG-Cri (...)
121One way that this can be realized is by claiming that the index-copying mechanism that takes place between a [-definite] NP and negation does not merely permit but forces LF-raising to Spec/NegP in Colloquial Welsh.40
122In this paper we have explored some interactions between sentential negation, Case, agreement, and (in)definiteness within a framework which postulates that the inventory of functional categories available cross-linguistically includes Negation Phrase. We have seen evidence of the head status of sentential negation and of the availability of an additional Specifier position associated with negation that is a potential Case-position for NPs. While in all three languages studied the head of NegP can only Case-license an NP if a [+verbal] element has raised into it, I have argued that the negative head in Russian is inherently [+Case], allowing it to Case-license an NP in its Specifier even when the verb is not a Case-assigner.
123Case-licensing in Spec/NegP is associated with changes in word order and Case-marking in Russian (Genitive Case rather than Accusative or Nominative Case), by changes in the realization of agreement and by agreement features on the negative head itself in Literary Welsh, and by the possibility of dropping an otherwise-required preposition in Colloquial Welsh. I have further argued that there is a relationship between the head of NegP and indefinite NPs that derives from an index-copying mechanism which is restricted to [-definite] NPs. Thus, not only the realization of Case-marking but also other restrictions on which NPs can be Case-licensed in the specifier of a given maximal projection may be determined by the properties of its head.