-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathQUESTIONS
More file actions
204 lines (137 loc) · 6.7 KB
/
QUESTIONS
File metadata and controls
204 lines (137 loc) · 6.7 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
REFL-acc
Reflexive pseudopassive
Naveliko se [je] riba jučer jela
((In-large-quantities REFL) [AUX]) (fish-NOM (yesterday eaten))
Fish was eaten in large quantities yesterday.
REFL transform the verb's COMPS list, just like passivization:
[EATER(nom), EATEN(acc)]
becomes
[ EATEN(nom)]
While passivization can be treated as a lexical transformation at
the point of recognising passive participle morphology,
pseudopassivized verb has active morphology, and is only transformed
when put together with the REFL. However, in this scenario, "riba"
(fish-NOM) cannot be recognised, since at the time of binding of
"fish" and "yesterday eaten", the information from REFL is not yet
available.
Alternately, allow the fish to be bound to the nominative slot? On
pure syntax level, deleting a non-nominative slot would then suffice
- But how to expunge its index from the verb's meaning when REFL is
finally seen?
Or - generate all verbs in both ways, just in case? One of them must
be bound to pseudopassive REFL, the other can't. This might be the
best option, but it does double processing time for each active
verb.
Desiderative REFL causes a similar transformation:
[EATER(dat), EATEN(nom)]
However, the dative is also otherwise always an option, in the
benefactive/possessive meaning, and thus does not cause the problems
above. However, the problem of connecting the correct
(desiderative) meaning where it is intended still remains. Also, the
desiderative reflexive abolishes the nominative subject.
a) Jede se žgance
b) Jedu mi se žganci
Reciprocal REFL does too
It modifies the meaning of the verb, and deletes the accusative
slot.
Lexically required REFL and true reflexive
They do not require special considerations. Lexically required verbs
will carry an appropriate REFL feature that needs to be satisfied by
binding with a REFL. True reflexives will just bind with a REFL as
an argument.
RECIPROCAL
Ivan(nom) i Ivana(nom) udaraju jedan(nom) drugoga(acc)
Ivan(nom) i Ivana(nom) rukuju se jedan(nom) s drugim(ins)
Upoznajem Ivana(acc) i Ivanu(acc) jednog(acc) s drugim(ins)
John and Joanna hit each other
John and Joanna shake-hands each with other
I introduce John and Joanna each with other
The reciprocal position will be filled with "drugi" in its proper
case, and immediately before that constituent, "jedan" will be
inserted, with case corresponding to the antecedent's. Gender and
number will be the same for "jedan" and "drugi": gender corresponds to
the general rule for collections in Croatian (masculine if mixed);
number is pragmatically determined (can be both singular and plural,
depending on whether the antecedent can form subgroups or not), and
thus outside the scope of syntax.
LONG-DISTANCE DEPENDENCIES
In cases where COMPS arguments are missing, they can be put into the
STORE at the top of the clause, if there is exactly one COMPS missing.
For example:
a) Vidim djevojku koja jede jabuku
I-see girl who eats apple
"jede jabuku" has noun-nom still in COMPS. It can be transfered to
INHER\REL, which will be bound to "koja". "koja" is CASE\nom, MOD\noun
and TO_BIND\REL\(noun & CASE\nom).
b) Vidim jabuku koju djevojka jede
I-see apple who girl eats
Same; "koju" is CASE\acc, MOD\noun, TO_BIND\REL\(noun & CASE\acc)
c) Vidim djevojku čiju jabuku jedem
I-see girl whose apple I-eat
"čiju" (whose) is MOD\[noun & CASE\acc], CASE\acc, TO_BIND\REL\(noun &
CASE\acc). TO_BIND is inherited from either HEAD or non-HEAD, so "čiju
jabuku" gets bound by HEAD-MOD schema, still having TO_BIND - and then
that binds the trace.
d) Vidim jabuku koju mi djevojka jede
I-see apple which to-me girl eats
Same as d). "mi" clitic does not have a COMPS entry, so it is bound as
a modifier. TODO: how does the possessive/benefactive meaning enter
here?
e) Vidim djevojku s kojom jedem jabuke
I-see girl with who I-eat apples
"s" (with) is COMPS\(noun & CASE\ins). "kojom" is CASE\ins,
TO_BIND\REL\(noun & CASE\ins), MOD\noun. "s kojom" is bound by
HEAD-COMP schema, and "s kojom" thus has TO_BIND\REL\(noun & CASE\ins)
and MOD\noun (both of which would have to be inherited from "kojom").
f) Koga vidim da jede jabuku?
Whom I-see that eats apple?
"koga" (whom) binds the noun-animate trace at the sentence top. TODO:
How is the sentence top detected (also important for clitic attachment
rules)
g) Čiju jabuku vidim da jede?
Whose apple I-see that he-eats?
Same as c), but with a question-variant of "čiju" instead of the
relativizing one.
h) Jabuku vidim da jede.
Apple I-see that he-eats.
"jabuku" is topicalized here; the trace generated at "jede" sentence
top is bound at the overall "vidim da jede" sentence top.
PREPOSITIONS - MARKING-MARKED or HEAD-COMPLEMENT?
Long Distance Dependencies example e), where most important
information must be inherited from "kojom" (who), suggests "kojom" is
a head, not "s" (with). Also, it would be more intuitive specifying
verb complements (CASE\gen & MARKED\od, instead of PFORM\od_gen).
On the other hand, how would non-complement prepositions contribute
meaning, and become modifiers?
CONJUNCTIONS
Intuitively, coordinating conjunctions are symmetrical. However, this
approach might work: in A CONJ B, CONJ modifies A and complements for
B. Except for semantics, again.
CLITICS
Da li si joj ga se rijesio?
Da - particle, clitic host - y/n question marker
li - particle, clitic - question marker (obligatory with "da" marker)
si - AUX ("you were"/"you did"); AUX-PRED or HEAD-COMP
joj - benefactive ("for her"); HEAD-MOD
ga - genitive (ablative) ("of him"); HEAD-COMP
se - REFL, required by the lexically reflexive verb; HEAD-COMP
rijesio - verb participle ("got rid of") - HEAD.
How are the particles bound? As HEAD-MOD?
NOMINAL SENTENCES AND SCRAMBLING
Ivan je covjek
John is man
is not same as
Covjek je Ivan
man is John
and especially
Mesar zeli biti astronaut
butcher he-wants to-be astronaut
is not the same as
Astronaut zeli biti mesar
astronaut he-wants to-be butcher
ADVERBISATION OF ADJECTIVES (and pasv. parts.) IN NOMINATIVE NEUTER SINGULAR
UNRESOLVED SENTENCES
Povijesno gledano, Zagreb je izrastao iz dva naselja na susjednim
mod> mod> comp> <aux head <mod <comp
brdima, Gradeca i Kaptola, koji čine jezgru današnjeg Zagreba,
njegovo povijesno središte.