The Category P Features, Projections, Interpretation
The Category P Features, Projections, Interpretation The Category P Features, Projections, Interpretation
143 ‘directional’ in (56), (57) is P C . Accordingly, it checks the structural Case (Accusative) of the nominal. 36 4.3.3.2 The unaccusative behavior of unergative verbs of motion: As opposed to typical Directional verbs such as send, verbs such as rac (‘ran’) or nafal (‘fell’) can be construed as Directional, if a Directional PP is added (59), or they can be modified by a Locative PP adjunct (60): (59) a. ha-kelev rac la-gina the-dog ran to+the-garden “The dog ran to the garden.” b. ha-matbe’a nafal la-ma’im the-coin fell to+the-water “The coin fell into the water.” (60) a. ha-kelev rac ba-gina the-dog ran in+the-garden “The dog ran in the garden.” b. ha-matbe’a nafal ba-gina the-coin fell in+the-garden “The coin fell in the garden.” Note that ran, as opposed to fell, for instance, is an unergative one-place verb. However, it has been noted by several researchers (Hoekstra and Mulder 1990, Ackema 1995, and references cited therein) that unergative verbs of motion such as ran, exhibit the typical unaccusative properties (e.g. auxiliary selection in Dutch), when combined with a Directional PP (59a). 37 Roughly speaking, a verb is classified as unergative if its subject is an external argument (e.g. ran), but as unaccusative if it is merged internally, surfacing as the subject upon syntactic movement (e.g. fell) (Perlmutter 1978, Burzio 1986). 36 But see Reinhart and Reuland (1995) and reference cited therein where Accusative in the Directional constructions is assumed to be inherent, rather than structural. 37 The arguable change from unergative to unaccusative is accompanied by the aspectual shift from state to event (Hinrichs 1985).
144 Modification by possessive datives is used as a reliable test to detect internal arguments in Hebrew. Borer and Grodzinsky (1986) observe that possessive datives can only modify internal arguments. Hence, they can serve as possessors to subjects of unaccusatives (61a), but not to subjects of unergatives (61b). (61) a. ha-sfarim naflu le-rina the-books fell to-Rina b. *ha-kelev šaxav le-rina the-dog lay to-Rina The verb rac (‘ran’) indeed patterns with unaccusatives when it occurs with a Directional PP (62a), but with unergatives when it is not (62b). In (62a) the possessive dative le-mi (‘to whom’, ‘whose’) modifies the subject (the dog, i.e. whose dog), indicating that it is the internal argument of rac (‘ran’). In (62b) this reading is unavailable, showing clearly that the subject is the external argument of rac (le-mi (‘to whom’) can only modify the locative adjunct in the garden, i.e. whose garden): (62) a. le-mi ha-kelev rac la-gina to-whom the-dog ran to+the-garden b. le-mi ha-kelev rac ba-gina to-whom the-dog ran in+the-garden Given this, the fact that the same verb is both unergative (60a), (62b) and unaccusative (59a), (62a) is puzzling. Let me sketch briefly how the puzzle can be accounted for, based on the mapping generalizations in Reinhart (2000, 2001) (Tali Siloni p.c.). Let us assume that a verb like run (and arguably also verbs like sleep, stand or sit) does not assign an Agent theta-role, but rather it assigns a Theme. This is supported by the fact that this theta-role can be realized by non-human DPs (The program runs smoothly; The time runs quickly). Further, it is reasonable to assume that run has two lexical entries: (i) run: Theme ([-c-m]) (ii) run: Theme ([-c-m]), Goal ([-c])
- Page 111 and 112: 92 P device involves the syntactic
- Page 113 and 114: 94 Note that the options in (103) d
- Page 115 and 116: 96 adjacency requirement between th
- Page 117 and 118: 98 would be [-c] PP-verbs in Englis
- Page 119 and 120: 100 Appendix A: Residual issues The
- Page 121 and 122: 102 (A.6) a. mabat-o nadad (motion)
- Page 123 and 124: 104 Given its theta-grid, hikša is
- Page 125 and 126: 106 Appendix B Table 1. 70 Hebrew P
- Page 127 and 128: 108 Table 2. Some properties of the
- Page 129 and 130: 110 67. serev (le-) [+c+m] [-c] - +
- Page 131 and 132: 112 4. Locative, Directional and Da
- Page 133 and 134: 114 is distinct from both the Dativ
- Page 135 and 136: 116 Let me illustrate briefly the e
- Page 137 and 138: 118 inability of the Dative PP to d
- Page 139 and 140: 120 In principle, (16) can have eit
- Page 141 and 142: 122 from natan, but from the embedd
- Page 143 and 144: 124 Both (22a) and (22b) are possib
- Page 145 and 146: 126 (ii) Binding In the Hebrew Dati
- Page 147 and 148: 128 Let us assume that in (29a) the
- Page 149 and 150: 130 4.3 The Directional P Zwarts an
- Page 151 and 152: 132 (36) a. dan šalax praxim (le-r
- Page 153 and 154: 134 The incompatibility of Dative p
- Page 155 and 156: 136 (45) ha-tiyul le-hodu haya me
- Page 157 and 158: 138 rather an (elided) NP modified
- Page 159 and 160: 140 Consider now the English and Ru
- Page 161: 142 the Accusative Case in (56) is
- Page 165 and 166: 146 use (63c), me- (‘from’) def
- Page 167 and 168: 148 combination with a path denotin
- Page 169 and 170: 150 4.4.1 Evidence for the Small Cl
- Page 171 and 172: 152 4.4.2 Projections of a Locative
- Page 173 and 174: 154 (81) a. on pologayets y a na Sa
- Page 175 and 176: 156 4.4.3 Modification by Locative
- Page 177 and 178: 158 c. *ha-sefer še-/ašer al ahav
- Page 179 and 180: V. 48 The remaining alternatives, (
- Page 181 and 182: 162 b. ha-ec (še-) ba-ya’ar kara
- Page 183 and 184: 164 (99) a. ha-sefer (??/*hu) al ha
- Page 185 and 186: 166 Appendix: On some differences b
- Page 187 and 188: 168 (A.2) VP Agent V’ V VP Goal i
- Page 189 and 190: 170 Consequently, they are predicte
- Page 191 and 192: 172 5. P pred in object gap constru
- Page 193 and 194: 174 . Extending the proposal to Eng
- Page 195 and 196: 176 . Based on previous work, defin
- Page 197 and 198: 178 . specifier of an NP is the pos
- Page 199 and 200: 180 . possibility that in principle
- Page 201 and 202: 182 (19) a. the destruction of the
- Page 203 and 204: 184 . 5.2.5 Adverbial modification
- Page 205 and 206: 186 b. ha-yeled kal havana the-boy
- Page 207 and 208: 188 . Before I discuss the lexical
- Page 209 and 210: 190 . Passivization is taken to inv
- Page 211 and 212: 192 . already ΘSat ARB (see (37)).
144<br />
Modification by possessive datives is used as a reliable test to detect internal<br />
arguments in Hebrew. Borer and Grodzinsky (1986) observe that possessive datives can<br />
only modify internal arguments. Hence, they can serve as possessors to subjects of<br />
unaccusatives (61a), but not to subjects of unergatives (61b).<br />
(61) a. ha-sfarim naflu le-rina<br />
the-books fell to-Rina<br />
b. *ha-kelev šaxav le-rina<br />
the-dog lay to-Rina<br />
<strong>The</strong> verb rac (‘ran’) indeed patterns with unaccusatives when it occurs with a<br />
Directional PP (62a), but with unergatives when it is not (62b). In (62a) the possessive<br />
dative le-mi (‘to whom’, ‘whose’) modifies the subject (the dog, i.e. whose dog),<br />
indicating that it is the internal argument of rac (‘ran’). In (62b) this reading is<br />
unavailable, showing clearly that the subject is the external argument of rac (le-mi (‘to<br />
whom’) can only modify the locative adjunct in the garden, i.e. whose garden):<br />
(62) a. le-mi ha-kelev rac la-gina<br />
to-whom the-dog ran to+the-garden<br />
b. le-mi ha-kelev rac ba-gina<br />
to-whom the-dog ran in+the-garden<br />
Given this, the fact that the same verb is both unergative (60a), (62b) and<br />
unaccusative (59a), (62a) is puzzling.<br />
Let me sketch briefly how the puzzle can be accounted for, based on the mapping<br />
generalizations in Reinhart (2000, 2001) (Tali Siloni p.c.).<br />
Let us assume that a verb like run (and arguably also verbs like sleep, stand or sit)<br />
does not assign an Agent theta-role, but rather it assigns a <strong>The</strong>me. This is supported by<br />
the fact that this theta-role can be realized by non-human DPs (<strong>The</strong> program runs<br />
smoothly; <strong>The</strong> time runs quickly).<br />
Further, it is reasonable to assume that run has two lexical entries:<br />
(i) run: <strong>The</strong>me ([-c-m])<br />
(ii) run: <strong>The</strong>me ([-c-m]), Goal ([-c])